COMPUTER ANALYSIS OF THE FUTURES MARKET
Charles LeBeau and David W.Lucas Editors of Technical Traders Bulletin
IRWIN Professional Publishing
Translation from English VD Gibenko Translation Editors AA Liman, A.M.Ilin Literary Editor IM Dolgopolsky LeBoCh., LukasD.V Computer Analysis of the Futures Markets: Per. from English. - Moscow: Publishing House "ALPINA" 1998 - 304c. ISBN 5-89684-002-0 (Rus) ISBN 1-55623-468-6 (English)
1
The book is written by two prominent specialists in the field of analysis of the financial market Charles LeBeau and David W. Lucas. They are widely known for their striking success in futures trading, detailed knowledge of technical analysis and extensive experience to create trading strategies. "Computer Analysis of the Futures Markets'' is a step by step guide to build and test trading systems and shows how to apply subtle and sophisticated technical studies, which are used by traders to operate in the futures and currency markets around the world. This book is designed for professional traders who already have work experience in different markets, as well as those who have sufficient theoretical knowledge in the field of technical analysis, but lack skills. This book is an invaluable guide to work and help avoid costly mistakes UDC 339.13.017 BBK 65.42 ISBN 5-89684-002-0 (Rus) ISBN 1-55623-468-6 (English) © Charles LeBeau & David W. Lucas, 1992 © The Russian translation, design. "Publishing House" ALPINA ", 1998
INTRODUCTION In this book put lots of work, and it is obviously demonstrates. I have read most of the books on technical analysis failed to lay a solid foundation for those trading techniques and methods that I used in the future. In my home town. New Orleans, any Cook book describing a good recipe for Creole dishes, starting with the instructions: "First, prepare the sauce." The sauce - this is the basis, the foundation of a prescription, and should be paid great attention to its development and use. The chef knows that the dish will fail unless it is cooked right framework to support it. This book, written by the masters of LeBeau and Lucas, is rich in its "sauce." The authors had a combined history and rationale for each of its concepts with detailed technical studies. I have not had a chance to read a book the other, where the basics were so carefully chosen and well laid out. I read with pleasure. Of developed using computer technology research discussed here only the most useful and important. This indicator, which have repeatedly proven their usefulness, passing the test of time and thousands of users. Another score in favor of the authors is that they are not included in the review of those techniques that are on the side of the utility, and the use of which suspiciously tied to the quixotic interpretation of the user. Despite the fact that I've never seen the office of David and Charles, I imagine that on a wall or screen printed in large letters "KISS" (Keep It Simple, Stupid! - Keep it simple stupid before!). This is exactly what they did with his book. Without the slightest deviation from the complex issues that bogged down most of the authors of books on technical analysis, they were able to keep the presentation of the material simple and well organized. The reader is given a full rationale of the basic techniques, and then shows how to apply each study. Chuck and David is not lost sight of the vital information, omits most of the authors on how to avoid dangerous and false trading signals with their costly mistakes and deviations from the correct direction. If you are an active futures trader using the computer in their work, or intend to become one, this book will prepare you to face the real world of futures trading. Tim Slater The President CompuTrac Software, Inc.
From the authors Over the past 15 years to help futures traders have developed a highly intelligent software for technical analysis. Charts that traders had earlier been forced to draw by hand, now updated automatically on a computer screen, which can be instantly reconfigured. Fast and powerful personal computers are now available to anyone with sufficient capital and to cover the risks of futures trading. Cheap, available for immediate use modern software packages allow futures traders to quickly and easily calculate indicators such as Stochastics, convergence and divergence of moving averages, parabolic stop and turning points, 2
and indicators of direction as well as many other informative and technological research, which can be quickly displayed on the monitor by pressing one or two keys. Unfortunately, there is a significant failure between the instructions that accompany the analytical software, and what the user really needs to know to effectively trade. Most descriptions of the software are only intended to teach the user to obtain the correct technical study on the screen and then, at best, offer a paragraph - the other on the main objectives and the use of indicators. This book attempts to bridge the said failure and explain in detail the most popular and useful technical indicators that evaluate computer. Most importantly, we will offer you the practical advice on how to properly use these indicators to futures trading. We are professional traders and consultants registered in futures trading. In addition to cash management in the futures markets, we publish a monthly educational sheet "Technical Traders Bulletin", whose purpose is the exchange of knowledge and ideas between professional traders. Many of the techniques and strategies described in this book are taken from these publications. We are actively traded futures for over 20 years and have used computers for the technical analysis of the availability of the first commercial programs. In the early days of computer analysis of the data and equipment costs us more than $ 3,000 a month. Our computer terminal in Los Angeles had to be connected to the main system on the East Coast through the annoyingly unreliable telephone lines, which, when they worked, gave us access to a small handful of basic technical information. Over the years, we have repeatedly becoming modernized our equipment and software trading, gradually increasing knowledge and experience. A huge amount of what we realized we went through trial and error, we passed through, after spending thousands of trades in futures markets. We found that it is always easier to learn from your mistakes than from successes. Profitable trading is usually the result of doing many things right, while the losing trade is most often the result of poor performance of only one thing. Since the error is easier to isolate, they teach us valuable but expensive lesson. In this book, we offer our readers the opportunity to learn what they have learned on their own without going through a costly trial and error, passed us. It's pretty much the book recommendations "how to do ...". We first give a step by step explanation of how to build your own personalized trading system. Then give detailed instructions and examples of how basic computer technical studies can be incorporated into your trading plan. We also explain how to use a computer to test your trading system to identify its strengths and weaknesses. Finally, we will offer you some examples of typical day trading systems and strategies that have been introduced by traders who use computers. Along the way, we point to a lot of mistakes and dangers that can be avoided. Many of these errors were painful and expensive lessons, and there is no need to repeat them to our readers. We hope that you will benefit from the practical advice and valuable advice offered in this book. The set of graphs and tables with real markets will be particularly useful in the study of how to analyze the indicators on your monitor. Careful study and application of the material in this book will allow you to make best use of your computer and help you improve as a trader. Charles LeBeau and David W. Lucas
Acknowledgements It would be impossible to mention all the people who directly or indirectly have contributed to our knowledge of the trade in the futures market. It is no exaggeration to say that each of us had to deal with thousands of traders n learn something about each of them. A partial list of those who helped us in the study of futures trading only in the aspect of writing the book, includes (without a specific order) Richard Tyulza, Grayson Whitehurst, John Gilmore, Fred Miller, Ed Maeder, Perry Kaufman, Henry Heshafta, Pat Melloya, Carol Brookins, John Herrick, Richard Kapsch, Tom Madden, William Davis, Jim Neysona, Walt Bressert, Steve Neeson, Ralph Vines, Roy Reeves, Joe Di Napoli, Phil Zachary, John Bollinger, Newton Zinder, Dennis Drappe-ra, Hal Saunders , Charles Harlow, Frank Pusateri, Bob 0'Konnora, Harry Makoff, Don 3
Powers, Arthur Skleru, Dick Rashshena, Harold Sussman, George Lane, Larry Williams, Terry Young, Humphrey Cheung, Gary Sometimes, Steve Kane, Mike Hallorena, Earl Hadad , David Winter, George Regsdeyla, Steve Noutisa, Spencer Davis, Bruce Babcock, Neil Veyntrauba, John Lane, Jim Sibbeta, Greg Garrotta and Bill Ohama.
Our leaflet "Technical Traders Bulletin" was created with the kind help of people from FutureSource. We continue to use their excellent data and software. We are also indebted to Tim Slater of Computrac Sofware, Chris Cooper of Technical Tools, Bill Cruz of Omega Reseach, Tom D'Angelo of the Pro-Manage Software.
Susan and Robert Charles LeBeau engaged in the creation of charts and data collection for them, while the authors of the manuscript polished, detached from the outside world. They deserve special thanks for their hard work. This book would never have come into being without their help. Special thanks to Alan Leonard (doctor) for medical assistance and moral support to our consultant on software and hardware Larry Grodin, who pulled us out of the countless computer crashes (but which we have at least two strokes in golf for the codec Xvid nerves), Joe Ulloa , who first interested in David's magical world of technical analysis, David Johnston, who was a teacher for more than twenty Chuck ago and who taught us the importance of discipline in the management of the funds. Special thanks are also numerous subscribers "Technical Traders Bulletin" for their support, suggestions and trading ideas. Most importantly, we would like to express our gratitude to our timeless families for their loyalty and patience. It's pretty much their book.
Content Foreword 5 6 authors 8 votes Table of Contents 9
Introduction 17 The use of personal computers 17 Building your own system 18 Trading - this is not easy 18 Finding the Right Tools 19 Test before you trade 19 Sharing ideas for day trading 20 We do not know all 20 The real purpose - to make money 22 This is not a book for beginners 23 Caution the reader 23
4
Chapter 1: Building a system of 25 Introduction 25
Why do I need to build a system? 25 The fee for the benefits of 26 Identify the problem, then solve it 27
Problem 1: The definition of suitable markets for trade 28
Liquidity - the key 28 Avoid emerging markets 29 Liquidity should be monitored 30 Problem 2: Determination of the trend 31 Tools trend detection 31 Let this be just 34 Problem 3: Setting the time of occurrence of 35 At the ready, aim, fire 35 Pick indicator 36 Patience is rewarded 36 Problem 4: Set the stop loss 37 Nearest stop loss compared with a distant 37
Ideal stop 38
Next stop 39 Problem 5: Assignment of outputs 41 Messages from outer space 41 A short review of 42 Popular exit strategy 43 The compromise Exit Strategy 44 Income from random entry 46 Issue 6: Setting the time of re-entry 47 "Safe" re-entry 47 The use of oscillators 48 Why not stay in position? 48 Coordinate exits and re-entry 50 Issue 7: Monitoring the system 52 The lower boundary of 52 Historical performance tests on 53 A short review of 56 Suggested Reading 57
Chapter 2: Technical Studies 59 Introduction 59
Types of indicators 59 Learn to use the indicators 60 The reduction of mathematics to a minimum 61 The exchange of ideas 61
Directional Movement Indicator DMI (DMI - Directional Movement Indicator) and Average Directional Movement Index (ADX - Average Directional Movement Index) 62
The concept of DMI 63
The calculation of directional movement (M - Directional Movement) 63 64 Calculation ADX Testing the performance of DMI 67 Using ADX 68 Growing ADX 71
5
A falling ADX 74 Problems ADX: Spikes 75 Problems ADX: The delay of 76 Day Trading with ADX 77 Bands (Bands), Envelopes (Envelopes) and channels (Channels) 78 Section 1: Trade with envelopes 80 Bollinger Bands (Bollinger Bands) 81
Trade rules envelope 82 "Optimal" percentage for the bands 85 Trade within the envelope 86 Section 2: Trading breakouts channel 88 Selection of the time value of 89 Reduced risk by introducing a neutral zone 91 Breakthrough of the channel as confirmation of 91 The Commodity Channel Index (CCI - Commodity Channel Index) 92 Overview of the main theories Lambert 92 Some positive results of testing 95 Using as an indicator CCI 96 long-term trend Using the 99 day CCI Several observations 99 Avoid jerking 100 The divergence (Divergence) 101 Divergence between technical Research and Markets 102 Trending markets, compared with 104 non-trending Major trading rules 104 Serial divergence 105 Divergence related markets 106 Model 110 as
Moment (Momentum) and rate of change (Rate of Change) 111 Rate of change (ROC Rate of Change) 114 Torque signal - trend following 114 Torque signal - following the trend against 116 Long-term trade with 118 points Trade with divergence time 120 The use of the time of other indicators 120 Moving averages (Moving Averages) 122 Simple Moving Average 122 Weighted averages 125 Exponential Moving Average 125 Systems of the same moving average 128 Dual moving averages 130 Triple moving averages 131 Four moving averages 132 Displaced Moving Averages (DMA - Displaced Moving Averages) 134 Finding the filter 139 What medium to use? 141 How do I get the system moving averages?
141
6
Trading method convergence - divergence of moving averages (MACD - Moving Average ConvergenceDivergence) 143 A brief overview of the basics 143 Trading on the MACD crosses 145 Using levels repurchase / resale 146 MACD trend line 147 Look for divergence 148 Combining signals 149 Trade with the trend 150 Parabolic SAR (Parabolic) 152 The origin of the parabolic system 152 Acceleration change 154 Wilder about the values of AF 156 Trading on the parabolic system 156 Directional parabolic system Kaufman 156 Another Parabolic Trading System 157 Percentage of R (Percent R) 160 Larry Williams Percent R (Larry Williams'% R)? 160 General terms of rule 160 Fixing income 162 Divergence 164 Tic-tac-toe (Point and Figure) 165 A short overview of the basics 165 Determination of cell size and the reversal of elements 166 Reading between the lines of the trend 167 "Calculate" your income 170 Smoothing false breakouts 170 A brief review of 170 Relative Strength Index (RSI - Relative Strength Index) 173
Spurious oscillations (Failure Swings) 174 RSI divergence model 175 175 weekly charts
Daily charts 175 Filter entries RSI 176 Re-entry to the RSI 179 Fixing the revenue with RSI 179 The slow stochastic oscillators (Slow Stochastics) 182 184 Time Periods When to use stochastic oscillators 184 Divergence 186 The left and right crossing 188 Knees and shoulders 188 The hooks and loops - Warning Model 188 Bearish and bullish setup 190
Fixing income 190 Volatility 192 Measuring volatility 192 How the volatility of 194 Comments and variations 195 The disadvantages of systems based on the volatility of 196 Recommendation 197 Quick exit 199
Volume and open interest (Volume and Open Interest) 200 Trading with a volume of 200 7
Open interest 202 The interaction volume and open interest 204 Studies of volume and open interest 206 Suggested Reading 208 Chapter 3 System Testing 213 Basics 213
Why should I test? 213 Software 214 The elements of the trading system 215 Expect the worst 215 Dangers optimization 216 What's really wrong? 217 Optimize or not to optimize 217 How to avoid the adjustment to a curve 218 Selection of the test period, 219 Selection of data for testing 222 Slippage and commission 222 Types of testing 224
Simple optimization 224 The cumulative anticipatory testing 224 A simple anticipatory testing 224 Measuring Performance 225 Sharpe ratio (The Sharpe Ratio) 225 The ratio of Stirling (The Sterling Ratio) 226 The ratio of Kalmar (The Calmar Ratio) 226 The geometric mean of 226 Testing for specific results 227 Comprehensive income (Net Profit) 227
Number of trades on a test set (Number of Trades in theTest Sample) 228 The highest winning and losing the largest trade (Largest Winning and Largest Losing Trade) 228 The maximum number of consecutive wins and about the payoff (Maximum Consecutive Winners and Losers) 228 Losses from peak to trough (Peak-to-Valley Drawdown) 229 Winning percentage (Percent Winners) 230 The ratio of average win to average loss (Ratio of Average Win to Average Loss) 230 Total return and the maximum loss (Total Return andMaximum Drawdown) 230 Volatility and the likelihood of failure (Volatility and Probability of Ruin) '231 Testing of entries, exits and stops 233 Testing of 233 entries Testing Methodology 234 entries Analysis of test results 235 Testing Procedures 236 Moving average 236 Channel Breakout 237
Crossing the boundaries of the stochastic oscillator with 238 The jump in the stochastic oscillator 238 Relative Strength Index (RSI - Relative Strength Index) 239 The Commodity Channel Index (CCI - Commodity Channel Index) 239 Moment (Momentum) 240
8
Volatility (Volatility) 240 Random entry 241 The importance of the outputs 242 Testing outputs 243 Testing Methodology outputs 243 The reference system 245 Parabolic Stop and Reverse 245 Support / Resistance 245 Tracking RSI (Relative Strength Index) 246 Key reversal 246
Stop watching 246 Volatility 248 The slow stochastic oscillator 248 Revenue targets 248 Random outputs 248 Conclusions 249 Testing stops 250 Methodology 250 Stop the initial risk 251
Dollar Stop 251 Support / Resistance 251 Unprofitable exit 252 Breakeven stop 252 Stop watching 252 Dollar tracking of closing 252 Dollar tracking the peak or trough 253 Conclusions 253
Create a simple trading system 256 Objectives of the trading system 256 Account size 257 ortfel 257 The software and the data 257 Adjustment to the curve and optimization 258 Risk control 258 Technical Studies - entry 259 Technical Studies - logical outputs 260 The first test 260 The next step 261 ADX as a filter 262 Further testing 263 Conclusion 266 Suggested Reading 267 Chapter 4: Trade within 269 days Introduction 269
The fee for attendance in business 269 Limited chances 270 Selection of markets to trade during the day Consider spreads 271 Maximizing revenues 272 Our rebuttal 272
271
9
Method envelope 5-25 273 The system "Hi MOM" (High torque) 274 Intermarket divergence, three-dimensional technique Ohama 276 Hooks% to 278 Kein One-minute charts with stochastic oscillator 280
Fulcrum 280 Gaps prices discoveries 283
RSI divergence 285 The system of "Knife" Sibbeta 287
Divergence of the stochastic oscillator 289 Key reversal and stochastic oscillators 290
Appendix 292
Introduction It is better to dare mighty things, to achieve glorious triumphs, even though interspersed with failure than to rank with the weak in spirit, who can not enjoy any of the soul, nor suffer much, because the live in the gray twilight, where no wins neither victory nor defeat. " Theodore Roosevelt, 1899
The use of personal computers The purpose of this book - to help the reader in the use of the personal computer and analytical software to make trading decisions in the futures markets. A personal computer is not a prerequisite for trading futures, but can be very helpful, and most sophisticated traders consider the computer as an indispensable tool. Transferring data from futures exchanges by high-frequency signals via satellites fast, accurate, and relatively inexpensive. Software and hardware are inexpensive and becomes every year more and cheaper, faster and easier to use. Computers can be a blessed time keepers when used correctly and destructive eaters time if misused. They allow us to quickly save and restore an almost infinite amount of data, and these data are considered from different points of view. We can study the table that represents the long history of the price of one of the hundreds of futures markets, and then by clicking on the button to isolate the detailed step by step information about transactions that have occurred in the last few seconds of trading in real time. The computer allows us to obtain any set of price data and manipulate it countless ways. We can flatten it, accelerate, enhance, compress, distort, move, stack, store and delete. Our possibilities are virtually endless, and this together with new opportunities lies the problem, what exactly are we looking for, and how do we know when we find? We hope to help you organize your thinking so as to allow the use of your personal computer to focus on the essence of things.
10
Building your own system In the first chapter of the book "Building a system of" we will offer a minimum set for a complete trading system. To you to decide exactly what you want in your personal trading plan. We just try to insure that you do not miss any underlying components. Business plans are the same as at home, it is best to design them to suit your personal preferences. Correction of errors in the basics of how to design homes and commercial plans can be very expensive in the future. Our dominant philosophy in publishing the leaflet "Technical Traders Bulletin" is that a speculator must be carefully prepared plan to trade the markets, such that it is fully consistent personality traits trader and his willingness to take risks. The significance of this is difficult to overestimate. Average customer trading system created for someone else, soon gives up the system on which he spent the money, even if that successfully sells. The reason for this lies in the fact that the system was designed and built by someone who had no idea about the personality and preferences of the person who will be using it for real trading. The best trading systems are always created by the users themselves. Before you start, be honest with yourself. Only you know what you need to do to make you feel comfortable and your work has been the most productive. Ask yourself the following questions: • I prefer to follow every fluctuation of the market, or should I trade on the long run? • How long do I want to spend every day for the development of tomorrow's strategy? • How much capital can I afford to lose in the event of poor trading decisions? • Will I be able to withstand the stress of trading short period of time, or I would be more comfortable to step back and look at the markets from a distance? • If I want to trade on short time intervals, whether the other will gain my work at this time, or the conflict can be devastating for my trading?
Trading - this is not easy
It is also very important to decide how much effort you want to devote to the pursuit of profit. Do not make mistakes, and create and trade on a profitable system is hard work, which also can lead to emotional stress. Our observation suggests that many people who have spent their lives in order to achieve success in their chosen careers, approach to futures trading as a simple way to purchase additional income. The published statistics show that the vast majority of people fail. We need you to say with confidence that you can make money even solid money in futures, but do not expect it to be easy. The most successful traders have devoted a lot of time and market forces. Do not expect to succeed with less.
Finding the right tools In the chapter of the book "Technical Studies", we describe a variety of indicators that can be created with many popular software packages. We'll show you how and where in the trading system should be used each of the plurality of indicators and technical tools. Obviously, if we were building a house, using a set of carpentry tools, we would need to know what each of them. If we tried to nail a saw, we would quickly concluded that the saw 11
- a useless tool. Similarly, if we try to apply the stochastic oscillator as a signal to enter or exit the market is in a state of a strong trend, we also realize that it was a useless tool. The chapter devoted to technical research, describes the basic tools of computer analysis and explains the types of markets in which each of them would be the most effective. One warning: we tried to write a description, and discuss in detail a great number of technical indicators, but we have neither the space nor the competence sufficient to cover all the available technical tools.
Test before you sell The personal computer is also a revolution in the field of creating and testing trading systems. Back in the mid-1980s, it was necessary to be an experienced programmer to test your heart dear conclusions about you developed technical research or trading system. The only alternative to self-programming was to hire someone who would do it for you, that is capacious for time consuming and expensive procedure. There is no need to be a programmer, but some traders still prefer to write their own software. The new readyto-use software displays most of the work required of the field testing, allowing the trader to test the mean endless mutation of virtually all trading strategies that one can imagine. The chapter "Testing System", which describes the basic procedures for testing, will tell you what can and can not do the testing and explain why, contrary to the belief of some traders, testing is not always a panacea.
Sharing ideas on day trading The last chapter of "Day Trading" is intended for a large number of traders who prefer an emotionally rich world trade over short time intervals. At that time, as the basic principles have been successfully used by traders in futures unchanged at different time intervals, day trading birth to a whole new group policies, and they are often quite different from those used for long periods of time. Most covered in this chapter techniques have been sent to our subscribers "Technical Traders Bulletin". Many day trading systems are relatively subjective, leaving much room for interpretation. We have tried to give a more specific description possible. Since we really do not sell the majority of these systems, we make no representations as to their effectiveness. However, we can point them in one aspect common to strategies for long time intervals: found a direct relationship between the level of experience of the trader and the complexity of the trading system. Contrary to what you might expect: the more experienced trader, the simpler system.
We do not know everything As we mentioned, this book is not just limited to their volume also contains the natural limitations of competence and experience of the authors. Our experience is limited for many obvious reasons, and a few not so obvious. Over many years of trading we operated the basic premise, which consists in the fact that futures prices and schedules these prices do not have a neat structure. We could not see any fundamental principle, which would be specified direction of the market. This philosophy is the general disorder of markets has led us to restrict our analysis of search trends or the lack thereof, and the use of appropriate indicators and strategies which should allow for the possibility that the trend may begin or end at any time. We do not believe any that certain future prices can be 12
accurately predicted, nor in what can accurately predict how a certain turning points whatsoever method. What we really believe in it so that significant revenue can be obtained by immediate recognition of trends in the process of their movement and quick to take advantage of the turning points when they appear. Technical analysis, as we understand it, is in the methods of detecting and measuring the strength of the trend, and as quickly as possible the moment of discovery, when the trend may be changing. Our goal in technical analysis was and is closely monitoring the current prices and not predicting any specific future prices. Our understanding of the nature of the markets has saved us from gaining experience in cycles, waves, astrology, relationships Fibonacci, Gann angles, and many other methods, which suggest that the market is an inherent order. We have seen many examples where traders were making money on the assumption that there is a certain market order. We do not argue with them and will not discuss the fact that many of these traders were very lucky. But we are inclined to attribute their success to a good form of money management and disciplined risk control, rather than to the validity of their theories of time or prediction methods, One drawback of the computer is such that it allows complete analysis of past data which can be found in almost any number of repetitive patterns and observation. The computer gives us the ability to produce such a comprehensive analysis on any set of numbers that we can now find the model cycles, waves, and other supposedly recurring relations not only in the prices of futures, but also on sets of random numbers. We have never seen examples proving that this is something represents, but coincidences are bound to arise when a sufficient number of variables used in massive amounts of data. These cases do not prove the existence of any real cause-and-effect relationship. If in fact there was a certain scheme, the underlying cost structures, the discovery and application of such knowledge would be quickly destroyed all the futures market space. In addition, if the market is in some way ordered and pre-determined prices of some unknown controlling force, the trader who will crack the code, or determine this law, will never get the loss-making trade. If someone "knew" what will happen in the future, no one else would not trade with him. We assume that there is some disorder and there are some trends. There are also some periods of serial correlation. We're not trying to add their views to the debate going on this controversial topic, but we want to explain the absence of many very popular and perhaps working technical theories in our presentation of computerized trading techniques. We decided to limit our work by the technical tools that we really used (except for day trading strategies) and to link practice experience with the observations that we have picked up, studying commerce with the use of indicators. In the real application of our technical approach in slightly different from the methods that presuppose the existence of predictable patterns in the behavior of the market. The main difference lies in our focus on measuring what is happening, rather than what is likely to happen. We are always alert to that lucky traders who can draw income from the waves, cycles, astrology and other assumptions predictable patterns were highly skilled and disciplined enough to wait for the actual price movement will confirm that their assumptions were correct. If their success depended solely on the preliminary conclusions or predictions and waiting for acceptance by the market was not mandatory, it would be quite expensive. We also noticed that the inputs and outputs of these predictors are often surprisingly similar to our own. The main difference lies in giving preference dominant cycle or Gann angle, which allows these traders to learn something in advance, while the analysis is the result of a careful study of price action without the use of predictions. 13
As for accidentally hitting the bull's-eye and other direct astrology predictions, anyone who makes enough predictions, is doomed to be right once in awhile. In fact, the published records of these predictors are sometimes impressive, but somehow disappear from there all the predictions that could not be realized. If their underlying assumptions were correct, they would have never been wrong. We have seen the adoption of the planetary influence on futures traders, but how can the influence of the planets control the futures tsynok when we know that there should always be the same number of buyers and sellers? What planet Mars selectively decides that it will affect the trader A, and not on the Trader? Should not the position of the planets affect all traders so that we would all buyers and no one seller?
The real purpose - to make money Futures trading is often described as a game, but do not forget - winning in this game is measured in dollars and cents. A trader who correctly identified and numbered on the wave Elliott wave pattern or correctly detected the presence of the last cycle, did not win the game. He did not win anything at all. This "luck" in the balance of the rooms on the last waves never bring income. Obviously, there is no right or wrong approach to futures trading. Fortunately for us, the traders, to be right is not an end in itself. Indeed, very few really well-known in our business, and even less can be proven. We should be the first to admit that what is stated here, plus a lot of what is to follow, is more of a personal opinion and conclusions than facts. This book is devoted to finding the best ways to analyze what happened in the past, and how to properly keep track of what is happening in the present. We believe that technical analysis is primarily concerned with accurate and detailed study of this, which will allow us to earn income in the future. If you are looking for this method of predicting the future, you will not find it, here or anywhere else.
This is not a book for beginners For those of you for whom this is a new subject, here are some guidelines on the philosophy and terminology. We try not to sink to the jargon, but some failures are inevitable, and we want to be sure that everyone fully understands what he is reading. But this book is not meant to be a first for new traders. Now many good books available that are better suited for inexperienced traders. We are not going to stop and explain basic terminology, how long and short positions, spread, twitching, the trend and the like. We are writing this book for more experienced traders and do not want to spend their precious time at the definitions of the main trading dictionary. Also, to save time and space in this book, we basically give examples that relate only to the buying side of the market. We do not stop every time to explain that the selling party - it's the same thing, only in reverse. As long as we are not directly mention the opposite, for our readers is safe to conclude that a sell signal - it is a signal back to the purchase.
Caution the reader We have tried to be as accurate as possible, but there are always one or two errors that creep up here and there. We're not talking about typographical error, but the error in the technique, procedure, or even logic. Futures trading - it is not science itself is not and never will be. Many of the things that we thought we know, a few years ago were not able 14
to withstand the relentless computer-based testing, and we had to change our thinking and strategy. Much of what we refer to in this book were carefully tested, but much was not. We continue to learn and we reserve the right to change their minds and attitudes in the future. Someday you will find that good traders understand - what is wrong now, then it may be correct. This book - the best thing we did, because we never wrote anything to just sell books and win a popularity contest. In fact, we are probably guilty before a few friends, questioning their trading practices and beliefs. The difference of opinion - that's what the market is doing. What we have written - this is what we believe in, and we sincerely hope that this book will help traders to improve their methods and results.
Chapter 1
Construction of the system
Introduction Why do I need to build a system? Futures trading systems can be successful only if it is followed with religious fanaticism, and in order to do this, you need to feel comfortable and be confident in your system. Without confidence, no trade will take place uniformly. The process of building and painstaking testing your own trading system will help to achieve the necessary level of confidence before going to happen the first real trade. Create your own trading system not only helps to develop the elements important for credibility, but also allows you to customize it to implement your own personal preferences. For best results, each system must be designed to meet the temperament and preferences of the trader who will apply it. A trader who is developing its own trading method should be familiar with the pace of trade as well as the nature of the risk inherent in the procedures that need to take place. In this case, the trader is more likely to follow the system in a critical situation. If two traders use the same system and it belongs to a class of systems that face frequent losses, a trader can take the losses and continue trading while the other may lose confidence and get out of the game. One trader will be the winner, the other - to lose. One trader will claim that it was a great system, and the other is that the application of this system was a catastrophic mistake, and they both are right in their assessments. As we have seen, due to the fact that futures trading is a very personal thing, the best system for one trader may be totally unacceptable to the other. There are countless elements of the trading system, which come into play personal preference. For example, some traders feel insecure, a strong trend for buying or selling after a long decline. These traders may prefer to design a system that buys only after the fall, and selling after consolidation. Many traders who hate twitching resulting from the closest stop to prefer a system with free stops. Some traders will require high activity of their system, because they are eager to direct action in the market, while others will prefer very quietly, carefully enter the market only when the opportunity seems to be almost perfect. Professional sales consultants are faced with an even more difficult task, because we have to develop a system that should satisfy not only their personal preferences, but also the expected preferences of their customers. Professional systems are becoming more conservative, that is, combining a modest income with minimal losses. Corporate and 15
private investors are probably best served by such systems is because it usually neurotic skeptics who are more likely to take back their money or change the consultants at the first negative signals. Private traders, which have the advantage of developing their own trading systems, it can leave a cautious approach and look for a way to make money with all possible speed, paying little attention to the severe psychological and financial blows inflicted setbacks. Through personal development and testing of their systems, they will be better prepared to face the inevitable losing periods that are less confident traders will lead to the termination of such an aggressive approach. If you happen to be one of those aggressive traders, just make sure that your system is designed exactly the same answers as your tolerance for losses, and a passion for winning.
Fee for benefits Construction trading system consists of interconnected successive solutions, each of which has its advantages, but also requires a fee for them. The system designer must develop a reasonable set of alternative solutions, and carefully weigh the pros and cons. The secret to ultimate success will depend on your ability to accurately select the procedure that will give the desired results at minimum cost, while being coordinated with your personal preferences regarding the characteristics of the system. For example, to avoid the effect of "twitching" can be achieved by expanding your stops. However, the price for this is a disadvantage that you will suffer great losses. An intelligent designer of the system will attempt to overcome this problem by selecting a solution that will avoid most of the "twitching" with a minimal increase losses. When it comes to the development of systems, there are no gifts. Assume that there is a fee or a lack of benefits for everyone, and spend analysis to determine the exact size of the board, and then decide whether it is in order. Identify the problem, then solve it. To be successful, as the creators of systems, we need to master the two components of their construction: precise definition of the problems at each stage and the selection of the most suitable solutions of the many available. Since there are many solutions to each problem, we have to choose the option that is most comfortable in terms of personal preferences, and goals of the system of payment for the decision. There is a very some workable solutions that require for itself a reasonable price. As we have emphasized, first of all it is necessary to define the problem. The following is a list of the major challenges that must be faced trading system, and some possible solutions. We hope that this simplified approach is a step by step system developer through the creation of a workable and profitable trading system, which you can follow and trust.
Problem 1: The definition of suitable markets for trade More than a hundred different futures contracts offered only on the U.S. Commodity Exchange Act, and the new contracts appear almost every month. In our opinion, to date, we have too many markets to keep track of them all, and how to effectively sell to them. Even taking into account the need of a PC will need too much time and trading capital to trade for 30 or 40 markets, so that the attempt to treat more than 100 markets and is not discussed. The first step in building a practical trading system should be a study of futures markets for a reasonable number of contracts. Even before we start to worry about trends, methods of entry and many other elements of our trading system, we need to set some minimum trade standards so as to separate the markets for which we want to follow. We 16
do not want to spend our valuable time and resources by analyzing the markets in which we would not trade even if a signal is received from our best indicators.
Liquidity - the key The main factor taken into account when selecting markets, should be liquidity. Can we enter and exit the market quickly and efficiently? We do not want to condemn himself for the expensive filling orders or to monitor how prices move in a wide range at low volume, especially if the amount represents the our transaction. We want to be sure that we are dealing with markets, where the spread between the prices of supply and demand is reasonable for the number of contracts that we are going to trade. We want to avoid markets with unusual sequences of price and where large gaps are the rule rather than the exception. The best measure of liquidity - this volume and open interest. We have proposed to set minimum standards, such as the daily volume is 5,000 contracts in the contract month, when we want to trade, plus 20,000 total open interest (for all contract months). If you are a trader, who prefers large volumes, or professional advisor, you will, no doubt, want to raise these minimums even higher. We propose to use volume, along with an open interest because it is often a large open interest may arise as a result of lengthy long positions and hedging. Because of its open interest of inactivity this adds a little to the daily liquidity (see Fig. 1-1). Computer data shown in this table, make it easy to keep track of the volume and open interest. Now we have still another problem to be solved before entering. We need to protect themselves from large losses by setting stop losses. The judgment of hindsight can sometimes be a useful tool in determining the stopping points. John Sweeney, a former editor of "The technical analysis of stocks and commodities" (John Sweeney, Technical Analysis of Stocks and Commodities), proposed to measure the maximum negative price action last winning auction to determine how far it was necessary to remove a breakpoint in order to keep all winning trades and losing away. This method has its advantages, if you carefully compare the worst results, which should include an analysis of the impact of freer all the stops on the losing trades. Perhaps you would be better to put a stop close and miss some profitable trades. Also remember that this method - the judgment retroactively, which does not allow changes in the resistance, which is likely to arise in the future. It is also important to keep your stop outside the randomness of the future, not the one that was in the past. However, if you believe that the future is likely to closely reproduce the past, Sweeney's approach makes sense and maybe getting better all the methods that we have reviewed.
17
In addition to liquidity, we also need to consider other important properties of the contract, such as historical volatility and availability of accurate fundamental and technical data. Markets are having in their stories trends broad scope would be preferable quiet, narrow markets rarely get the chance to even substantial income subject to the availability of accurate data we have. Free access to accurate and timely technical and fundamental data is essential, even when we are not using the data in our own analysis. The fact that such data are available, will attract other traders and increase market liquidity.
Avoid emerging markets We would also warned traders from entering into new markets. In spite of sufficient volume and open interest, make sure and wait for the new market will be held every two or three cycles of the supply before adding a new product to your portfolio possible. You never know what you are trading and the real price of this as long as do not exactly understand what can be delivered on futures contracts. Some of you may remember the futures contracts for diamonds on the long-defunct West Coast Commodity Exchange. Significant supply problems related to the quality of diamonds, thrown out of the diamond business traders in futures and caused the collapse of the exchange itself. We do not recommend you to become trailblazers or guinea pigs for new contracts that regularly appear with great fanfare only to quietly fade away after a few months. Have participants and hedgers do the dirty work and forced to contract. They have a legitimate interest in the passage of these contracts, as well as the necessary experience and capital to trade them in the incubation period. Liquidity standards 5000 and 20000 of the daily volume of open interest will help you avoid almost all of these suspicious markets.
Liquidity should be monitored The application of these standards, the minimum daily volume and open interest will allow us to filter out of 100 something around 20 markets, which is acceptable for daily tracking number. Remember: When changing the fundamental and technical situation, price, volume and open interest will be set accordingly. The volume will tend to grow in emerging markets, while the trend-less and falling markets will lose volume. Do not worry about missing any significant price movement. Notable price movements quickly attract volume, so that you will enter early enough, if it starts to grow a big trend. This means that the volume and open interest should be checked regularly and we encourage you to 18
review the liquidity of markets at least once a month. You will find that your set of markets will change over time depending on how the volume and open interest are increasing or decreasing relative to your minimum standards. And while these principles volume and open interest will save you from many quiet Detrend markets, thereby saving you money and allowing you to focus on your technical skills with a true market potential for higher earnings. If you regularly track the volume and open interest, you'll find that expiring contracts and other seasonal factors can cause drastic changes in the open interest. Typically, these factors do not cause the product to cross your level of open interest. Daily volume is more likely to fluctuate at the intersection of your level, so it is better to average volume over the past 10 days or more before you decide to add or delete from your market portfolio. Now, when you appeared manageable number of markets to work, you can move on to the next problem - determining trends in these markets. In the meantime, make sure your program is well set up to track the volume and open interest. If you do not get the daily volume and open interest on your computer, then periodically check it for financial publications.
Problem 2: Determine the trend Whenever used word "trend", it must be determined relative to the time period. In this case, we are looking for trends that should last three to four months or longer, so we use the basic technical research could be based on the daily, weekly or even monthly time periods. Remember that at this point we do not use the study to enter the market, and determine its direction. Once we have established the direction of the market, we will start working on the definition of prime time entrance. Now we are concerned only with one simple but very important question: whether the markets are moving upstream, downstream or sideways trend. (Yes, we are considering the lateral movement as a trend.) There are many simple technical indicators, and even more fundamental, such as the analysis of supply / demand, which can help us in determining the direction of the trend. However, fundamental analysis, along with many other topics of discussion is beyond the boundaries of this book.
Tools trend detection Let's look at a few common technical studies that may be used to solve the problem of determining the direction of the market. We always got good results with main applications of simple moving averages. In particular, we used the 3-day moving average in conjunction with the 12-day moving average and 9-day moving average in conjunction with the 18-day moving average. For example, if the 3-day moving average is above the 12-day moving average, we would say that the market is in an uptrend. Conversely, if the 3-day moving average is below the 12-day moving average, we would have concluded that the market is in a downtrend. As you can see, it is possible to determine the direction of only up or down, but you can not define a sideways trend, using only one or two moving averages (see Figure 1-2 on the next page). Possible, however, to construct a tracking sideways, if we use more than two moving averages. For example, using the popular combination of 4 -, 9 - and 18-day moving average: when 9 above 18 but below 9 4, we can conclude that the market has been in a sideways trend. If the 9-day moving average is below the 18-day moving average, but the 4-day moving average went above the 9-day moving average, we can conclude that the trend was also side. The upward trend can be determined only when the 4 was above 9, 19
and 9 was higher than 18. The downtrend is defined as 4 below 9, and 9 below 18 (see Figure 1-3 on the next page).
20
In addition to the moving average of all types, classic and quite acceptable direction indicators are: trend line, linear regression, the SAR, research tic-tac-toe, as well as indices of direction. In much the same as moving averages, many of these indicators do not define the lateral markets to do this requires a combination thereof. One of the most effective ways to determine the lateral trend requires the use of multiple indicators, such as indicators rassog-lasovyvayutsya, the market is considered to be non-trend or side. We can always find a sideways trend, when we use two indicators that can not fit in, or to confirm the signals in one direction (see Figure 1-4).
Let this be a Some traders are developing special trend indicators for each market in which trade, trying to create the perfect system for commodity-based goods. For example, they may choose different combinations of the moving averages for each market. In our opinion, the selection of technical studies that determine the trend reduces to a simple and convenient. From this point of view, our study selected should be easy to understand, without too much subjectivity and ambiguity in interpretation. Instead of trying to develop a very complex technical indicator that would do everything in our system, we chose to split the construction of our system into functional elements, and then picked up a simple but effective technical means for each function. Now let's assume that we are using two trend indicator, and when they are consistent in direction, then we find a trend in this direction. And when they do not agree, we say that the trend sideways. Once we have established the direction of each market, our system is a trend following works depending on the direction of the trend. If the direction is upward, then we want to adopt a strategy of buying only as long as the trend will not change. If the trend is down, 21
we will use the strategy to sales. On the market side, we can not use trend-following strategies. We have a choice: either to stay away from the market side, or use a counter trend strategy that buys on dips and sell on the climbs. We do not recommend reversal strategies that can not determine the market side and always move in the markets from short to long positions and vice versa. Such reversal of the system tend to have a constant twitching in sideways markets and they have no hope of success as long as the markets do not enter into long-lasting trend. Now that we've sorted out our 20 markets in the direction we are ready to move on to the next stage - the definition of the time of inclusion.
Problem 3: Setting the time of entry Most traders can not understand the complexities of building a system and just trying to find the Holy Grail indicator that will do everything. Such traders are willing to believe that one perfect indicator will identify all the trends, set the time of entering and even signal the exits and turns. Would not it be great if the markets and were really so simple? Popular practice to place reliance on a single indicator, which produces a series of signals, inevitably doomed to failure because, when the nature of any of the elements of the system changes, the system will fail. We believe that it would be more prudent to isolate and test each of the problems, and then carefully select appropriate indicators for each specific task. By implementing the approach of multiple tasks / solutions, we expect to develop a flexible dynamic system that will survive in the real world of ever-changing market conditions.
At the ready, aim / fire Artful entrance to the position can be compared with shooting a gun at a target. First we have to aim (finding direction). Then the cock (ready), and then gently pull the trigger (enter the market). At any moment markets consist of at least three trends. The first trend long (weeks or months), which we use to determine the focus of the market. The second trend is the average length (last few days), we now need to determine, using more sensitive indicators. Latest trend - the most short-term price movement (yesterday and today), which we will use for precise input. In the process of solving the problem of determining the trend we have offered a few of the many possible indicators determine the direction of: various moving averages, charts tic-tac-toe, long linear regression trend line and a combination of indicators. Since the long-term direction of the market has been defined - is our next task is to find the mediumterm indicator, which gives a series of signals within the long-term trend. We need a series of signals, because the first intermediate signal occurs before the long-term indicator will allow us to trade in this direction. Note the clear sequence: short signals occur first, then the medium and, finally, long-term signals. By the time we define the long-term trend, the first intermediate and short-term signals already arise. Thus, we want to use the intermediate and short-term signals to be repeated several times within a long-term trend.
Pick indicator A list of useful intermediate indicators may include crossing DMA (Dual Moving Average - double moving averages), channel breakouts, moving averages, parabolic signals, violations of the trend lines, research tic-tac-toe, and any number of methods of pattern recognition. Every trader has his favorite indicator, and any of them may be as good as another. We must remember that we are building a system around a combination 22
of indicators, so that the importance of a separate indicator is normally suppressed in the total system. All you need to do is to find an indicator that you trust and who will give a series of short signals during a long trend. Real input starts market activity, followed by the intermediate signal. Furthermore, there is a choice run mechanism. For example, place an entry point to a new peak or trough traffic stop, or select a point beyond the boundaries of today's peak or trough. You might want to be very cautious and wait for the entrance to the series of peaks and troughs. If you are less patient, you can select the input on the first closing in the right direction. The important thing to remember here, is what is needed to confirm price action signals to the rest of your indicators and allowed the market to start your entry. Most good traders make profits immediately after the start. Our goal is to go in sync with all three trends from the beginning.
Patience is rewarded Testing by the Group of Trading System Research and Development (User Group System Writer Plus), showed that the results of the double moving average crossing would be dramatically improved by expectations of a breakthrough peak of the day on which the intersection occurred. In applying the tactics of expectations of a breakthrough to enter the total return on a theoretical account on the testing period five and a half years has increased by 177 percent (from 22 percent to 62 percent). Revenue for the trade jumped by 275 percent! The researchers tested 56 combinations of moving averages in 10 different markets in the period from January 1984 to June 1989. The research team came to the conclusion that waiting until the market itself will admit them to trading, increased revenue per trade, reduced the required capitalization, lowered the number of trades and significantly raised the total income.
Problem 4: Set the stop loss Any experienced trader will confirm that the orders to stop losses is essential to avoid catastrophic losses. Traders who speculate on markets without interruption losses, are doomed to fail, the only thing that can change, so this is the time in which they will go bankrupt. The stop loss is similar to the contributions to the insurance policy and should be viewed as a necessary cost of business.
Nearest stop loss compared to a distant Stop losses are generally divided into two categories: near and far. Ideal stop should be one of those that are located far enough to barely extend beyond the boundaries of random or from a technical point of view, meaningless price movements, and at the same time close enough for a comfortable control risks when trading. We found that these common options are mutually exclusive, that we are forced to reduce the search to stops that are set either very close or very far away. Let's look at the "pros" and "cons" of each of the options. Close stops offer the obvious advantage of low losses at each position and restrained the overall risk on a portfolio of open trades. However, this procedure leads to a weakening of the financial and psychological discomfort of many of the stops on the auction, which could be very profitable if the original position allowed to run its course. We assume that traders can get around this problem by simply defining a convenient method of re-entry, which will bring them to the market in the initial direction just in time to catch 23
most of the remaining potential income. However, as always, is a compromise. This logical step to return to the market will inevitably lead to increased activity, which significantly increases the transaction costs and the costs of slippage. A system using a near halt, face a disadvantage, as expressed in an increased percentage of losing trades, but have the privilege of smaller average losses. A system using distant stop will tend to increase the percentage of winning bidder to the system with close stops. Stopping is not faced with the problem of re-entry and is in control of slippage and transaction costs. This completes the picture significantly increased the average loss on trade and significantly increased overall risk to the portfolio. It seems that we are faced with trying to choose the lesser of two evils, or an acceptable compromise between the two equally unpleasant possibilities.
Ideal stop Appropriate procedure stop loss could be developed, if you try to put a stop just over the border random price spikes. In the event that one of these stops work at a time when the trend still continues to move, we need a method of re-entry, which will bring us back to the trade when short-term trend will resume its movement in the direction of the long-term trend. This procedure seems to be a workable compromise between too close and too far stops. Setting stops outside the randomness should help avoid most of the troubles and losses from frequent twitching. The method of re-entry will help avoid disruptions due to omissions of any significant price movements. It sounds simple enough, but the definition of chance - this is one of the foundations of futures trading, and if we could do it accurately enough, we would not even need to stop - all our trades would be advantageous. We can not quite work out perfect, but acceptable procedure, which will include the basic principle of the job stops, allowing us to avoid most of the problems associated with random fluctuations in prices. As one possible approach would be to use the standard deviation of prices from the moving average, and then place a stop a few steps away from the standard deviations from the moving average. Not all software systems offer such a standard deviation band (now often called "Bollinger Bands" - "Bollinger Bands" on behalf of a technical analyst FNN, who popularized the band as a technical tool. Refer to Figure 1-5). As a practical matter (and perhaps as effective) alternative to the complex approach of the standard deviation, we can use the average daily price range as the minimum distance for setting stops that will help us to avoid most of the small fluctuations in price, resulting in jerking. We can simply set a 5-day or 10-day moving average of peaks or troughs, and then place our initial stop at the minimum distance to be equal to the distance between the moving averages. As long as the market moves favorably, the stop may also coordinated this distance. This technique helps to avoid what we call "random fluctuations throughout the day," because it keeps stopping far enough to avoid the daily fluctuations. In order to stop us, you will need an abnormal fluctuation within a day or a series of hostile daily price changes. Maybe this method is not perfect stop, but it can be very useful in the sense of finding the minimum distance to stop, to avoid unnecessary jerking. Other acceptable methods of setting stops that you may wish to examine are the points on the graph, such as the levels of support
24
and resistance, the peaks and troughs of the last days, parabolic stop and all sorts of envelopes or trend line. Since there are no perfect stops, there is no need to be distracted by overly complex techniques for solving the problem of where to house them. In fact, we have tested many methods of setting the initial stop and found the usual number in the U.S. works just as well as more complex procedures.
The next stop Whichever method you choose, it is important to be consistent and disciplined. For example, consider the results of a trader who started with $ 500 and a stop after five consecutive twitching lost $ 2500 and missed five potentially profitable moves. After that, he decided to use a free stop and lost $ 1,500 on the next trade. Now he experienced the disadvantages of both methods, losing too much money on the last trade, without having the advantage of getting the potential revenue in the first auction. If any of the $ 500's or $ 1500's stops used unchanged in this period, our example would have produced a much better result than the failure, which was caused by the inconsistent approach. You may not, without good cause at the same time to use the closest stop, and in the other - distant. And except for the changes in the stability of the market is probably not there any good reason for a significant change in the stops!
Problem 5: Assignment of outputs As a result of our research, we came to the conclusion that traders spend too much effort trying to find methods to determine the time of input markets. Somehow got the mistaken belief that success depends on the time of inclusion, and that all the rest is a matter of will. Traders are concerned about this so that the search for the perfect entry system has become like a quest for the Holy Grail. Unfortunately, the truth is that the entry is one of the least important part of a complete, well-designed trading system. We argue that the real key to income is in knowing how to get out. We know a lot of traders who make money in spite of their absurd methods of entering, never realizing that their precious strategy I add very little, if any, are added to the impact of their trades.
Messages from Space
25
We once met a trader who assured that receives signals from the input of the mysterious creatures in outer space. He argued that receives these messages with the help of his "interplanetary cell phone" made of Coke bottle sticking out of the neck of a piece of broken radio antenna. This happy (or unhappy) trader actually made money because of the fact that he had a good knack for determining the correct outputs of trading. He could not afford to lose money and suffer other traders sarcastic smile, sitting in the operating room, so quickly closed the money-losing trades. He used to accuse atmospheric phenomena or some cosmic interference that distorted his secret message. When he chanced upon winning the auction, he extended this success as much as possible, and such a way, could brag to their colleagues the right message from outer space, and scoff at their seemingly futile attempts to earn his money, studying charts and fundamental information. He severely criticized the conventional methods of trading and admired by his own success. He was utterly intolerable when found himself on the right side of the market. This successful trader in the habit of cut off their losses and allow income to flow, so he made money. His success has hit trading floor, which made fun of talking bottle of Coca-Cola and meets her miraculous trader. For all his eccentricities, this trader unconsciously followed an excellent exit strategy that allowed him to earn money. However, if you asked him, he would have sworn that his success was defined solely by the input signals, which he received from a bottle of cola. It is not surprising that many of today's popular entry system based on the theories are even less suitable than the message of cola bottles. Stop and think, what's the difference between getting a clear message from a computer connected to a satellite dish and receiving messages from imaginary bottle of cola? If you believe in what you are doing and act according to reports, then you start with a roughly equal position, but a trader who will be the best in the output, will earn more money. Despite the fact that someone else can claim, those who achieve success in futures trading, have a good exit strategy.
A short review Let's take a look at where we are now. The first thing we did was viewed various futures markets and identified those that are called "traded". Then we come to determine the direction of each of the selected markets. We then discussed the problem of determining the time of occurrence and the associated target shooting with a pistol. Then we moved on to find a logical stopping point loss. So now we have an open position and a stop loss. We either make some money, or stop. The losses take care of themselves, if triggered our stop, so our main concern at the moment should be to find ways to maximize revenue. When trade takes place in the direction we expected, we are faced with a choice between getting a fast but steady income and continued trade in the hope of winning big. Most novice traders tend to lose money at a good percentage of profitable trades and a serious loss, while more experienced traders lose money by collecting a lot of small losses. (We confess that we have not turned off the intended path by studying the strategies of losing traders, and will be difficult to avoid the periodic references.) For a successful application of the system with low income should give a high percentage of winning trades, but we are tempted every day the old adage: "Better a bird in the hand is worth two in the bush." The rapid acquisition of small gains rather than let them slip away and turn into a loss, helps to improve the largely averages. On the other hand, we have many times there was a possibility to allow income to flow further. How do traders in the face of such a contradictory situation?
26
It seems to us, the trader must have a minimum acceptable level of income that is directly related to the amount of money at risk when the stop loss is triggered. Most profitable traders whose experience we studied had a mean volume average more clearly wins of failures, typically for two to one. Given the slippage and commissions that must be included in the results, you may need to enter the ratio of income of about three to one for the final ratio of two to one on the set of trading. Those who follow the long-term trends, as you might expect, have the opportunity to receive a good value for the loss of income and the low percentage of winning trades. But despite this, their attitude income is not as high as would be expected, and reaches, at most, to four to one.
Popular exit strategy Let's discuss some of the most popular exit strategy. The method of entering and holding the position for a large income working for long periods if you do not mind the loss of large and painful losing periods that will enervate and very expensive on short trends. These methods are useful for trade only your own well-capitalized and require a huge confidence, experience and discipline. Unfortunately, these conditions and the lack of quality in most of us. The main component of the risk of this method chase for higher income is that a trader will almost inevitably ends up in the middle of the output period of severe losses. Look closely at the open documents of any professional consultant who follows the strategy of long trades, and you will see hundreds of closed accounts. This accounts departed customers, while the main indicators of consultants were quite profitable and impressive. The vast majority of traders can not stand the kind of big profits they missed, and they are psychologically unable to withstand the inevitable loss, regardless of how well they were prepared or educated. Be very careful when using this strategy for yourself, because you will pass the painful test of his trust and belief, which can be terrible. Less popular exit strategy, considered below, this method of aiming out as traders closed, reaching a pre-determined price targets. As we have stated previously, it appears that each strategy has its downside. We see some of the main problems associated with the ability to predict a particular purpose with any degree of accuracy. Anyone can point out some obvious support and resistance levels, which may cause the trend to waver, but with the exception of this general analysis, we doubt that a more accurate aiming is actually reasonably possible. No one knows where the market is. Trader using iron sights exits, gets an advantage in that it does not face the problem of monitoring the loss of large unrealized gains. On the other hand, they will certainly suffer from frustration that many prices have not reached the predicted objectives. The trader will also have to learn to withstand disturbances brought by the observation that has been received less income while you can get more with a little more patience. We consider aiming a workable exit strategy, provided that you have the knack of selection of targets and the ability to not look back at what could have been. You should also be careful to avoid the natural tendency to reduce problems of sequential revenue for higher averages. This seductive way can cause serious problems if you let the relation of average income to average losses too down.
Compromise exit strategy One of the best strategies we know of income - is a compromise that provides the advantage of rapid revenue and leaving at the same time, the possibility for big earnings. You are just using double trade account and receive income from one position to priberezhennoy price target, and the second position allow to be open in the hope of 27
winning big. This method requires a large capital investment as compared to one trading account, but it seems that he has obvious advantages. Fast income under one contract will always give you more freedom in the second, and you can not afford to be very patient. Putting the account of one win, you can give the second position is sufficient time to avoid a premature stop. As we made clear when we were not talking about the benefits of a particular strategy, we also have to carefully consider its negative aspects. For example not far to seek. The apparent lack of a dual strategy is that if you start in the wrong direction, you will be losing two positions instead of one. If you have a very good strategy to get in, and you're convinced that the majority of your trades will start in the right direction, then the dual strategy can be a great choice as an exit strategy. But before you take this strategy, be sure to check the results of your past and see if you can withstand the losses twice as high as those that occurred during the periods when your entry strategy did not fire. For a trader on one trading account, we recommend a method of exit, which gives some room for maneuver in the market (wider stops) as long as it does not become overbought or not to give an unusually big move in your direction. Then narrow the income stops to protect most of the profits, but at the same time be able to generate revenue, and further, if the market continues to move in the right direction. Sometimes we use a six-day RSI Relative Strength Indicator (RSI - relative strength indicator), which tells us when the market is overbought and we should raise the stop. For example, when the indicator of the relative strength rises above 75, and then drops by 10 or more points, we prefer our stops raise the level of the minimum price for the last three days trading and correct them rising market. Often, this procedure allows us to remain in a strong market and throws very close to the top. (See Figure 1-6.)
Another simple but very effective method is to use a stop tracking in all cases, and to remain in the trade as long as the stop is not going to work. Using this method, you will at all times know the exact amount of income that you can lose. Despite the fact that this method of exit may seem simple, he showed himself well on historical testing, and it is difficult to find a better way. The tracking station can be used in conjunction with the 28
description of the method previously repurchase / resale (using RSI) to be able to go out every time closer to the top.
Income from random inputs How good is your current exit strategy? A good way to check the return of your outputs will be joining the hypothetical trades at random, regardless of the direction or time. Then, place your stop out and apply your strategy to generate income. If, after the passage of 30 or more such trades your results (in terms of total revenue) is better than the absence of losses, then your exit strategies should be improved.
Issue 6: Setting the time of re-entry In the previous section we discussed the importance of properly setting the time out. Good outlet features a winning trade from losing, and perhaps this is the single most important element of any system. Unfortunately, in our attempts to get out of the market before the give away too much of our hard-earned income, we can often come out before the trend actually ends. In cases where the trend continues, we need a way of re-entering the market. Strong trends are rare, and are too valuable to lose them, so we have to be sure that we can return to the market, if our previous stop turned out to be premature. The nature of our multiple entries may be quite different from the main entries, because the market is now in the middle of a strong, well shown by the trend, where the fluctuations in the price is much higher than at the beginning of a trend. We're at a point where confidence in the direction of the market, and what we need here - it's a quick reentry in order not to miss the main motion.
"Safe" re-entry One fairly obvious strategy, which certainly brings us back to the market if the trend continues, is to buy at each new peak (or sell to each new trough, if our previous position was short) for general assumption that the market's ability to overcome one of these classic point of resistance is a sign of trend continuation. In recent years, with the acquisition of technical analysis more finesse, a break or depression on the top of the chart rarely leads to the expected wave of purchases or sales, as was the case in the "good old days" when the charts were compiled manually tool of the trade. Creation of new peaks or troughs usually entailed a lot of stops and losses caused to the market more traders, and the trend continued in his new "wave". It is not necessary that it will work that way in the future. Now the price movement is more likely to respond to indicators such as the stochastic oscillator divergences and intersections MACD, which may not be apparent nekompyuterizirovannomu eye. Despite the fact that the purchase of a new peak in the mandatory insures you from missing a lot of traffic, the lack of this approach lies in the fact that buying at the highest price in the field of view is rarely the best way to enter the market. If the market will continue to move as we had hoped, we may find ourselves in a strong interest and loss. Instead of waiting for a new peak before re-entry, it would be better to re-enter once held an unfavorable movement that forced us to come out for the last time.
The use of oscillators 29
Oscillators that determine the overbought resold or given, can work very well in the determination of multiple entries. Working with an example for non-trending market, let's assume that we were stopped at a profitable long position in the correction of the price that was more powerful than we could take. We could now observe the relative strength index (RSI) or stochastic oscillators to produce a signal of the end of this deviation. One of our favorite techniques is to wait until the Stochastic Oscillator falls below a certain level and then turn back. The fall of the stochastic oscillator to any value below 40, followed by a recovery, should initiate a usable re-entry. Usually trade for the purchase of a fall caused by the stochastic below 20 or 30, and then turn. However, since we are on a clear upward tren de, it is unlikely that the stochastic oscillator reaches very low levels. The stronger the trend, the higher the probability of a reversal of the stochastic oscillator. (If the Stochastic Oscillator falls only to 50 or 60, and then turn around, we probably will not be stopped in our original position and we do not have to worry about re-entry.) After the launch of a new trade for the purchase, we can supply our new stop loss below the level of Depression correction, and then raise it to our point of no damages when it reaches a new peak. These trends are dying slowly and with difficulty, so that the probability of getting a good trade for re-entry is quite high, especially if we can enter the fall, without waiting for the next peak.
Why not stay in position? You must be very surprising advice to leave the starting position, if we like the current trend. The advantage quickly manifested when our "correction" is not a temporary correction, but a complete change of the main trend. We will go to the market with higher income and not enter again. It's so close to the ideal output as possible while following the trend. The secret to the success of re-entry is to wait until the end of time correction and start quickly procured, once we understand the direction of the main trend. Waiting for the market to produce a new peak - it's too long wait, but we need to ensure sufficient power, indicating that the correction is indeed over. We're talking about a very delicate moment that requires careful thought along with the availability of sensitive and reliable indicator. As an example of how sensitive can be an indicator of re-entry, we recommend a method of using a very short-term oscillator, such as a three-day Relative Strength Index (RSI) as a starting signal re-entry. (See Figure 1-7.)
30
Typically, a three-day RSI is so often skips it is a bit as an indicator. Since this is a very sensitive indicator, any correction is strong enough to stop us in our original position, lower the three-day relative strength at a very low level. When the RSI turn back the mark of 50, we can conclude that the correction is over. Consequently, we buy the next day, when the market comes from the peak of the day. RSI had raised to a value of 50. RSI technique gives us two signs of trend continuation (meaning 50 and confirmation), and at the same time it is fast enough to get us back on the market well before the new peak. Others counter trend indicators like Stochastics. Percentage of R and the Commodity Channel Index, may also be used in this way. Percentage of R - is a sensitive indicator that will work almost as well as a three-day RSI. The idea is to use one of these indicators for the end of the correction signal. You need to install a more sensitive indicator than usual, because we want to measure is short-term correction, not the actual trend.
Coordinate exits and re-entry Another useful topic related to multiple entries, is that we can expect the best results if repeated occurrences of sync with our exit strategy, so that the price movement, including our way out, we automatically start the selected indicator of re-entering. For example, when we were stopped by a long position, the value of the stochastic oscillator should drop the mark of 40, Percent R should drop to -90 (-100 - The minimum scale of interest R, and 0 - max) and RSI should be less than 50 depending on the what kind of indicator, we plan to use for our re-entry. They will be launched in order to give us the signal of re-entry, when turned up with these lower values. If our indicators are used correctly, we do not risk missing any significant trend continuation. If the trend continues, before our indicators will be used, we risk missing the movement. Remember, the exit strategy - the most important strategy, so use the best possible strategy and then set the sensitivity of re-entering the indicators that are triggered after the exits. Because of our preference sensitive signals re-entry, we will from time to time subject to certain antsy and losses. (When we did not achieve any benefits, there will always be 31
negative.) It is assumed that after only one or two consecutive losses from repeated occurrences or we go back to the trend or our indicator of the main trend will change direction or go to the side. In any case we will not have to endure a long series of twitching with repeated occurrences. Maybe it's just our whim, but we do not always like to watch the leaves most of our unrealized gains. It seems to us that the pursuit relatively quickly fixed income strategy will be suitable for most traders. But be warned in advance: this technique is fast income you may be sorely disappointed if you do not skoordiniruete it with a well-planned re-entry so you do not miss a lot of traffic. In the next section we will give some useful tips on how to keep track of your system, so you can determine when it is in need of repair, and where it should be discarded. Gains and losses are not the only and the best way to measure how well actually running your system. Many errors can be identified before any any serious damage.
Issue 7: Tracking system Are all systems fail from time to time? We would like to know! We know that markets and market conditions change, but probably human nature and human qualities, such as fear, hope and greed, do not change. We can argue about the question: "Are all systems fail?" on either side. It seems the most sensible approach is always to assume and prepare for the worst. Thus, let us assume that in spite of all our hard work and testing, the best possible system, which we can only imagine, would one day fail. (We define failure as the loss of so many either funds or trust, that we can not follow the system.) With the assumption of the possibility of failure looming in the background of our work, our best defense is a failure to recognize the symptoms as early as possible before the moment we will lose money or patience. We need to develop a well thought-out an early warning system so that we could either change the system or abandon it for a sufficient reason.
The lower limit The natural desire of the majority of those who follow the system, just to focus on the lower level of performance. We make money or lose money? The problem with this logical approach to the lower boundary lies in the fact that all systems are periodically lose money, so it is obvious that only the abnormal loss-making period is serious enough to bother us. At this point it may be too late. To avoid this, we need to develop an informative and objective set of predefined criteria unrelated to the lower boundary, which can alert us to potential problems long before any serious damage. Imagine a typical situation where the system has shown a modest income at the usual month of trading. So far, so good. A closer look may reveal that within a month there were 14 trades and only 2 of them were successful. Fortunately for us, the proceeds of two successful trades blocked the costs of 12 unprofitable. This is a good example of how a simple analysis of the lower bound can let us down. In the end, it was a profitable period, so why bother? Should we be concerned about the fact that only two of the 14 recent trades were winners? Maybe yes. Or maybe not. What can we expect in the sense of the winnings to the losers? Are two of the 14 winning trades something we can expect from time to time, or is it a warning that our system may have crept along the seams? We need to be prepared for such a situation and, more importantly, should be able to consider it in perspective. How about the activity in the last month? Was the 14 trading more than what we expected? It is normal or abnormal? What is the probability of continuation of such activity ratios? 32
As you can see, tracking system requires more work than just a casual glance at the lower limit of performance. We need to make a comprehensive assessment, to think carefully and then try to make up some of the expected standards of performance, to get something to compare with our current performance.
Historical performance tests The best way to develop indicators by which we will evaluate our current performance, is a neat collection of relevant data on productivity at a time when we are still in the testing phase of our system on historical data. The data obtained during testing can then be compared with the current data in real-time trading. As an integral part of the tracking process is important to be able to determine the source of the problem, after we had been warned about her possible appearance. Returning to our example of the 14 bids for a month, it would be very helpful to know that 9 out of 12 losing trades were held in currency, and upon closer inspection, we received three consecutive losing positions related to the yen, the Swiss franc and the German mark. A closer examination may reveal that perhaps there was a central bank intervention or other unexpected factor that made all nine losses on our different currency positions. Now it is especially important to know that our past experience and testing have shown that we can expect five consecutive defeats for the currency, and that losing three in a row is not an unexpected event. Knowing this will lead us to the conclusion that there was essentially nothing wrong with the inputs and outputs of the system. We may also decide that it is unwise to have a lot of foreign currency positions at the same time on the same side of the market. If our historical data show that we could expect a 5 defeats in a row in trading one currency, we should be ready for 15 defeats in a row in the trade three currencies in the same direction. It turns out that the problem has more to do with the choice of portfolio and product range, with the job market than time. The following statistical studies should be conducted on the historical data and then applied to the current action. You may find it helpful to gather the average or normal expectation for each statistical research and noted the range of extremes, so you can quickly move the figures for the future. For example, our historical data on T-Bills could look like this: The frequency of trading for the period - an average of 1.5 trades per month, with extremes from 0 to 5 trades per month (worst case on tests and in real time). Percentage of wins against the losers = average 32 percent of the winnings, if I win 8 of trading on the worst period and 6 wins out of 7 trades at the best period. The longest series of consecutive defeats = 7. = 5 consecutive wins. The average gain for the lucrative trade = S1420. Maximum gain = $ 5330. Average losing trade = $ 490. The biggest loss = $ 2,700. The biggest loss account when trading Treasuries = $ 7880. The time required to recover from the greatest loss = 11 months. The longest period for the creation of a new peak in account when trading Treasuries = once went 21 months without a new peak on the account. This is just a small example of the wide variety of data that could be useful in tracking system. You could improve and expand this list of my own thoughts about what you would like to know. The collection and maintenance of these data it seems a lot of work, and the way it is. But the work done by the trading system should be almost entirely mechanical, 33
so that your free time can be spent on an objective monitoring system instead of a simple observation of the process of trade. Monitoring the system without predefined objectives of this observation leads traders to finding reasons to rewrite the system or change it without an objective necessity. Most traders have a tendency to make mistakes, making too many unnecessary changes to their system after a series of losses. Losses may well fall within the range of normal expectations and be called by a factor that is not the fault of the system. Most traders never worried about the definition of normal and abnormal performance (except for the lower limit of the results), and thus, panic often occurs when there is no reason for it. Let's go back and take a look at the other side of the month when we earned a little money, but received only two of the 14 winning trades. Suppose our thorough analysis revealed the following: of the 12 losses, six were consecutive losses on treasury bonds while the remaining six were split between five different markets. The obvious problem is the cumulative losses of unprofitable trading treasury bills, which amounted to a total of $ 2400 (an average of $ 400 to trade with a maximum loss of $ 850). A close look at the data tells us that the number of trades on T-Bills for the month (six) were abnormal and actually set a new extreme value, surpassing the previous extreme at 5 trades per month. However, the overall loss ($ 2400) as well as the average loss ($ 400) were small compared with the previously collected data. Maximum loss in sequence ($ 850) was not peak. Six consecutive losses were extreme of, but close to it. Conclusion: We have been through a very unusual period in trading Treasuries, which attracted attention, but we would not have to replace the trading system at this point. We filled up our expectations interval to trade treasury bills and now have six trades per month as extreme activity. The level of activity should be monitored carefully (do not wait until the end of the month, and then later we find ourselves in front of 12 consecutive losses). If an unusually high level of activity continues, we will try to understand whether there is a fundamental reason that would lead us to believe that this chopper was only a temporary factor that will disappear by itself. If it turns out that the problem lies in the system itself, we may want to use a slower indicators or add an item to confirm the trade Treasuries in an attempt to reduce the activity and twitching.
A short review Monitoring system - the last element in our disciplined approach to its construction. At the very least, we hope that our intuition helped readers to identify the main issues that should be considered. There are many solutions for each of the seven issues that we presented, and we offered only a few possible alternatives. In addition, our proposed solutions, serve more to illustrate our way of thinking and approach. These techniques may not be for you the best solutions, and perhaps they are not the best for us. We will constantly seek new and better ways to solve each of these problems, and also to act and advise you. But before you start searching for answers, you need to understand the nature of the problem and assess the need for and benefits of their decisions. As you can see, building a system takes more than just finding the perfect indicator that we like. If it were so, we would all be much better off. Remember - each offers the advantage comes at a price. Make sure you have identified all the subtle flaws before decide that a particular solution is the best. When you are confronted with a roughly equal ways of solving problems, prefer simple complex. Try to be logical and objective at all times and do not let your emotions and unfailing optimism or pessimism take over. When making plans, always assume and prepare for the worst and be thankful, if the worst happened. When it happens, thank yourself for insight and preparedness - you will survive and get rich. 34
Recommended reading Sweeney, John. "Using Maximum Adverse Excursions for Stops." Technical Analysis of Stocks and Commodities 5, pp. 149-50. Wright, Charles.lt The Magic of Setup and Entry. "System Trading and Development Newsletter l.no.1 (October 1989), pp.2-5.
Chapter 2: Technical Studies Introduction Types of indicators Every successful trading system must have an objective iterative method for entering and exit from the market. These methods can be divided into several types. This is a classic stock charts Edwards and Magee, the benefits of which can be attributed to the simplicity and accessibility of understanding, and the shortcomings - excessive subjectivity. This is followed by a very sophisticated mathematical techniques such as autoregressive integrated moving average (ARIMA - autoregressive integrated moving average) or Fourier spectral analysis. We have seen in the non-obviousness of what is mathematically complex models set a time superior to any other types of analysis. There is a surprisingly large number of traders who want to believe that the basis of the market there is a certain structure, which, if open, will lead to wealth. These traders are using some very popular methods such as Elliott Wave, Gann analysis, and even astrology. These strategies have faithfully defended the small but close-knit group of traders fanatics who call mainly to trust and can offer a bit of data or logic to support their beliefs. They work long and hard, silently suffering their losses, believing that their failures are determined solely by their own lack of ability to determine the parameters of the truth, which they believe is hidden in the markets. Scattered wins give them the means of subsistence, while the loss becomes a punishment for their lack of work or lack of money spent on the opening securely guarded secrets that control the market. Unfortunately for them, these mysterious secrets were revealed only in part on expensive seminars and private consultations with the current guru (which is as big-hearted friend of humanity wishes to share his revelations just a few hundred dollars). We believe that if there really is some underlying structure and controls markets, it is obviously not yet disclosed. But most of all we believe that someone would have found the main secret to wealth, he is unlikely to share them with us for what else the money. Finally, in our listing types of indicators, we come to those which are the subject of this chapter. This is a group of computer generated technical research is a relatively simple indicators, usually displayed on the prices. Far from the secret to wealth, this is the usual plain and simple methods that will never be able to open any new truth about the markets. 35
These relatively well-known indicators have been developed in order to give us the easy to understand the signals that help you enter or exit from the market. The best part has already been written about all of these indicators. But we feel that much of the literature in this area is too academic bias and gives little in terms of practical application and management. We do not have a bone to pick with the academic analysis, and we actually indebted to him for what he taught us. However, we feel that most traders do not see many techniques beyond the obvious practical applications and limitations for each indicator. Also, they do not see and the many ways in which these techniques can be woven into the trading system.
Learn how to use indicators Novice traders often tell us that they do not have a clue what he was saying most of those wavy lines on the screens of their monitors. Even experienced traders can find these bends a bit frightening. We will try their best to patch some holes views information and advice based on our experiences and the experiences of other traders, which they shared with us over the years. You'll find a few common threads that run through the different types of analysis. Each of these has its place and has certain useful features in a well-designed trading system. Some technical indicators work best in trending markets, while others perform better on non-trending. Although all of these indicators are based on low, each calculated differently. For us there is no better or worse, or even a loved one technical study. We'll tell you what we think is the best and most efficient use of each indicator, as accurately as possible. There will be plenty of options without right or wrong answers. So, in the end, you decide which one you want to use and how. Except for the direct study of the volume and open interest, which we will look at the rest of the study are derived solely from the price. In essence, they are all included in the available commercial software for technical analysis.
The reduction of mathematics to a minimum In most cases, the mathematical basis of each trial is relatively simple. We'll try not to go into the math, but in some cases it is important that users understand how and why the calculations lead to a certain result. This is important for two reasons. First, you should understand exactly what light does and what does not. Second, you will find that many of the analytical software packages on the market differ in the way calculations of what should be one and the same technical study. Downtime RSI oscillator from one program may look completely different from that expected identical study to another application, notwithstanding the fact that data can be derived from a single database. The reason may lie in either a programming error, an incorrect understanding of the formula by which it is considered an indicator, either for personal predilections software developer. In most cases, the differences are minimal, but we urge you to be prepared for possible anomalies. If programming is not only outrageously bad, these variations are not in any way invalidate the technical studies. This book is not meant to be comprehensive, where we try to tell a little about everything. Several of these have already been written. If we missed your favorite technical indicator, it is because we do not dare to speak authoritatively about the research that we actually did not use and did not feel comfortable with them. In our knowledge, there are gaps, perhaps some of the important ones. We will not discuss the methods or 36
indicators clearly implying that the method or strategy is defective. We are confident that we are affected only the top layer, but we have no intention of trying to write about something beyond the boundaries of our competence.
Exchange of ideas We do not want someone decided that we invented or come up with all the methods of analysis and techniques, which are described in this chapter. We have extensive experience in trading, research and testing, but many of the best strategies fall into our field of view with the traders and subscribers to our leaflet "Technical Traders Bulletin". One of the main purposes of the leaf - serve as a venue for the exchange of these trading ideas. As the editors of technical publications, we regularly listened to the ideas of traders from all over the world. Essentially, all of our subscribers are actively involved in the futures markets, but a few traders in securities that are prepared to do futures. The majority - is trading advisors, brokers, bankers, dealers FOREX, home, technical analysts, and the like. The others - these are private investors who have a computer and an analytic software. We are grateful to everyone who has ever called and passed on the information, opinions, experiences and ideas, even if they thought that at the time of their proposal was not material. We have tried to gently separate the wheat from the chaff as far as possible and see yourself what works and what does not. As you will see, we sometimes dare to express views that may be unprovable one way or another. However, we tried to add enough of the facts and logic to give weight to our opinions and conclusions.
Directional Movement Indicator (DMI - Directional Movement Indicator) Average Directional Movement Index (ADX - Average Directional Movement Index) The vast majority of profitable trading systems include some form of trend-following, but most of the time are not on trend, strong enough to bring revenue worth. For the reason that successful traders use tactics to obtain small losses and allow income to flow, non-trending markets seem to bring only small losses. As a result, those who follow the trend, and usually lose money most of the time in most markets. Their cherished dream of the success is due to accidental finding of the market with the trend strong enough to bring big profits. A general method for "finding" the big trend is to invest in various markets hoping to get on one of the most profitable markets. Unfortunately, such an investment unprofitable markets adds more than winning. The usual procedure of investing is to find the best market results through contact with a couple of good markets in forced patience of a wide range of bad ones. Fortunately, there is a very practical solution to the problem of defining and measuring direction of the trend of the market. The correct interpretation of the Average Directional Movement Index (ADX) allows traders to significantly improve the effectiveness of finding good markets and bad clipping. We probably spent more and more research work with 37
ADX, than any other indicator, because it found the ADX surprisingly valuable technical tool with many practical applications. In order to give our readers a complete picture of the ADX, we must start with a basic explanation of directional movement indicator (DMI), which is used for ADX.
The concept of DMI Directional movement - is a concept that J. Welles Wilder Jr. first described in his book "New Concepts in Technical Trading System" in 1978, in the classic work on technical analysis, which we wholeheartedly recommend. (See "Recommended Reading" at the end of the chapter.) Direction indicator (DMI) - is a useful and versatile technical study, which has two great features. First, DMI itself is a good indicator of market direction. Second, one of the derivatives DMI is an important index Average Directional Movement (ADX), which not only allows us to define markets in a state trend, but also provides a way to assess the strength of trends. The calculation of directional movement (DI) is based on the assumption that when there is an upward trend, today's price peak should be higher than yesterday. On the contrary, when there is a downward trend, today's lower price must be lower than yesterday. The difference between today and the peaks - a movement up or + DI. The difference between today and yesterday's valleys - a movement or down-DI. Internal days when today's peak or trough at most of yesterday, essentially ignored. Positive and negative DI separately averaged over the period of a few days and then divided by the average "true range". The results are normalized (multiplied by 100) and are shown as oscillators. For readers with mathematical inclinations we have included detailed calculations. Fortunately, we are now able to produce the necessary indicators of only three or four taps on a computer keyboard.
The calculation of directional movement (DM - Directional Movement) 1.Napravlennoe movement - is the largest part of today's price range, which is beyond the boundaries of yesterday's range.
2. External days will have a + DM, and-DM.
Use more.
3. Inside days have zero DM. 4. Limit days will be DM, calculated as the diagrams shown above. For example, the upper limit of the day (first diagram) + DM is the difference between A and an upper limit, reached at day C.
Calculation ADX 1. Measure the directional movement (DM). 2. Measure the true range (TR - true range), which is defined as the higher of: a) The distances between the current peak and the trough today. b) The distances between the current peak and yesterday's close. c) The distance between today's and yesterday's closing depression. 38
3. Divide the DM TR for the direction indicator (DI-directional indicator).
DI = DM / TR Result may be positive or negative. If it is positive, it is the percentage of the true range, which rose for the day. If it is negative, it is the percentage of the true range that goes beyond day. + DI and-DI are usually averaged in the time period. Wilder recommends 14 days. Then we have the following calculations: + DI14 = + DM14/TR14 or-DI14 = -DM14/TR14 + DI and-DI - Two of the three values are usually displayed as DMI. Third - it ADX, obtained as follows: 4. Calculate the difference between the + DI and-DI. DI DIFF = | [(+ DI) - (-DI)] | 5. Calculate the sum of + DI and-DI. DISUM = | [(+ DI) + (-DI)] | 6. Count the Directional Movement Index (DX). DX = (DI DIFF / DISUM) * 100 DX 100 normalizes the value so that it falls between 0 and 100. By itself, DX is usually very volatile and will not be published. 7. Count the moving average for DX Average Directional Movement Index (ADX). Typically, the smoothing is performed on the same number of days, and that the calculation and + DI-DI. 8. Further smoothing can be done calculating the time derivative of the ADX type called a ranked Average Directional Movement Index (ADXR-average directional movement index rating). ADXR = (ADX t + ADX t-n) / 2 where t - today and tn - a day that began with the calculation of ADX. Displayed on the computer screen as an oscillator, directional movement moves up when the + DI is more-DI. If + DI is less than-DI, the movement is downward. Since the divergence of the two lines directed movement increases. The greater the difference between the + DI and-DI, the more the focus of the market or the steeper trend. Wilder used the 14 days based on his calculations, because he believed 14 days of an important half-cycle. We think that there are more optimal periods of time, depending on what you intend to do with DMI and ADX. Studies on the computer monitor DMI usually occur in the form of three lines: + DI,-DI and ADX. (Some programs are for convenience ADX separately.) As we said, the results of calculations DMI normalized (multiplied by 100), so that the line will vary in the range from 0 to 100. An important indicator ADX is derived directly from the + DI and-DI measures the amount and trend of the market. The more ADX, the more directional movement of the market took place. Accordingly, the lower the ADX, the movement of the market was less directional. Note that when we say "directed" we can have a direction in mind, both up and down. ADX does not distinguish between rising and falling markets. It is important to understand clearly that ADX measures the trend and not its direction. For ADX is completely normal clearly come at a time when prices are falling, because it reflects the rise of its increasing strength of the downtrend. Other oscillators, + DI and-DI, indicate the direction. When the + DI with-DI crosses and goes higher, the trend is up. When the + DI with-DI crosses and goes lower, the trend is down. The farther apart then the line, the stronger trend. (See Figure 2-1) 39
In his book, Wilder also describes the calculation of the ranked average directional movement index, or ADXR (average directional moving index rating). This is simply the sum of ADX early period (say, 14 days ago) and today's ADX, divided into two. This additional smoothing ADX Wilder was made to reduce the fluctuations to the extent that ADXR could be used in the calculation by the method of comparison of markets, called 'index of product selection (Commodity Selection Index). From our point of view, ADX
was initially quite smoothed and additional smoothing optional. In fact, for our purposes, anti-aliasing, which was held for ADXR, reduces the effectiveness of the indicator.
Testing the performance of DMI It was published quite a few tests DMI and ADX. The results were generally better than those of most other indicators. Here are some examples. Bruce Babcock DMI tested and described the results in his book, "The Dow Jones Irwin Guide to Trading Systems" (see recommended books at the end of the chapter). Testing the DMI, Babcock was part of a long position at the close, when the general 40
direction of movement was positive. When the general direction of movement is negative, the system, by contrast, has introduced a short position, Babcock test results showed that the five-year period on a 28-day DMI was profitable on a wide range of markets. However, the internal losses were significant, because it is not used stop. The system tested Babcock was the most simple use of the indicator, and many of the basic rules Wilder had been violated. It is important that the proposal on the use of Wilder's expectations of a breakthrough when entering the top or bottom of the day crossing the DI was ignored (we found that Wilder's recommendation with respect to occurrence significantly reduces the jerking). In testing Babcock income undertook to clear intersections, and no attempt was made to generate income before. The fact that under these conditions, DMI showed a significant income, just amazing! Despite the fact that we do not recommend trading on DMI so Babcock testing showed that a fairly long DMI could be a useful indicator of the time of entering the job. A more realistic test / optimization was conducted by Frank Hochheymerom of Merill Lynch Commodities. Hochheymer tested two cases: Case 1, which followed the basic rules of Wilder, and Case 2, which is simply sold at the intersections. In most markets, were used 11-year data. Since this test was also optimized, he tested the + DI and-DI by independently changing the number of days used in each of them (that we do not recommend). Not surprisingly, the case of one who has followed Wilder's proposal to join the purchase or sale at the level of the previous peak or trough of the day, proved to be more profitable. Optimization periods DI showed that the best time intervals were in the range of 14 to 20 days. Our independent testing ADX on a different set of data confirms the profitability of the range of 14 to 20 days, with the best results shown in 18 days. In "The Encyclopedia of Technical Market Indicators" Colby and Myers (The Encyclopedia of Technical Market Indicators, Colby and Meyers) was held very interesting test DMI with built ADX. They were at the intersection of + DI and-DI, ADX only when raised. They came when the ADX dropped or otherwise arise intersection. They only tested New York Composite weekly data using intervals of 1 to 50. Top income was received at time intervals of 11 to 20 weeks. They noted that many of the indicators tested their method DMI had the lowest losses and deserves further study. At first glance it may seem that Colby and Myers followed the trend, trading only on the rise ADX. However, since they used trades based on the + DI and-DI after rising ADX, the system was more than counter-trend method because rising ADX was the result of the presence of the trend prior to the current intersection. When the + DI and-DI lines crossed after rising ADX, it was a signal to trade in the opposite direction of the trend, which was measured rising ADX. We find the ADX moderately useful as an indicator of the set time, in spite of some positive results of testing / optimization mentioned earlier. DMI is an indicator of trendfollowing, and subject to the same weaknesses that are common to any form of trend following. When the markets are not in trend, + DI and-DI lines intersect in different directions constantly, producing a painful twitching after another. These are sensitive indicators that give good results in trending markets, but just this sensitivity and leads to antsy when the market gets into a sideways trend. However, we are very excited about the use of a derivative ADX DMI as a filter, which will help to choose the most successful trading method for each market at any time.
Using ADX We suspect that the ADX indicator is often neglected because of the obvious drawbacks associated with the lack of correlation with price movements. Someone 41
casually exploring the rise ADX during the fall in prices, could conclude that the indicator will give false signals about the direction of the market. It is extremely important to understand correctly from the beginning, that in itself is not ADX indicates the direction of the market. ADX may fall during a rally and grow during their fall. Appointment ADX is to measure the strength of a trend and not its direction. To determine the direction of the market you have to use additional indicators such as DMI. (See Figure 2-2.) Some technical analysts attach great importance as an indicator of the level of ADX strength of the trend, and they would argue that the value of 28 indicates a stronger trend than the value of 20. We found that the direction of ADX much more significant than its absolute value. Changing up, for example from 18 to 20,
shows a stronger trend than the negative change from 30 to 28. A good general rule can be formulated as follows: while the ADX is rising, any value of ADX above 15 indicates the trend. We recommend that you familiarize yourself with the ADX and share it with your favorite technical indicators. You will soon discover certain levels rising ADX, which produce outstanding results with your favorite indicator. One indicator works well when ADX rises the mark of 15, and the other - when the ADX rises above 25. When ADX begins to decrease at any level, this is evidence that the market has gone away and forms a sideways trend. We explore the importance of growing and falling ADX more detail and suggest appropriate trading strategies. (See Figure 2-3.)
42
Growing ADX Growing ADX indicates a strong trend on its way, and is intended to include trading strategies trend following. Technical indicators for which higher incomes need strong trends, such as crossing moving averages and breakthrough methods should work just fine. Almost any method of following the existing trend should give excellent results in a supportive environment, the predicted rising ADX. (See Figure 2-4.) Keep in mind that the growing ADX also provides valuable information about what trading technology can fail. Knowing what to do, can be just as important as knowing what to do. For example, a popular trading technique uses oscillators repurchase / resale, such as RSI or stochastic oscillator, and is looking for sell signals when the market is trading at overbought. However, if the ADX is steadily growing, this should serve as a warning of what is a strong upward trend, and the oscillator signals for the sale to be ignored. When the ADX is rising, the indicators repurchase / resale tend to approach a particular extremum and remain at that level, giving the repeating alarms to trade against the trend. If you follow the signals of the oscillator, the losses can be very significant. The fact that the ADX is rising, do not necessarily mean that we can not use our favorite oscillators. It just means that we have to accept signals coming in the direction of the trend. (See Figures 25 and 2-6.)
43
44
A falling ADX A falling ADX indicates the trend-less market, where the counter trend strategy should be used in place of a trend-following methods. Oscillators repurchase or resale, which give buy signals in the hollows, and with an increase in the sale are the preferred strategies at a time when the market is in a trading range. Indicators such as stochastics and RSI, should give the right signals when the price fluctuates within a limited region of its trading range. Due to the fact that buying and selling in the hollows on the climbs can get, at best, a very modest income, many traders prefer to trade only in the direction of the main trend. In such a case it would be best to just ignore the signals following the trend during the fall ADX. Of course, the ideal would be to have a profitable counter trend strategy in addition to a trend-following strategy and apply each of the methods being coordinated with the direction of ADX. (See Figure 2-7.)
45
Problems ADX: Spikes We would have been a disservice by claiming that the ADX will solve all the problems that can only come across a trader. At ADX has its drawbacks. One of the problems is that for a long period (we prefer 18 days, as mentioned earlier), which is best applied in most markets, ADX suddenly changes direction, taking the form of spikes. Spikes usually occur at market peaks when prices suddenly go from a strong uptrend to strong downward. ADX source of the problem is that he can not recognize a new downtrend. ADX will still be included in its computation historical period with a strong movement in the positive direction, taking the same period of time, new data with a strong movement in the negative direction. The conflict at the entrance, ADX will fall some time as long as the old movement in the positive direction will not be expelled from the data, and then ADX start growing again because of a new downtrend. In a market that has produced a spike, ADX can not warn of a trend in time, it will not allow to catch most of the rapid downward movement. (See Figure 2-8.)
46
We will try to find a solution to this problem. One possibility is to switch to a shorter period ADX, when the market is at a level where you can expect a spike. We noticed that some markets often produce spikes (eg, metals and grains), while others tend to produce flat tops (treasury bills and securities). ADX very good job with flat tops with no such problems, which arise on the thorns. We prefer to refrain from any subjective classification of markets, if at all possible, so that we continue our search for a more objective decisions. Fortunately, market depressions rarely take the form of spikes, and ADX makes a very timely work to identify trends upward in the course of their development.
Problems ADX: The delay One of the characteristics ADX, which may result in a problem is that it is slightly slower than many other technical study. When ADX begins to rise, many trend-following indicators already give a signal to the entry. For example, + DI and-DI cross before ADX begins to rise. More than likely that during this early signal to the entry ADX was still falling, so the entry will be ignored. In practice, this situation becomes a signal growth ADX time to enter the market in the direction of the trend. Faster technical study is able to determine the direction of the trend, and ADX is used to set the time of inclusion. While the trend faster indicators may provide additional signals entering that if ADX continues to grow, you need to follow. You will find that it will take some thought and planning to coordinate the ADX with other technical tools. We consider the delay as a small price you have to pay in order to avoid costly twitching that may arise in the case of entry into the trade during deflection ADX. However, the delay time can be set depending on the characteristics of the market and the individual preferences of the trader. Several markets are more likely to be able to trend than the rest. For example, the currency markets were moving well in recent years. In markets that have shown good performance trending, ADX time period can be shortened to work faster signals. If the delay of the inputs frustrates you, shorten the period ADX. If you are upset twitching, save ADX period of 18 days. Latency is not a problem when using the counter trend strategy during the fall ADX.
Day Trading with ADX
47
Perhaps because of the distortion caused by large gaps between yesterday's close and today's open, ADX does not work so well when applied to graphs with a period of less than one day. Using the 5-minute chart and ADX with a period of 12 gaps between the opening and closing can be destroyed after-hours trading, and the ADX give regular information about the strength of the trend for the first hour. However, many day traders prefer to use the 20-minute or 15-minute charts, and in this case it is difficult to avoid possible distortions of DMI and ADX, caused by the gaps between the opening and closing. Most often, a standard 18-day ADX can provide valuable long-term information that helps in day trading. A day trader should pay attention to the presence of any trend indicates a growing ADX, and to enter into short-term trades only if they are in the same direction as the trend. When ADX falls short term trades can be carried out in any direction. Virtually any method of day trading can be improved by examining the first ADX to determine the existence of a trend. (See Figure 2-9.)
In short, we believe ADX one of the most useful technical indicators. When we sell on our control programs, we usually prior to further analysis of the first thing considering ADX. We find that the measure of trendiness, extracted from the ADX, is an invaluable guide to choosing the best strategy for each market. A simple but very important information provided by ADX, can increase our percentage of winning trades by a significant amount. Many of our test results following the trend only in the case of lifting ADX clearly demonstrate its importance and value. Waiting lift ADX often means detain about our desired time of entry, but the belief in the success of a mandatory trade together with the obvious benefits of reducing the number of losing trades are more important award. In addition to its usefulness for inputs, ADX can be an exceptional tool in determining the time of the outputs from the auction. Important model, marked by Wilder, it's a short term top or bottom, portends cross lines + DI,-DI and ADX. The turning point of the market often occurs with the first pass down the ADX line, after crossing the ADX bottom up first + DI, and then-DI. We agree with the conclusion of Wilder that this turn down can be a good time to fix the revenue trend following, or at least the closure of most contracts that are part of the lucrative multikontraktnoy position. (See Figure 2-10.) ADX can be very useful to get out another way. When ADX falls, it shows that we should receive a small income instead of letting income flow. When the ADX is rising, it shows the possibility of higher earnings, and therefore we should refrain from premature 48
aging. The possession of precise indications of when to take a small profit, and when to wait for higher incomes, can be a great asset for any trader. It is rarely mentioned ADX application can be as important as its use in selecting the vehicle entry.
Bands (Bands), Envelopes (Envelopes) and channels (Channels) Trading for the envelope formed by straps around the moving average (or around a similar indicator) - a well-known and effective method of smoothing short-term twitching common to all systems of trend following. Most technical traders have to use this method once. Essentially, all of today's software packages provide the ability to demonstrate basic forms envelopes around the moving average. The manifold channels of interest and envelopes can be used as the basis profitable trading system. We divide our discussion into two parts. The first part will focus on the envelopes, which are formed by the installation of the bands around the moving average. The second part will be devoted to the channel breakout systems.
Section 1: Trade with envelopes 49
Envelopes can be as simple or sophisticated as you want to make them so. The most simple - a simple moving average with stripes on each side, is calculated as a percentage of the moving average at the interval. For example, the 10-day moving average with the bands removed by 5 percent from the average. The area inside the two bands theoretically acts as a buffer zone to contain costs within themselves, mainly when the market is in a trading range. Start the trend will be indicated when the price band strikes. At the end of the trend correction or prices will drop back into the band in the direction of the moving average. (See Figure 2-11.) Another simple example is the use of the envelope points to the absolute value on each side of the moving average. For example, the 10-day moving average of U.S. Treasuries may be surrounded by boundaries, which are 1 and "'A,, points, or $ 1,500. Usually this fixed dollar envelope will be used more for Offering excellent control of risk, rather than as a method of entering into new auction.
Variations of the above two envelopes almost innumerable. For example, the moving average can be smoothed exponentially, or otherwise. Percentage of prices contained within the bands may vary depending on the purpose for which the position is considered a long or short, the slope of the study in the direction of increased volatility in the direction of the trend. For example, in the rising market of the strip can be placed at the level of 5 per cent above the moving average and 10 percent lower. Another possibility - using moving averages peaks and valleys as the strips on each side of the moving average of closing. Bands thought to contain within themselves and determine the price movements within the trading range. Any breakthrough beyond the bars must signal the start of a trend, because prices do not wander inside the envelope.
Bands Bollinlzhera (Bollinger Bands) 50
Enough already popular and recent additions to the list of research channels are alpha-beta band, now more nazyvaemy Bollinger Bands (after John Bollinlzhera, market analyst at CNBC / Financial News Network). Alpha and beta bands Bollinger bands are statistically defined around short-term moving average. Computer software first calculates the simple moving average, and then consider moving standard deviation from the mean. (See Figure 2-12.)
Ballinger typically uses a strip having two standard deviation on each side of the moving average. He explains that the two standard deviations theoretically contain the vast majority of the subsequent data. He also points out that the calculation of the standard deviation uses a number of deviations from the average price, making the calculation is very sensitive to short-term price changes. Bands are expanding rapidly and interact with the volatility of the market, becoming sensitive to the recent market movements. Recommended installation - 20 days with an envelope of two standard deviations. Both values change frequently depending on the test of the market and the purposes for which the band used. Bollinger bands are usually used along with other technical studies to determine the trend turns on the stock market. If the prices are close to the lower limit and other research confirms the rotation, therefore, be safe to buy. For example, the divergence RSI may be used to confirm the lower bottom of the envelope. (See Figure 2-13.) Instead of two standard deviations from the moving average, which is probably only suitable for securities trading, alpha-beta bands can be set at any distance from the moving average. Typical installation bands alpha-beta for the futures market is only one standard deviation on either side of the moving average. The basic concept of statistical output width of the envelope is that the current study volatidnost specific market determines the location of the bands. The use of these bands self-aligning means that volatile markets will automatically receive a broad envelopes and less volatile markets will have a more narrow envelopes. One caveat about the statistically-defined bands: the theory of selection of sample claims that 30 data points - this is the minimum required amount to obtain a statistically significant result. Bollinger commonly used 20 days or less. We can say that the bands contain some of the resulting data, but it does not mean that they are using any normal statistical procedures. Bollinger clearly warns against the establishment of any statistical conclusions beyond the simple observation that the band will contain most of the recent market movements.
51
Trade rules envelope There are almost as many trading rules for envelopes, and how many of the rules for their construction. Rules are based (or should be based) on the idea that the envelope contains a significant amount of market price movements and the movement of one of the bands is a deviation from pricing behavior, and to such an event to react.
The following are the traditional trade rules for envelopes: 1. Enter the market in the direction of the break in the time of crossing the strip. This signals the beginning of a possible trend. 2. Go out and change positions when crossed the opposite lane. (See Figure 2-14.)
52
We recommend using the intersection, based only on the border closings for bands to avoid some twitching. Prices will often jump out of the border during the day, but closed within the bands. Alternatively 1. Come to the market, where the band is crossed as indicated above. 2. Come out of the market when the moving average is reached between the bands, but do not change positions. (See Figure 2-15.) Both sets of rules will ensure that the underlying trend is not to be missed. The first set of rules is a major and gives regular reversing system.
We experience the irresistible skepticism toward reversing systems and prefer the second set of trading rules. The second set of strips are also used to enter, but the output is used for the moving average. If the prices are within the bands after the cessation of trade, the market in the neutral zone and new trades will not be until the new breakthrough. Another reason why we prefer the second set of rules is that the theoretical risk for each trade reduced to the distance between the strip and the moving average, instead of the total distance between the strips.
"Optimal" percentage for the bands The "right" value for the moving average and the surrounding envelope illusory. The most extensive testing that we've seen, covering the period from 1960 to 1978. It appeared in an article by Irvine and Ugriga in the December issue of "Review of Research in Futures Markets" (see "Further Reading"). The authors used a breakthrough bands for coming in and going out at the intersection of the moving average inside the envelope. (They used a second set of rules mentioned above.) Then they optimized the system to find the best combinations. Below are some figures that they have found the most profitable:
53
Even a cursory glance at today's markets shows that these values are no longer optimal, and they only serve to illustrate the futility of optimization (some important caveats regarding optimization are discussed in this book devoted to system testing, Chapter 3). As well as moving averages, envelopes work well in trending markets, and not very good at changing direction frequently markets, and the "best" envelope changes over time. A common optimization in order to find the correct value is useless. We recommend to trade with the trend when the price jumps beyond the envelope, and use counter trend methods when it is inside.
Trade within the envelope Logical and effective technique, discussion of which we have rarely met, involves the use of strips as indicators of repurchase / resale in such a way that the trade is conducted within the band, not the outside. We have used this method very successfully, when the markets were in a sideways trend. Trade rules are clear and relatively simple. Buy as soon as the price touches the lower band. If the trade goes against you, which will close at the lower boundary of the strip, quickly get out and take a small loss. If the trade starts moving in your direction, as so often happens, stay profitable positions and change the trading position on the top of the strip, applying the same rules, just the opposite. This method seems to be effective because it combines the tactics of taking small losses and higher incomes, with a trading strategy of buying and selling in the hollows on the peaks. The real challenge is to make sure that you are in a sideways trend. How do you know whether a market is in a limited trading range or trending? Objective method is to use the ADX, as we described in the section on DMI / ADX. If ADX is falling, trade inside the envelope can be very advantageous. If the ADX rises, the market is in a trend, and you'd better use the envelope following the trend. (See Figure 2-16.)
54
Section 2: Trading breakouts channel In addition to envelopes, which are determined by the distance from the moving average, there are channels defined by the peaks and troughs in a certain time period. The simplest of these methods is the channel - the usual reverse system that is always in the market. For example, the upper band is formed by a peak in the last 10 days. The lower band is formed by the lower price for the same number of days, and both bands form a channel. The channel is changed in width, while the old peaks and valleys will be replaced with new ones. The system gets into a long position when overcome by the higher price, and remains in a long position, as long as there is no broken lower price. Since it is a reversible system, then when the long position is closed, there is a short. There have been many tests showing that the channels are one of the most effective and affordable trading tools. Perhaps the most famous of these was a study testing the channel break spent FrenkomHochheymerom of Merrill Lynch about 10 years ago. Besides Hochheymera, many other respected traders celebrated the dignity of trade channel. For example, Richard Donchian, despite the fact that he is best known for his work with moving averages, is also known for its trade in the channel using the four-week rule. (See Figure 2-17.)
55
This system was made public Donchian in the 60s as a "Week Rule." He used a four-week time axis, buying when prices overcame a four-peak, and selling when prices break through the four-week trough. Testing has shown that the system gave a reasonable profit, despite the fact that from the enormous losses periodic simply breathtaking. As you can imagine, the risk taken in each period was virtually unlimited and depend only on the size of the four-band chain. Because the system is not applied the stop at the aggregate score total portfolio risk was enormous. (See Figure 2-18.)
Very popular and expensive system, promoted in the 80s, was basically the same as the methods and Donchian Hochheymera, except that the temporary grid periodically overoptimize for each product. After many years of very profitable trading losses in 1988 were so severe that many users of the formula were forced to stop its use. In defense of the system can be said that 1988 has brought misfortune / aphid many trend-following methods.
56
Selection of time values What is the optimal number of days to be used in the construction of the channel break? Merrill Lynch study Hochheymera gave the following number of days for the optimized channel breakout systems:
As might be expected, with optimized study, these channels have been extremely lucrative six-year testing period (1970-76). However, even with the advantage of optimizing only 42 percent of the trades were profitable. It should also be noted that the losses were very substantial and that the four-channel for silver trades made in 1866 (more than one trade per trading day). Create a channel breakout system with optimized values for each market simple, but in our experience, these systems fall apart rather quickly. As shown in a test method Donchian Bruce Babcock, a single value can work and be profitable for all of the portfolio (see Babcock's «The Dow Jones-Irwin Guide to Trailing Systems»). In fact, with the exception of S & P 500, the system will bring excellent returns. William Gelecher in his wonderfully witty book, "Winner Takes All: A Privateer's Guide to Commodity Trading" tested the 10-day breakout system on 10 different products for the period 130 weeks. The results showed that this simple 10-day channel generated revenue to deserve the respect of 24 per cent per annum. (By the way, we do not know yet whether the book is published Gelechera, but if you see her, buy. This is one of our favorite activities.) Lucas, Brorsen and Irwin tested 12 different trading strategies in the 12 actively traded markets in the period from 1975 to 1984. Three of the tested trading systems are systems of channels. The main channel system, they explored, brought the annual income and had the highest total return (33.4 per cent per annum) of all the systems in their study. Directional Parabolic was in second place, and one of the modified systems channel was third. It is interesting to note that those are optimized time pe - period for the channels that they have published, differed significantly from the developed Hochheymerom in his study. Our testing and experience in the channel breakout, which is quite extensive, shows that 18 days is a good interval that works in many markets over long periods of time. Frankly, we are of the opinion that almost any interval in the range of 10 to 30 days will eventually be profitable. The losses will be different in magnitude and occur at different times depending on changes in these intervals. Reduced risk by introducing a neutral zone Constructive way to reduce the losses inherent channel breakout system, without having to sacrifice too much of the potential income has been developed manage cash our 57
friend in Southern California. His system used different time periods for entry and exit. Weekend lane for his method were shortened to half of the time period of input bands. For example, if a signal to buy soybeans is a breakthrough peak of the last 20 days, the internal point of the channel and the exit point will be at the level of depression the last 10 days. This addition to the Donchian system has the advantage of strong reduction of the overall risk of the portfolio. It also leads to a reversible nature of the system, creating a neutral zone in which there is no trade. This is to give effect to reduce twitching in volatile markets and keep most of the revenue due to the accelerated outputs. Unfortunately for the control of cash and its customers, the invention is able to overcome most of the problems associated with a loss, but not all of them. After several years of outstanding performance of one of the large losses that are a characteristic lack of channel systems, collected his tribute, and this man had gone out of business.
Breakthrough of the channel as confirmation We found in our testing that some types of trend following entries can be significantly improved by using a recent breakthrough in the channel as confirmation. Contrary to popular belief, the new peaks and troughs are not always a bad place to enter the market. Furthermore, if gold goes from $ 300 to $ 800, it always sets new peaks. Breakthrough this is another way to ensure that the trend is strong. It seems particularly effective if a breakthrough is a breakthrough as a reverse previous in the opposite direction, and not just set a new peak or trough.
The Commodity Channel Index (CCI - Commodity Channel Index) The Commodity Channel Index was first described by Donald Lambert in the October issue of the journal "Commodities" (now - "Futures") for 1980. Despite the 11-year history of the CCI and its presence in almost all software packages focused on futures, we know some traders who do use it. We do not know why, but suspect that one reason may be a lack of literature on this indicator, as well as Lambert's insistence in tying CCI to the theory of cycles. Despite the references to the cycle theory, the original article Lambert is probably still the most accessible explanation of how to use the CCI. Like most of the technical studies, CCI requires some understanding of its origin to its effective use. Mathematical and statistical concepts behind the CCI, a little difficult to understand at first viewing, because its formula is more complicated than that RSI, MACD and stochastic oscillator, which can be more or less intuitive. CCI partly a statistical formula that makes difficult to demonstrate the relationship between the graphs of changes in price and the indicator chart, obtained as a result. CCI provides a convenient formula for the number that statistically shows how far prices have gone past the rolling average. If prices have gone far enough, it is believed that the established trend and trading signal is generated. We tend to partition technical studies into two groups, those which are best used as a counter trend indicators such as RSI and the percentage of R, and the ones that go well beyond the trend, moving averages. CCI is an indicator of trend following.
58
Overview of the main theories of the Lambert CCI formula calculates the simple moving average of the average daily prices of [(peaktrough + + Close) / 3], and then calculates the average deviation. The average deviation is the sum of the differences between the average price of each period and the simple moving average. The average deviation is then multiplied by a constant (Lambert offers 0.015) and divides the difference between today's average price and the simple moving average. The result is represented as a single number that can be positive or negative. A trader can change the number of periods used to calculate a simple moving average. As you might expect, the shortening of the time span makes the code faster and more responsive to small movements of the market, while lengthening the time span index slows and dampens market volatility. CCI on the computer screen is usually displayed as a bar graph or an oscillator that fluctuates in different directions around the zero mark. Since the index measures how far prices have moved away from the moving average, CCI enables us to measure the strength of a trend. In theory, the larger the value of CCI, the stronger the trend, and so should be more profitable trade with the trend. (See Figure 2-19.)
Lambert originally developed CCI to find the start and end of the expected cyclical seasonal price patterns. He felt the need to have an indicator that could determine where the loop starts and ends. It seems a contradiction cyclic theory, because if you know that there is a cycle, you have to know where it starts and where ends, otherwise no cycle. A clear need to type CCI indicator indicates that the imaginary cycles had to be far from certain and non-recurring.
Lambert made the variable part of the formula with the moving average, so that the user can adjust in some way to the estimated length of the CCI cycle. His research showed that for best results the length of the moving average used in the CCI, must be less than onethird the length of the proposed cycle. However, tables of test results in this paper show 59
that the moving average five periods worked best, regardless of the length of the cycle (another sign of weakness Lambert assumptions about the existence of cycles). CCI uses a simple moving average, instead of the exponential, so the prices of the past will be rejected and will not affect the results. 0.015 certain arbitrary constant used in the formula CCI, was added to the index scale so that 70 to 80 percent of the values fall into the channel between -100 and +100 percent percent. The premise of the Lambert was the fact that fluctuations between the boundaries of the channel are considered random and of no value to trade. He proposed to establish long positions only when the CCI exceeds +100. A significant drop in the mark of 100 is considered to be a signal to the output of a long position. Terms of short positions are the same: to sell below 100, buy back above -100. (See Figure 2-20.)
As we mentioned earlier, Lambert has done research that showed that the length of the period of CCI must be installed by an amount less than one-third of the length of the cycle. He tested a number of different period lengths, finishing number 20 as a standard, but suggested that this number was chosen for each market individually. (We do not dispute that the length of the period should be set in such a way as to satisfy the historical data.) Twenty is the value used by default for the majority of the CCI program.
60
Some positive test results Colby and Meyers in his book "The Encyclopedia of Technical Market Indicators" tested the CCI on the weekly prices New York Composite, using the original trade rules., 0ni tried to find the optimal time periods. This procedure seems to us to fit under the curve, but the results are interesting. The most lucrative period of time tested proven to be very long - 90 weeks. However, in the range from 40 to 100 weeks yielded good results, and can be easily and profitably today as the 90-week period. Our warnings regarding optimization can be found in the third chapter. Colby and Meyers pointed out one important aspect of the 90-week CCI, which should not be surprising. CCI on the period of 90 almost always misses the early phases of starting a trend. In today's market of securities skip the early stages of a trend often means the omission of most of the potential income. The first studies of Lambert showed that more than a short-term CCI will likely outperform or coincident than a lagging indicator, and Lambert used the time period from 5 to 20 days. To adjust the time delay produced by the 90-week CCI, Colby and Meyers decided to ignore the + / -100 extremes and use the intersection of the zero line for more early signals for entry and exit. They called this display "zero" CCI and found it much more profitable than the original + / -100 signals. As an aside, let us note that in spite of the fact that when testing the trading system using the concept of zero CCI weekly data NYSE Composite Colby and Meyers got better results than the popular 39 - and 40-week moving average system, now protected by many traders of securities market it does not mean anything.
Using the CCI as an indicator of long-term trend Monthly CCI can be very effective as an indicator of long market trends. Take a look at the following monthly charts with CCI signals with a period of 20 to the zero line in place of the mark + / -100. The first graph (Figure 2-21) - on the Japanese Yen. In addition to the sequence of trading signals should note two other features of this graph minute.
61
First, the faster growth of CCI from 0 to 100, and the more the detected trend them. Second, the rapid fall after CCI for 100 usually indicates that the trend will expire and that revenue must be protected switching stops at this point. The graph treasury bonds (see Figure 2-22) are the use of trend lines CCI for early exits.
We recommend that you try to use the monthly CCI 20 period for long-term directional movement while using a short-term indicator for setting the time entry and exit in the direction of the monthly trend. This strategy would be particularly effective in the rapid growth of CCI from 0 to 100. After a month peaks CCI would be wise to consider the possibility of suspending the trade in this market as long as the CCI starts to grow again. The situation is similar to the monthly CCI, can be seen on the weekly chart. The rapid growth of between 0 and 100 should definitely mean a steady trend. Try to use the weekly CCI to set the time of trading in the direction of monthly charts when CCI is in a period of growth. Come out when the weekly CCI creates a peak or when another indicator warns you that the intermediate trend is losing strength. As an alternative strategy may be to start trading small lots, when I first crossed the zero line, and then add the item to the acceleration CCI and strengthening trend. Begin to close the position when the CCI stops, indicating that the market ends the movement. Trade set of positions based on the weekly charts will obviously work best in markets with slower movement and controlled risk, where the big long-term positions are preferred. (See the graph in Figure 2-23 Eurodollar).
62
Using fluorescent CCI Our study showed that 20-day CCI, used by itself, does not work well in most markets. Its main disadvantage of skipping the beginning of the strong trend can be really negative trait of fast and volatile markets. This slowness can be overcome by using a 10-day (or even short) CCI or by entering the zero mark. But the faster methods are extremely vulnerable due to frequent twitching. We can always set the CCI to meet each market, but we are almost sure that it's just adjustment to the curve and do not recommend this method. We recommend for daily trading CCI combination with other indicators. Since one of the problems CCI is its tendency to make mistakes in assessing the volatility trending markets, it seems logical to tend to the views of DMI / ADX as a duplicating trend indicator. If the ADX rises, the market is in a trend, and it can begin to trade the signals of CCI. If ADX is falling, so fickle market, and it should not be trading for at least an indicator of a trend-following kind of CCI. Come out after the CCI will peak and will move further in the direction of zero. An alternative strategy would be to use the output stops at the last peak or trough after the correction of CCI. Our testing has shown the usefulness of each of these basic approaches.
A few observations Our research has shown that in a general sense CCI is a tool in much the same with the ADX, which can help in assessing the overall market trend. As we stressed earlier, faster than the growth of CCI, in fact there is a strong trend market. Despite the fact that mathematically possible upward movement of CCI during the absence of such a 63
movement of the market, it is unlikely in practice. Remember that the CCI can provide important information trader, even when not giving signals to entry. If the market is most of the time within the range of + / -100, it shows no trend, so you should avoid this market or use a counter trend strategy. We have found that the best markets for trading are those where the CCI recently produced multiple spikes protruding for 100 in the same direction. We also observed that the first trade against the trend set CCI is usually very unprofitable. If the market was able to trend and showed a series of movements CCI on one side of the 100 range, as we have just described, do not expand trade in the direction of the first movement of CCI, which pierces the 100 mark in the opposite direction. A short passage on the opposite side of the range, it's probably possible to add a new position, and not a demonstration of turning the trend.
Avoid jerking We also observed that the technique is often recommended by us wait for an acknowledgment after receiving a trading signal is the exclusive method of eliminating most of the twitching when using CCI with a faster period. We found that when CCI produces spikes after + / -100 level is almost always better to wait for the confirmation signal before taking any action. When the CCI rises above 100, wait for the market to produce a significantly higher close before buying. We noticed that much of the break of 100 was provided only odnodvuhdnevnym event, especially on a short-term periods. Confirmation method of entering it possible to avoid most of the twitching and at the same time caught all the big moves. Technique also allows us to confirm the switch to a faster CCI, which is needed to overcome the problem of delay, do not fall for the twitching, as 64
you would expect. For example, the 10-day CCI demanding confirmation will produce a much more rapid signals and probably jerking produce less than a 20-day CCI, used the normal way. (See Figure 2-24.)
The divergence (Divergence) The divergence is not a separate technical research, as others in this chapter, but we mention the divergence so often, that the detailed description of this concept would be useful. Investigation of the divergence has a long history, dating back at least to 1890, when Charles Dow first to formulate what later became known as the Dow Theory. Dow Theory is based on the confirmations between the major indices Dow to generate signals that detect trends in stock markets. Prior to joining derivatives intermarket communications technology research, research volume and open interest are in fact the only tools for the study of divergence. They are still very popular and are the basis for a huge portion of interesting and practical work (eg, three-dimensional table and the Titanic Bill Ohama for stock markets). Today, the majority of futures traders much more interested in the divergences between technical research and the downstream markets * than the classical divergence of the Dow Theory. It is difficult to give a detailed description of divergence. A possible reason for this may be that these divergences - a relatively subjective types of trading signals. Similarly, the classical charts Edwards and Magee, easy to detect divergence in hindsight, but it is always very difficult to see during deployment. It is also very difficult to efficiently calculate test trading models based on divergences. * The term we understand the downstream market or schedule to which this technical study (indicator). (Prim.perev.) Divergence between technical research and markets
Divergence between technical research and markets Divergence is usually best are identified with some types of oscillators such as Stochastics, RSI or MACD. Let's just say: when the market creates a new peak or trough, and the technical study, built on it, per se does not, then we have a divergence. By definition, this means that as a technical study and the downstream market will exhibit peaks awl or depressions that can be readily recognized. (See Figure 2-25.) In practice it is not always easy. Most large formation of divergence, which correctly showed the turn of the market, will contain a number of smaller divergences
65
which seem to be insignificant for consideration, but eventually look more important. Several small spurious signals often have to correct each of the main signal. The problem, of course, is to recognize that a divergence which carries sense and that which is not. Most technical analysts use some other kind of chart to confirm, such as the classic models or analytics based on multiple divergence, comparing the different technical studies and the belief that the figures they will find safety. We do not believe that there is an advantage in anticipation of the divergence of multiple indicators. Due to the fact that the oscillators are usually similar in their actions, multiple divergence will occur as often as solitary. You can see a divergence in trend-following indicators such as DMI and even moving averages, but we have a feeling that they are not as correct as those that occur when you use oscillators. (See Figure 2-26.)
66
Trending markets in comparison with non-trending One of the ways of the false separation of divergence is to determine whether you are in a trending or non-trending market and the interpretation of the signal accordingly. There are indicators that help measure the trendiness, such as ADX Wilder, linear regression. Even just watching the graph can be estimated trendiness. The main difference between the two types of markets is that non-trending market trades on divergence can be made in any direction, while on the market trend divergence signals against the trend should be generally ignored (with the possible exception when attempting to catch the main peaks and trough). The direction of the trend, under the assumption that the trend can be understood by means of simple indicators, such relatively long-term moving average.
Major trading rules Divergences - extremely dangerous signals, because you usually try to catch the peak or bottom of some type. The most important thing to remember when trading divergences, is that you have to wait until the divergence will be confirmed by a series of closure or closures in the direction of the new trend. Do not jump the gun. Most charts littered the remains of potential divergence that result turn around the trend continuation. If you hurry with the entry, the risk to be on the wrong side of significant market movements, especially if the divergence is at the peak or trough - it puts you on the wrong side of the break. (See Figure 2-27.) Better yields in trading on divergence, assuming that you have correctly entered depend on the type of market in which you are located. In trending markets, if you managed to get on the right side of the trend, your entry is no different (other than what was earlier) from a normal occurrence when you follow the trend. Set your output by using the technique of relatively free of outputs that will keep you on trend, but do not give too 67
much income - such as Wilder's Parabolic Stop or stop tracking a distant relative, or both of them together. If you caught a short-term peak or bottom of a volatile market, use a very close stop watching or trading purpose, or a combination of both.
Serial divergence Divergences often appear in the form of a series of closely spaced on a separate market. Obviously, only the latest in a series of divergence means something, and the longer the series, the stronger the signal. One observation says that the divergence going into three, the so-called A-B-C divergence and "three pointer to the Top" George Lane. Our observation suggests that the double and triple divergence occur in trending markets, while perfectly correct single divergence occur on non-trending markets. The behavior of stock market indices (March-April 1991) is an example of multiple oscillator divergence in a trending market. See Figure 2-28.)
It is difficult if not impossible to know when to take the first divergence as a signal, and when the wait. Many large market reversals in recent years have been predicted by single 68
divergences. Approximately the same number was determined by multiple divergences (see weekly chart of S & P / RSI in Figure 2-29). As you can see, there is often more art than science. Apparently, you need to act with the appearance of the first divergence and put up with losses.
Divergence related markets It is important to be prepared that the divergence between related markets or between the cash market and related futures market as much useful and valid, as well as the divergence between technical research and the downstream markets. As we argued earlier, the Dow theory is based on divergences related markets. Divergence related markets offer excellent entry signals, and should not be ignored. Stock indexes, in particular, often show a divergence near or directly on the peak of the market (see chart divergence Dow / S & P futures in Figure 2-30). Industrial Dow Jones reached new market peaks in early January 1990, while futures on the S & P did not confirm the new market peak (like many other stock indexes).
69
The assumption is that this was caused by the increase in the number of hedgers in the futures market, which lowered prices enough to show the non-confirmation. Be that as it may, failure to reach the peaks in both markets is good reason to believe that the market is getting ready for a correction. A similar phenomenon occurred in August 1990, when the Dow hit a new absolute peak. (See Figure 2-31.) In order to show that this does not necessarily apply to the exchange market, we have led the daily and intraday divergences in the oil market. (See Figures 2-32 and 2-33.) Other divergences offered to your attention - it's Treasuries versus treasury notes (we saw this combination, have been used successfully in day trading), soybeans, compared with soy oil or soy flour and divergence between currencies. The basic principle applied is to trade in the direction of the market, who would not confirm the movement. For example, if the soy flour has created a new cavity and no soybean oil, soybean oil necessary to buy. Or in other words, buy more powerful items to buy and sell signals of weaker signals on
70
sale. A possible exception to this rule would be when one of the contracts do not have sufficient liquidity. Then trade a contract with the highest volume.
71
Plant model George Lane highlights the divergence form, called them a "set of bulls and bears." (See the recommended literature.) These patterns are formed when the oscillator sets a new peak or bottom, and the price is not confirmed. Lane concludes, when Bear installation occurs after an uptrend, following consolidation may occur at an important top. (See Figure 2-34.) Use the reverse logic for bovine installations after a downtrend.
72
One of the subscribers to our leaflet tested this model and came, it seems, to the opposite conclusion. He made the observation that during the installation of the bear is often followed by an explosive breakthrough in the upward direction. We can see some truth in both observations that may lead to a combination of unusually profitable trades. If you buy on consolidation following the installation of a bear, as recommended by a subscriber, you can get very explosive trade on the rise. For this big trade in the direction of follow up summit, which was looking for George Lane, after which you can expect a lucrative trade in the downward direction.
Moment (Momentum) and rate of change (Rate of Change) Regularly communicating with other traders, we found that most of them use the time more any other instruments, except, perhaps, only moving averages. The moment is not always used them as a primary tool, but traders are closely watching him and used in conjunction with other technical studies to achieve a more timely trading decisions. Among the many reasons for the popularity point, you can call its simplicity, versatility, and the fact that it is considered a rare "leading indicator." In addition to simply reacting to the direction of price movement, a moment can change direction before it will make the price. Very few technical research can equip trader such a valuable leading indicator. Because of the versatility of which we spoke, the moment it is difficult to attribute to follow the trend or counter trend indicators. It can be used to demonstrate the direction of the trend, and may also give a very good warning repurchase / resale, making it a useful tool to counter trend trade. This seemingly simple indicator actually contains much more information than what is immediately obvious. The wealth of information hidden in the calculations, opens many possible uses of the moment. A full understanding of what we calculate is to help exploit the full potential of the moment. Moment Indicator provides us with an accurate measure of the speed of the market, and to some degree the extent to which the trend is still full. The calculation is simple: subtract the closing price n days ago from today's closing price. The result will be positive or negative, oscillating about a zero point or line. The formula is as follows:
73
where M - the moment, P t - today's closing price, and P tn - the closing price for n periods (usually n days) to Pt. The value of n - is the only part of the formula, which can be changed by the trader, and is most often used here is 10. Some software packages allow the user to select for periods of price discovery, the peak depression, close, and some other value prices. We see no reason to use anything other than the closure of the calculations. The result is a technical computing study which oscillates about the zero line (making it an oscillator). If the market moves up, the zero point intersects the line upwards, and in general, retain the slope upward. If the market moves down oscillator zero crossing points downwards, and in general, retain the slope downward. This all sounds simple enough, but the oscillator has more time or other and more complex properties. For example, the further away in the low P t and P tn, the greater the distance between the values of torque. When the market is moving rapidly in the upstream direction (we assume that the market is bullish), so does the time and. But when the market is close to its peak and closing prices are closer to each other, the time slows down considerably and the time line is horizontal or tilted down, despite the fact that prices may continue to rise. When the market peaked and there are negative values P t - P tn the time line will dive for the zero line. The moment of clear signals that the market rate is slowing. Formula measuring points not only speed, but also the speed with which the movement is slowed down. She describes how the speed of the market and the rate of change of speed when the market is close to the top or past its peak. With a further deviation of the market, become dominant negative moment, and his line will approach and cross from time to time zero line, signaling a change of trend from bullish to bearish. (See Figure 2-35.)
74
What makes time to respond in this way? In order to increase the value of the moment, and its direction was upward, the recent price value must exceed the older ones. If recent price values are the same as the older, since the line is flat, while the market is still moving upwards. If recent price less than the old, even if the prices are still rising, the pace of change will continue to slow down, and the moment will fall. Flattening and subsequent deflection line down ahead of time shows us that the normal price chart can not show. Moment gives us early warning of slowing the rate of the market and that the rate of increase in prices slowed down now. (See Figure 2-36.)
The rate of change (ROC - Rate of Change) Here we very briefly consider the rate of change, as most of today's software packages predostavlyaeet this indicator in addition to the time, but despite this they are essentially identical. The rate of change has the following formula: ROC = 100 (P t / P t-n) Level 100 is equivalent to the zero line chart of the moment. The only possible difference or advantage, which can be seen here, is that when you use the ROC instead of the moment, do not have to deal with negative numbers. Trade rules and practical application are the same for both indicators.
Torque signal - trend following When being used as an indicator of a trend-following, his most important signals are at the intersections of the zero line. When the line is crossed from the bottom up, the moment 75
is bullish. When the line is crossed from top to bottom, the moment is bearish. We would not recommend entering a position opposite to the direction of the moment. The number of zero-crossing point of the line depends on the time period, which was used to calculate the moment. Like other indicators shorter time values lead to that point will be faster and better respond zero crossing line. Longer values generally slow time signals, reducing the frequency of intersections. The smoothing effect of longer periods, obviously, is not the result of averaging more data because the formula does not provide an average of the closing prices. The simple logic is that if there is a trend longer it takes to return to the price set 40 days ago, rather than a return to the price of the 10-day statute of limitations. We know traders successfully use a wide range of time periods from 10 to 40 days. Many traders working with loops, try to connect with a long period of time in the market cycle. Since the extension of the period of time will make the oscillator is less responsive, and the shortening can lead to twitching, some traders find it helpful to use a relatively short time and sensitive values, and then set boundaries above and below the zero line. They then use the border crossing instead of crossing the zero line as a signal to the new trading. When the time ranges within the boundaries, this is not a signal for new auctions. This leads to the fact that the market is forced to "prove" their
movement before getting into position, all it eliminates a lot of twitching caused by the frequent crossing of the zero line. (See Figure 2-37.) Keep in mind that the most significant gains can be obtained, and when the time and the prices are accelerating. As we have described, the slope of the time will be reduced when the rate of price growth will slow down. Clear and effective use of time would be not entering into new trades as long as the time line does not bend in the direction of the trend. When the moment is moving back towards the zero line, the trend is by definition weakens or disappears, so that trade in this area could prove to be useless. (See Figure 238.) 76
Torque signal - following a counter-trend Since the torque measuring acceleration or deceleration of the market, it is very useful as a repurchase indicator / resale. When the market reaches the top, time aligned and often begins to fall well before the peak of the real market. A similar discrepancy in the directions will appear on the market and depressions. Assuming no significant change in the volatility of the market, the line is drawn on the long-term chart, connecting the peaks of the time, parallel to the zero line and the line connecting the trough point, also parallel the zero line will represent the zone repurchase / resale.
77
The main trading strategy here is selling right at the break of the upper zone, with the placement of protective stop above the recent peaks, and purchase immediately at the break of the lower zone, with the placement of a stop under the recent depressions. Income can be obtained when the opposite zone. (See Figure 2-39.) This counter trend strategy will be productive if the recent market action is taking place in a certain price range, but if the market will make a significant breakthrough, it obviously fails. We have seen a formula that incorporates an attempt to cope with this problem by normalizing the time so that it is always fluctuating in the range -1 to +1 or -100 to +100. This can be done by dividing the values of the time on a certain constant value. We do not see much value in this approach. The normalization of the values of the oscillator will not prevent the market to make a breakthrough, if there are conditions for this. Normalized time to act is largely similar to the RSI or some similar indicator in a trending market. The values will be grouped at the top or bottom of the scale and provide continuous signals to buy or sell. Standard unnormalized moment will continue to rise or fall on the theoretically infinite levels, confirming the continuation of the trend, and encouraging the trader does not use a counter trend strategy.
Long-term trade with the moment One of the most productive uses of time is to determine the long-term trend, and when the trend is defined, the auction must be carried out in this direction. This rule should 78
significantly increase profitability by sieving unprofitable trades going against the trend. Moment not only tells you about the direction of the trend, but also gives an idea of its strength. This valuable information will insure you out of trouble. (See Figure 2-40.) Our study shows that in most markets, 25-period time, based on the weekly chart is surprisingly reliable indicator of long-term trend. Trend trading is particularly advantageous when the line is being rapidly removed from the zero line. However, be very careful to follow the trend, when the time reaches a peak, and the time when the line is deflected back towards the zero line. The logical combination of technical research in this case would use the time to find a long-term trend, the medium-term moving average to enter the auction, when the moment is strong, and more short-term counter trend indicators such as Stochastics or RSI, to generate income when the time is reduced. Meyers and Colby in his book, "The Encyclopedia of Technical Market Indicators" (see recommended in literature end section) in one of several test time optimized rate of change of about 20 years data NYSE (rate of change of the indicator is substantially identical at time .) Their trade rules were simple: buy when the indicator upward cross the zero line, and sell when it crosses her down. Holding the position after the first crossing, passing the peak, and the closing only after crossing the opposite may be of academic interest, but to us it seems to disregard the basic properties of the moment (or ROC). Trade with the moment as a turning method ignores the fact that the deceleration time is a signal to the exit from the market, or at least for the transition to a method of trading that differs from that which would have applied if the market continues to move up. Not surprisingly, the results have been disappointing total income, and losses quite severe. One very simple test of time was also conducted by Bruce Babcock and described in his book "The Dow Jones - Irwin Guide to Trading Systems". He tested the 10-day and 28day time using a simple method of turning the intersection without stopping. The results were breakeven, which is encouraging, given that neither by time nor by any other oscillator should not be trading in this way. Check out the testing system (Chapter 3), where we show similar results of our own testing time.
79
Trade with divergence time We have always observed that the divergence between technical research and the downstream market, or between related markets often provides effective trading signals. The divergence between the oscillator such as the time and the downstream market may arise when the market and the moment of creating a high peak, then both depart, and then creates a new market peak, which is not supported by the new oscillator peak torque. The theory essentially consists in the fact that the divergence indicates a weak market support and that he will not be able to continue climbing after a new peak. Divergence of price and time are quite diverse, 10-period time, based on the daily charts reveal plenty of divergence and a number of significant commercial opportunities, especially in the case where a longer 25-week point in a phase of his fall. Our standard caveat concerning trade with the divergence, said: Wait for the divergence is fully confirmed before entering the market. Early entry can be reasonably expected to leave you on the wrong side of the trend of the market (see Figure 2-41).
The use of the time of other indicators Many technical studies, which we mentioned in the book, is measured in some way by the market. This is usually expressed as slope of the line obtained in the calculation of the study. For example, a moving average, which is in a strong trend, usually a sign of an underlying strong trend of the market. The steeper the slope, the stronger the trend. The exact determination of the degree of force can be very subjective, if we consider only the technical study, however, if we take the time or the rate of change of the indicator, we can objectively calculate the strength of the trend. This opens up new possibilities. We can
80
filter slabotrendovye markets and focus our efforts in markets with unusually strong trends. Or, if the market is not in the trend, we can buy on dips and sell on the climbs. We believe that the moment has a lot of standing and its application can be considered as one of the most useful technical studies available trader. Has a rich imagination and inventive technical analyst should find plenty of interesting applications for this indicator, which is ahead of price rather than follow them.
Moving averages (Moving Averages) Probably the most real money is earned today using moving averages, rather than with all other technical indicators combined. Because they can be used for any purpose, such as finding a long-term monthly trend setting stops for day trading as well as for many other things, moving averages, were the subject of more debate in the technical literature 81
and other sources than any other technical research. One reason for the acquisition of such popularity is that when markets are trending these simple small lines operate worse or even better than LEDs that require for their doctorate calculation and interpretation. Moving averages smooth out the fluctuations of the market and short-term volatility, allowing the trader to understand in which direction the market moves. It is equally important to know that what they do not. As long as you do not draw a them as an oscillator, they will not provide absolutely no information on the repurchase / resale. They are indicators of the trend-following in its purest form. They always show the direction of the trend, but do not measure how strong or weak trend. Their purpose - to determine the direction of the trend, and then flatten or suppress its volatility. Moving averages cope with these important tasks simply and well. There are so many various types and variations of moving averages, it is pointless to try to list them all. Most of the species were created in the 1970s, when the moving averages were considered very tricky and advanced technical analysis tools. Many talented and resourceful technical analysts have spent most of their time, bringing new ways of using and improving moving averages. The interest in them has been amply rewarded: 70s were a time of markets that are in a constant state of trend and moving averages have worked exceptionally well. Most of the more sophisticated types remained unclaimed until (such as "modified accumulate" or "srednemodifitsirovanny" Maxwell). The three main categories withstood the test of time: simple, weighted and exponential.
Simple Moving Average The simple moving average is calculated by adding and averaging a set of numbers representing the action of the market for a certain period of time. Calculation typically includes closing price, but can also be calculated from the peaks or troughs average of all three. The oldest data point is dropped with the advent of a new, consequently, the average "slips" and follows the market. The line connecting the daily average will smooth the recent market fluctuations.
82
Set representing a large number of historical data and create a smooth line. For example, the chart shows the 50-day moving average. (See Figure 2-42.) As you can see, most of the time prices are on one side or the other of the moving average. Also, when a trend develops, acquires the slope of the moving average in the direction of the trend, showing us his strength. The longer-term moving average to smooth out all minor fluctuations, and show only the longer-term trends. Short-term moving averages show a more short-term trends in the long-term damage. Reduced data set representing only more recent data, will create a more sensitive line. A graph showing the 5-day moving average, covers the same schedule for the 50-day moving average, a 5-day reflects data
83
much more accurate, following every little price fluctuation. Short-term trends are easy to see, and trends that have become apparent through the 50-day moving average is much more difficult to determine. (See Figure 2-43.) Long-term and short-term moving averages have each been applied and disadvantages. Despite the fact that the 50-day moving average is a trend, it always remains at a distance from the actual price and changes direction much less than the price. In practice, based on the length of a moving average trading system will slowly move in and out of the market. Slow entry misses a substantial part of the beginning of a trend, and slow output donates most of the income. On the other hand, the 5-day moving average comes in and out quickly, but not in harmony with the main trend and also often on the wrong side of the market, as well as on the right. Another interesting property of the simple moving average (and many other technical studies of this type) is that they also act old price ejected from the averaging as new. Unexpected turn of the moving average may mean that prices have turned fresh. It could also mean that the fresh prices do relatively neutral, but substantial costs have been thrown from the other end of the data. This is not necessarily a bad thing. After all, the purpose of the moving average is to smooth the data. But to this effect should be prepared. This phenomenon can sometimes explain what seems inexplicable change in the moving average or other indicator.
Weighted averages
84
Simple Moving Average assigns equal weight to each price used in the data series. Some traders, believing that the recent price is more important than the older (and probably with a view to partially overcome the problem with the data described above), prefer to create moving averages, which are quick to react to the latest data, and slowly on the old ones. Weighted averages assign more importance over recent data by giving different weights prices every day. This is typically done by multiplying the latest data on some predetermined number (e.g., number of data points used in the moving average), adding to the overall computation result, and then multiplying these less fresh minimal number of data and so on. Line derived from it would be better to respond to the recent market activity than the simple moving average.
Exponential average skolzyashie Simple and weighted averages can respond only to a specific range of data selected for evaluation. The exponential moving average gives greater weight to the latest market action as well as a weighted moving average. However, the exponential moving average continues to recognize all of the data points that do not discarding the 5-day exponential moving average usually includes more than 5 days of data and can include data in the history of the futures contract. In fact, these moving average can be better identified by their true "smoothing constants" because the number of days of data in the same computing the so-called five-day average and 10-day average. The exponential calculation may have undesirable property exhibited by the difference between the moving averages, depending on the choice of the starting point, 5-day exponential moving average A trader may be different from the same in the trader if they started their calculations on different dates. For practical purposes, these two values are likely to be fairly close to the same time to cross the 20-day moving average, but this is not certain. Since our objective is to describe the practical application of indicators, not their calculations we omit the formula. Details of exponential calculations are quite numerous and well documented in previous studies, to which we have referred. (See Figure 2-44)
85
Despite the apparent sophistication of the weighted and exponential moving averages, almost every test that we have seen or held their own. showed the superiority of a simple moving average over the rest in terms of trading results. Our own research shows that the weighting of the data to emphasize the light of recent events makes overly sensitive, thus reducing the primary purpose is not to how high-zyaschey - smooth out the market action. Weighted and exponential moving averages generate more trading in the narrow, being in a trading range markets than simple moving averages. The result is usually costly twitching. This is to confirm the theory, which we have long held: any method of entry, which is the result of obscure calculations, is more negative things than positive. Futures trading is more art than science, and mathematical sophistication does not guarantee the profitability of the method. Despite the fact that these calculations are performed by simply pressing a computer keyboard, we recommend using only simple moving averages. Saved the complexity of your system for a more scientific applications, such as cash management and risk control. Trading systems moving average can be used as a single moving average, and any number of moving averages in various combinations. We used the trading systems of one, two, three or even four moving averages. Perhaps it can be more, but even with variations only three or four can already just be too complex, and as you've probably noticed, we do not see any advantage in using anything was more complicated than necessary.
A moving average system The simplest and often the most effective moving average is a single moving average. It is more useful as an indicator of a long trend than as a tool for day trading. For example, 86
Colby and Meyers in his book, "The Encyclopedia of Technical Market Indicators" single optimized moving average on 75 years of data NYSE, using a simple rotary system. They found that 12 months is the optimal number, significantly superior strategy of "buy and hold". According to our experience, this simple 12-month moving average is the best tool for the job time for the stock market. (See Figure 2-45.)
The basic rules of trade with single moving average is simple: buy when the prices (normally closed) rise above the moving average, sell when prices fall below a moving average. The result is a simple reverse system, which is always present in the market. We do not recommend this system of trade. No matter what you choose moving average, long-term use will be periods and periods of loss of income, and the overall result will fluctuate around zero minus transaction costs. It's probably best to use a single moving average as a filter trends. If prices are higher than average, trade only on the long side of the market, using some other more sensitive method for the determination of entries and exits. If prices are below average, trade only on the short side. (See Figure 2-46.)
87
Dual Moving Averages The most popular system uses two moving averages moving averages. They usually consist of a continuous medium that is used to determine a trend or more short average, which provides trading signals at the intersection with a long-term average. The most famous of these systems - 5-dnevnaya/20-dnevnaya Richard Donchian system, which, by the way, is not a simple reverse of the system, and uses an elaborate set of filters. (See Figure 2-47.)
88
The main signal is a double moving average crossing. Buy when the shorter average crosses above the longer, and sell when the opposite situation occurs. Can also be used as a crossing point of a trend reversal and trade only in the direction indicated by the trend, using other shorter-term methods for entries and exits. Most of the seen and our studies have shown that the system of double moving average, tend to be more profitable than other combinations of moving averages. The study also shows that all of the moving averages have long periods of gains and losses, depending on the market trend. In general, the system of moving averages are notorious for a habit to give away too much of the hard-earned income. Anyone who has traded on the system Donchian trend during the 70's had a regular and substantial income, due to the strong trends that had prevailed at the time. The same system suffered heavy losses in the mid to late 80's.
Triple Moving Average The most popular triple moving average is widely used 4-9-18-day method popularized by RK Allen in the early '70s. The third moving average opens up a large number of potential trading opportunities. In general, when the market hit bottom, the main evidence of a change in trend is the intersection of the 4-day 18-day. Confirmation tone - the intersection with the 9-day 18-day. When prices are at a peak, advance signal a possible change in trend is the intersection of the 4-day and 9-day. Generating revenue at this point will help to overcome the characteristic feature for systems moving averages expressed in the return of income. Reversal of the trend will end only when the 4 - and 9-day cross the 18-day. We like the triple moving average system, because they provide the advantage of a neutral zone, as opposed to continuous reverse trade generated by means of single or double moving averages. For example, on 04/09/18, when four crosses 9, we go out of our position and do not enter into a new, yet not cross the 18 9. We also like the ternary 89
system because the intersection of 4 and 9 is a mechanism for rapid revenue that solves some of the problems associated with too much of a return of income, which we mentioned earlier. We believe that a good trading system outputs must always be faster entries. The entries must be slow and selective, may require an extraordinary event for the entries in the trade. The outputs must be slow enough to allow the flow of income, but not fast enough to ultimately secure the bulk of potential profits. (See Figure 2-48.)
Four moving averages The use of four moving averages is not so strange and not as difficult as it seems. When used properly, the approach with four moving averages can bypass some of the problems associated with moving averages, without losing any of the advantages. The method uses four moving averages in sets of two. Two of the longer moving averages are used strictly as an identifier, and the trend is most easily applied when installed as oscillators. Two shorter moving averages are more sensitive and are used to set the time entries and exits (usually on the basis of the intersection), trading exclusively in the direction signaled by the long-term oscillator. Trading against the trend of dropping out, by definition. If there is an uptrend defined by long-term oscillator, the signal short-term intersections will be applied only long trades. On the other hand, will be only a short trading in a downtrend. Will occur during periods of neutral trend correction and lateral movements of the market, when short-term and long-
90
term moving averages are not able to confirm the direction. Twitching not be completely destroyed, but their number is significantly reduced. (See Figure 2-49.)
The displaced moving average (DMA - Displaced Moving Averages) One of the problems of the moving averages, as well as for any other trend-following studies of this kind, is their inability to snuggle close to trend, while retaining the desired smoothing effect. As we said, the slow moving averages smooth out fine, but always at a distance from the market at the time, as a fast moving average overly sensitive. The displaced moving averages help to correct this defect. They are created by conventional computing the moving average and projecting it into the future by means of displacement for a given number of days. As you can see in the accompanying graph (see Figure 2-50), the effect is to shift the moving average forward in time. To see the look of the displaced moving average, if your software does not have these features (DMA has long been known, Nona too common), draw two graphs: one with prices, and the other with the moving average, then put one on the other and view them on clearance. Slide the moving average to the right. You will see the result of the displacement moving average forward in time.
91
In the most popular use of DMA shifted forward by the same number of periods, which is used for its calculation, 3x3 DMA - it trehperi-odnaya Moving Average, moved forward by three days. (See Figure 2-51.)
92
Some traders prefer to reduce the time period of displacement relative to the moving average, 10x5 DMA - that's 10-day moving average, offset by 5 days. (See Figure 2-52.)
The most common use of DMA that we met is to use them as a short-term trend indicator. Joe DiNapoli, for example, is trading at spreads within the trend defined by the DMA. (See Figure 2-53.) Other day traders prefer to use them to make
93
solutions, which side of the market to adhere to during the day. For example, in futures trading on the S & P can only go in the direction of DMA, calculated at 30 or 60-minute periods, using a different, more sensitive method for the actual entry and exit from the market. (See Figure 2-54.)
94
Finding the filter Instead of blindly following the intersections of all, many traders use a variety of filters to determine the suitability of the primary signal. Filters are divided into two categories: pricing filters and temporal filters. Filtration signals Stores usually means pulling entering trade until the price satisfy some additional criteria. This can be determined by measuring the breakthrough for the moving average or the distance measurement, which was one of the other moving average after crossing. In this case, the trader is looking for confirmation that the moving average price was not accidental event, but is in fact a change in the trend. The new trade does not begin until as long as the price will not exceed the moving average is a minimum value. Another variant of this filter will be waiting when prices will rise by a percentage in relation to the moving average. The next option (which we find common) is waiting a predetermined period after crossing until the market reaches a new peak or trough in the last n days, which is a breakthrough channel. (See Figure 2-55.) One of our favorite filters or methods of verification is very simple: wait until the closing direction of the new trend.
95
Temporary filters use a certain amount of waiting is the time period after the intersection before entering the trade in a new direction. Many traders use moving averages have noticed that most of the jerking occurs very close to the beginning of a trend, and a slight delay entry will avoid most of them. The waiting period is typically one to five days. If the price remains on the new side of the moving average for a minimum time period, we conclude that the signal is correct. Obviously, the more the waiting period, the less twitching, but at the same time it may lead to late entry so that the main part of the movement will be missed. (See Figure 2-56.)
What medium to use? There is no right answer as to which combination of moving averages works best. We once saw on the computer calculated a large matrix containing the results year after year moving average crossover from 1 to 100, going back in time 15 years. The conclusion of this study was on a portfolio of products that work consistently moving averages, if you know in advance what specific combination to use for each product in each year. The most extensive known to us the published test results were presented by Frank Hochheymerom in the early 80's at Merrill Lynch. We ourselves have done considerable work in this area and have tested hundreds of thousands of possible combinations of moving averages. We believe that there is no magic answer. In virtually every case, the values of moving averages, worked well on past data, did not give good results in real trading. The method of testing does not really matter. However, in our testing, and in other works, where data was available real trading, a phenomenon repeated. However, it is quite obvious that almost any combination of moving averages is profitable in a trending market, and virtually non-existent combinations, 96
income on non-trending market. Thus, the solution lies not in finding the ideal combination of moving averages. The answer lies in finding a reliable system that will provide the markets in which the moving averages are generally profitable. Then it makes sense to trade in these markets manner, calculated to keep most of the revenue with minimal losses. Non-trending markets, we stress should be avoided or trade them with counter trend indicator type.
How do I get the system moving averages? Moving averages are the simplest and most elegant of the available research, following the trend. To a certain extent they can be very effective, but their restrictions can be substantial. Most markets spend most of their time in a sideways motion, rather than in a trending. Non-trending market can knock picked up in the most accurate way system moving averages. Here are some of our thoughts and conclusions about how to help the system survive moving averages. Try to limit your trade only trending markets. Diversification helps, but do not trade the same way in all markets. At any particular time, typically less than 50 percent of the market may be defined as trend. Most of the time their actual number is much smaller. Find a way to objectively determine the market trend and then use moving averages. We recommend DMI / ADX Uaildera as a reliable study, which measures how the market is directed or nenapravlen. The simple explanation is that when the ADX is rising, the market is in a trend, and when the drops - the market is losing focus. filter channel breakout, which we mentioned earlier, can also be effective. Long-term moving averages are basically reacting too slowly to be useful for the exits. Use an alternate exit strategy. We think that the most common mistake when working with moving averages is to apply one set of moving averages for the inputs and outputs. If you are using a slow medium, you will slowly go out and lose a large part of the income. If you use the faster moving average, you will have better yields, but you will find that you get when entering twitching. Even something as simple as tracking the dollar stop, usually better than the yield on the moving average with a possible loss of revenue.
Trading method convergence - divergence of moving averages (MACD Moving Average Convergence-Divergence) Trading method convergence-divergence of moving averages (abbreviated MACD) was developed in 1979 by Gerald Appel as a tool to set the time for the stock market. The fact that many traders use MACD exclusively to trade futures and options on stock indices, does not mean that it can not be successfully applied to other contracts. As you will see, there are some excellent examples of effective work MACD on various non-financial markets. Among other things, the price behavior is behavior rates regardless of which markets it occurs. We prefer to share technical studies into two categories: those that are effective in trending markets, and the ones that work best on an unstable or non-trending markets. MACD is best used as a study, following the trend. MACD trading method works 97
particularly well on long calm markets, where you can stay with the main trend, ignoring the weaker price movements. One of the best uses MACD will use it on weekly or monthly charts as an indicator of long-term direction of the market. Normally the use of MACD on non-trending markets do not lead to success. Look for divergence when the markets are not in trend.
A brief overview of the basics MACD is a combination of three exponentially smoothed moving averages, which are represented by two lines. The first line shows the difference between the 12-period exponential moving average and the 26-period exponential moving average. The second line (called the signal line) is a rough equivalent of the exponential 9-period moving average of the first line. Exponential values are used 0.15,0.075 and 0.20. MACD is usually displayed as a line or bar chart of the oscillator. Most software allows users to change the value in the calculation of MACD. Some systems require you to enter the exponential value, while others use the available number of periods for the three moving averages. We think that you should not try to change the initial values of the MACD to fit the curve of the data. It is important to note that the recommended Appel two different sets of values, one for buying the market, and the other for selling party. Both use a 9-peri-odnuyu (0.20) the signal line, but the combination 0.15,0.075 recommended only for the sell side. Buy-side values are 8-periods (0.22) and 17 periods (0.11). (See Figure 2-57.)
Using the formula of buying and selling formula may require a completely different mindset from casual users MACD. Always good to stick to the original, but you need to know about the mindset of the developer, when you use any technical study. If the default settings for the side of your software sales are set to common values, or if the software does not allow to change these values, you may find that you use the MACD is not, as suggested by its developer. Ideally, your computer monitor should be set up to display the futures price curve and two additional graphs, one for demonstration MACD buy-side and the other for the sell-side formula. You will find that buying formula a little faster and a little prone to twitching. Formula slower sales. It seems that the idea was to quickly buy and try to keep the position in order to allow income to flow. We think that could be applied to the construction of buy-side for both formulas, provided that it is inherent understanding of 98
their initial appointment. For markets other than the stock market, you can follow the standard formula sales as long as you do not need a faster and less reliable signal.
Trading on the intersections MACD The main signal MACD is crossing. Buy signals are generated when the faster line crosses above the slower and sell signals - in the opposite case. We want to warn you right away: in most markets, mechanical trading at each crossing MACD will result in frequent twitching and significant losses. You will quickly find that the narrow trading ranges destructive effect on the indicator, which gives a lot of false signals and collecting losses. Fortunately, there are additional interpretation MACD, which will help the trader to avoid jerking and other shortcomings. (See Figure 2-58.)
Using levels repurchase / resale MACD method can be used to determine the points at which the market is overbought or oversold and thus is subject to reversal. Considering the faster of the two lines MACD, Appel set of repurchase / resale for the S & P at the level of + / -2.50 on a scale MACD. For the NYSE index, it recommends + / -1.20. When the other line has reached these critical areas, the cross-generates a signal to buy or sell. Intersections which occur until it reaches an extremum may be ignored, and thus will remove most twitching. As a result, small study, similar levels could be detected for any market, located in a wide trading range with large price movements. Remember, MACD is best used as a long-term trendfollowing tool, rather than short-term trading timer. Signals arising in the middle area of the chart MACD, should be taken only if the other proven indicator confirms that the trade will take place in the direction of the trend. (See Figure 2-59.)
99
MACD trend line One very interesting way to use MACD is the mark of a jump before crossing by drawing a trend line at the MACD and then trade on the break of the trend line instead of waiting for the intersection. MACD trend line break may be preceded by an important change in the market and is an early warning signal of the rotation on the market. Intersections MACD, preceded by or consistent with a trend line break, have a much more important technical value than the MACD crossing themselves. Aggressive traders might consider entering the market immediately after the breakout of the trend line in anticipation of crossing MACD, while more cautious traders can wait for the real intersection to confirm. Remember, if you are trading, relying solely on the trendline breakouts, without waiting for the crossing, trade little justify itself if the intersection is not going to happen in the near future. You can get to the unfortunate situation, using a system MACD, and will be forced to find some other method to exit the market. (See Figure 2-60.)
100
Look for divergence From our point of view, the most valuable and revealing models are MACD divergence, especially in netrendovyts markets. There duplex structure which arise when the market takes a peak or trough, then successively a new peak or trough which can not confirm MACD. (See Figure 2-60.) Divergence - this is a very effective form of most technological research, and MACD is no exception. MACD divergence is more useful for indicating the continuation of the trend after the correction, rather than as a turn signal a trend. As with any other divergences, it is best not to jump the gun and wait before entering the market, while the divergence is fully seated. Nothing will make you feel so much like a fool, like getting into the trade in anticipation of divergence, which will never happen. Then you definitely will be on the opposite side of the strong trend. (See Figure 2-61.)
101
Combining signals The divergence arising in accordance with the break of the trend line and MACD crossing is a very strong signal. Typically, the trend line break occurs first, when the market begins to come out of the moment. The divergence comes next, usually coinciding with the double peak or trough in the market. Finally, the intersection of the market by the signals on the entry. Remember that by themselves crossing MACD is usually not significant, until they are joined by one or two schemes we have described. As we warned earlier, no need to 102
mechanically trade at intersections and wait while profits. Our testing has shown that, when MACD sell so get your break-even system minus the cost of transactions. These results are similar srezultatami testing when other technical studies used to trade in the same way. A similar approach should be taken with a grain of salt, even though you may have read about it. (See Figure 2 62.)
Trading in the direction of the trend The main trends can be identified by the use of longer-term MACD to confirm a shortterm MACD. Use the weekly MACD to confirm the daily charts MACD. Use daily MACD to confirm the intraday charts. Trade more than a short-term signals only in the direction of the longer-term MACD, and you will avoid at least half of potential twitching. Another approach is to use other technical studies, such as the medium-term simple moving average single or double simple moving average to find the intermediate trend. Then take the MACD only in the direction of the trend. (See Figure 2-63.)
Whichever method of trend you choose, it must determine not only the direction of the trend, but also its strength. The absence or the presence of a trend in the market can be measured by the ADX Wilder. Use the 18-day ADX, and if the ADX is rising, so the trend is strong and the MACD will give a good return. If the ADX falls, the market lacks the forces and conventional MACD traders are likely to suffer losses, but you have to look for divergence. This will help filter out a lot of non-trending markets, causing losses when trading MACD. (See Figure 2-64.)
103
Parabolic SAR (Parabolic) Parabolic systems cost / time are the technical research, trend-following, trying to overcome two problems common to most trend-following trading systems: the return of income due to the delay signal to the output and the inability to schedule at the time as a factor to calculate the stopping points. The parabolic formula solves the problem by narrowing the price gap with increasing speed stops when it reaches a new peak or trough. Parabolic formula also includes calculating the factor of time, allowing the stops to be removed for a short period and then bringing them inexorably regardless of price action. The result is a function of time is that prices should continue to move in the direction of the trend, otherwise the sale will be stopped. We believe that the parabolic system is a perfect technical tool when used only for outputs. We do not recommend using it for occurrences or as the back of the system, as the developer intended it.
The origin of the parabolic system The parabolic formula was first described by Welles Wilder in 1978 in his book "New Concepts in Technical Trading Systems" (see recommended literature). Wilder was looking for a system that could capture most of the gain in a trending market, without relying on any external methods of retention of income. Parabolic calculations result in a series of tracking stops that in case of activation signal a trend reversal. Stops are recalculated daily (or you're using for each time period) and become closer in the process
104
of promoting the trend. If the trend could continue, sliding stop will replace stand and begin a new time period. In the case of self-use. The parabolic system is the reverse of a system that is always in search of the market trend. We do not consider the parabolic system with one of the best trend-following technical studies, able to work independently. However, in combination with other indicators, it can be extremely effective. We believe that the greatest value Parabolic has when used as a method of placing stops. For the most effective use of it will be helpful to explain the nature of the various elements constituting the parabolic systems. As we said, Parabolic Wilder conceived as the reverse. Wilder
is the point at which the system is deployed, "stop and upheaval" (SAR - stop and reverse). As you can see in Figure 2-65, a series of dots SAR form a line similar to the trend line, but gaining the shape of a parabola, so that the point of the stops are close to the market. To calculate the SAR you first have to select a certain starting point. Wilder recommended to go back a few weeks on the chart and find a significant peak or trough to start computing. Most Computer Science will start on the left side of the screen. If the first few days have vvepx direction, the formula will suggest an upward trend. If the first few days are down, the formula will suggest a downward trend. For practical use, it does not matter which direction starts parabolic system, because in the end it will be on the side of 105
the trend. We recommend users the software variable width window to make sure that the system comprises a parabolic window of at least 100 data points. Without these minimum data point of the first SAR can identify abnormal trends. When the first point of entry and the first SAR established the formula for subsequent SAR following: SAR (tomorrow) = SAR (today) + AF * (EP - SAR (Today)). AF - it is a factor of acceleration (acceleration factor), and EP - extreme point (peak or trough) of the previous trade (EP - extreme point). Note that the price of the previous extremum and the acceleration factor are used together to keep the SAR points close to the trend. Price of the previous extreme EP is understandable. AF, however, is what makes the system unique Parabolic. AF - it is a weighted factor. Wilder used the initial value of AF, equal to 0.02. Then 0.02 AF increases each time when the price of creating a new extremum, resulting points on the acceleration schedule. AF does not increase until the new EP will be made, and it never rises above 0.20. Thus, the range of the acceleration factor of 0.02 to 0.20 in increments of 0.02. This is the default for most software packages, but sometimes they can be installed by the user.
Acceleration change Changing values manifest in AF zoom out or stops SAR, thus making the system more or less sensitive to market movements. If AF increases approaching and stopping the system becomes more sensitive. If AF is reduced, is cleared and the system is slower. The following graphs allow us to compare the values of AF, from 0.01 in steps of 0.01, and the values of AF, from 0.03 in steps of 0.03. As can be seen from these graphs (see Figure 266), the difference is striking. Almost always possible to find a set of initial values and step values for the parabolic system (and any research) that will show revenue when tested on historical data. We recommend using the default settings, the default setting. Try to avoid fitting the indicator under the curve data. For those who are interested, Colby and Meyers in "The Encyclopedia of Technical Market Indicators" Parabolic tested the system on 18 years of weekly
106
Data New York Composite, using it only as a recirculation system. She showed the average profit with standard values of AF and, of course, very high profit after optimization. They received the test results, raising the AF from 0.02 to 0.20. We rather believe the test results, use the default values, and skeptical of this kind of optimization. For example, if using Colby and Meyers data 15 instead of 18, practically certain that their optimum AF would be different. If they tested each year of their data series individually, AF for each successive year would be different from the previous one. You change the value of each year? Every month? We think that is much better to find a reasonable value, unchanged by time performance, instead of chasing the perfect embodiment of the idea of the incredible value for tomorrow. Wilder about the values of AF Wilder in his book (see the recommended literature) made the following important observation: "I've tried a lot of different factors acceleration and found that the consistent increase of 0.02 works best, however, if you want to give the system an individual character to change the points of stops, possibly used by other traders who use a range of incremental increase between 0.018 and 0.021. Any incremental increase in this range will work well. " It seems that Wilder was worried about too many stops at the same point of the market, and we share that concern. Some versions of acceleration can serve for optimization purposes, and for giving your stops differ from those used by the crowd. Remember, the parabolic system - well-known and popular study, perhaps significantly more popular than I expected Wilder, when he proposed to individualize the formula.
Trading on the parabolic system Despite the fact that the parabolic system solves one of the major concerns of the majority of trend-following indicators by placing the plants closer to the market, it still crashes when the market becomes unstable. Since the non-trending markets are more common than the trend. Parabolic becomes unprofitable in most market conditions. Need a filter that will reduce the occurrence in volatile markets and a timer entries, which will parabolic system to do what it does best - ask a stop in trending markets. Wilder understood the limitations of parabolic system and offered to share it with the Directional Movement Index (DMI) or the index of the selection of goods (CSI - commodity selection index), though he did not give specific recipes or rules.
Directional parabolic system Kaufman A good example of how to combine the parabolic system and the directional movement index, was cited Perry Kaufman in his book "The New Commodity Trading Systems and Methods". He combined two techniques in the system, which he called "directional parabolic." Here are summarized his rules: 1. Use the 14-day DMI. 2. If DMI signals an upward movement, take only long Parabolic trading. If DMI falls, take only short trades. 3. Enter the auction only if the systems are consistent. If they conflict, trade is not made. 107
4. Use the Parabolic stop just to get out, and not as the reverse. 1.
Exit the trade if the ADX rose above the + DI and-DI and then turns around. This suggests a weakening trend. (See Figure 2-67.)
Another parabolic trading system Here is an example of a simple trading method for parabolic system, which includes the ADX and helps to avoid many of the twitching. Method for the most part the same as in Kaufman, but it is simpler and relies on the close stop Parabolic outputs. It reflects the three most important principles that must be followed. At trade: trade with the trend, and check your risk, let the revenue flow. That's trading rules: 1. Use the 18-day ADX and DMI. Do not initiate trades when the ADX is directed downward. 2. When ADX begins to move up, go in the direction of DMI, only if the parabolic trend is consistent. If Parabolic conflicts wait and go when it unfold. 3. After entering the trade, use Parabolic Stop to exit. 4. After the release of re-entering when prices cross the parabolas again from the bottom up, believing that the ADX continues to grow and DMI shows that the trend is still valid. This time parabolic system is used to set the time of inclusion. (See Figure 2-68.)
108
109
The method is simple and logical, because the change in the upward direction ADX indicates the beginning of directional movement. DMI indicates whether this movement is directed upwards or downwards. Parabolic SAR provides evidence of the direction narrows stop accelerating, which control the risk at the same time retaining most of the revenue.
Percentage of R (Percent R) Percentage of R, often shortened to a mere% R, belongs to the family of oscillators, which includes stochastic and Relative Strength Index (RSI). Percentage of R, like others, was designed to identify areas for repurchase and resale in non-trending markets.
Larry Williams Percent R (Larry Williams'% R)? Larry Williams is usually associated with the percentage of R - display, presented in his book "How I Made One Million Dollars Last Year Trading Commodities". We believe that George Lane, best known for the development of a stochastic oscillator, was the real inventor of the Percent R, because it really is a stochastic oscillator that uses the price peak instead of the price depression in the numerator. The research is often referred to as the Williams Percent R because of its volume of work on the interpretation and promotion of the indicator. Since this is a stochastic oscillator, we will give a short comment and let us refer to the section on stochastic oscillators, for a more complete discussion of its origin and use.
General trading rules Ratio R is defined as a value between 0 and -100. The value tells you where a percentage of period-end closing price is within the range of the highest peaks and deepest valleys. The letter R in the name of the indicator indicates the range (range). When the closing price is at the top of the range, the market is seen as overbought. When the closing price is at the lower end of the range, the market is considered oversold. The normal time interval 10. We recommend that you use 10 or more as the time interval and the -10 and -90 as levels of repurchase and resale. There are three basic rules for the use of Interest R. Here are the rules for the buying side. Terms of sell-side opposite. You buy when: 1. Percentage of R reached 100. 2. It's been five trading days from the last 100 achievements. 3. Percentage of R again falls below 95. (A graphical representation of these rules, you will find the figure 2-69.)
110
Williams recommended the use of the 10-week moving average, or time, or both, to establish the direction of the market, and then only trade with the trend. Despite the fact that we have described the author's trading methods, we do not recommend them as a tool to follow the trend. Many other studies work best in trending markets. This does not mean that the percentage of R does not work (see Figure 2-69), but if the trend is strong, the value can stay at one end of the scale for days and weeks, while giving a lot of false signals against the trend and little or no signal in the direction of trend. Search multiple challenges trades that will give you the percentage of R in trending markets can be a fruitless and costly exercise. (See Figure 2-70.)
111
Fixing income Despite our negative comments regarding the use of R Interest in a trending market, it may be useful as a tool output. If you are logged using moving averages or some other trend-following method, waiting for the signal of the same indicator for the output often leads to loss of income. Because the percentage of R is very fast, it can be useful in protecting revenue. Some tips on obtaining interest income R: 1.If you are in a long position, use the zone Resale (-90) to determine the bottom of the correction in the trend. Come out or place a stop below the bottom of the correction. 2.Or wait for the percentage of R becomes overbought (-10). Then, if he zavernet down by 10 or more points, take profit or stop install close. (See Figure 2-71.)
112
Divergence Ratio R, either as the oscillator can be very effective when there is divergence. If the market makes a new peak or trough, and the percentage of R does not confirm that its new peak or trough, we have divergence. These signals work best when both the spike Percent R overbought or oversold. Like most divergence oscillator signals generated can be very strong, especially in markets with lateral movement. Despite the rapidity of interest R, you'll find surprisingly little divergence. Do not miss these models. They deserve attention. (See Figure 2-72.)
113
Tic-tac-toe (Point and Figure) Charts tic-tac-toe successfully withstood the test of time by applying at least 100 years longer than the software used today for their construction. This is a very difficult topic on which is written a great many books. What we intend to do here - is to consider some of the modern methods of research plots tic-tac-toe, which are relevant to the user of the computer. Charts tic-tac-toe can be a valuable tool for technical analyst because of the simplicity of the ideas embodied in their base. In much the same manner as moving averages, they allow a person working with the schedule, filter, secondary, or minor adverse price movement. The ability to destroy parasitic price data makes it easy to interpret the graph 114
only the price movements of the user deems relevant. By filtering out distracting noise of price and non-timeline graphic noughts and crosses represent the limit of simplicity - pure price movement. Lack of time scale and the absence of complete price data can be seen by most analysts as a technical flaw. But for those who wish to focus solely on price, or to those who use them in conjunction with other indicators, charts tic-tac-toe offer some unique advantages over conventional methods of charting.
A short overview of the basics Charts tic-tac-toe in our computer age typically consist of characters «X» and "O" placed in columns. The first chart builder tic-tac-toe, such as Charles Dow and RD Wyckoff, used only points (points) instead of X's and O's, while others preferred the chart builder to put prices (figures) in the cells, hence the English name of the method of "point and figure". The column is a series of crosses, rising prices, a column of O's - a series of declining prices. There is no time scale on the horizontal axis, characteristic of normal schedules. Each column can be presented any time. When you configure the graphics tic-tac-toe should be defined two important variables: the size of the cell and the reversal items. The mesh size depends on the amount of the price of items presented in each cell. One cell may be one point or 100 points or more (all from one mark to infinity). Spread elements depend on the number of cells used to determine when a column of X's unfolding in a column of Os and vice versa. Single-cell reversal means that the change in direction of the price of a single cell will give a new column crosses or zeros, applied to the right of the previous column. Most often set to a graphics trehyacheechny reversal. On the schedule for tic-tac-only price changes are applied to quantify the value of one or more cells. For example, if one cell is 10 points, and the last entry in the column crosses were at 110, 115 to the motion is not applied. The market should move to 120 or higher, before another cross will appear in the column. If we used a standard trehyacheechny turn, the prices should drop to 80, before we start a new column of O's. As long as prices remain between 81 and 119, the chart no new highs. (See Figure 2-73.)
Determination of cell size and the topping items 115
The cell dimensions and turns determine the sensitivity of the schedule and the number of columns and models that it will produce. Now, when we have computers that build schedules in seconds, we can very flexible approach to the selection of cell size and the topping items. Trial and error is a practical approach to finding the best values for each market. It was a difficult task, but tic-tac-toe graphics were optimized long before the advent of computers. When searching for the desired ratio should keep in mind some important principles. We have to take into account the risk and volatility. For example, if our plan is to limit the loss of $ 300 includes a stop, we will not use a cell size of $ 300 and trehyacheechny reversal, because we can be stopped, and without making a single entry on the chart. Consequently, the risk must be taken into account when setting up the schedule. The small size allows for better control of risk, when the schedule is used to determine the stopping point. Many builders charts tic-tac-toe using trehyacheechnye reversals as a stopping point after receipt of income. Smaller items are as good as they allow you to record small plotter price movements in the direction of the trend. Adding a new column to the crosses or zeros will capture more of the price movement, as the spreads are metered out in a more profitable level. We encourage our readers to use the following method to find the initial values of the cell and then modify them to suit your needs. First, set a pair of 10-day moving averages, one of which is based on the peaks, and the other - in the hollows. Calculate the current difference between two moving averages and take half of the difference in the value of the cell. Sometimes you need a little rounding, some builders graphics (but not us) prefer to round to the nearest Fibonacci numbers. After determining the size of the cells, we recommend using a standard trehyacheechny reversal. With this method you quickly discover that you can create a surprisingly workable schedules tic-tac-toe on the first try. (See Figure 2-74.) Some authors propose to set the initial size of the cell at the level of about 3 percent of the price in the market. Any method percent of the price is very rude and did not make assumptions on the volatility of the market. We prefer our method of secondary peaks, troughs, because it sets the volatility and gives good results in the graphs with different time steps. For example, the method of peaks, valleys will give great value in the cell for weekly data, while the percentage method will give the same size of the cell and for the day, and weekly data.
Reading between the lines of the trend The simplest way to use graphs tic-tac-toe is to buy every time you start a new column crosses, and vice versa, to go into a short position when you start a new column of O's. Dimensions of cells and reversal elements can be set for otfidt-rovyvaniya twitching. Charts tic-tac-toe form shapes as well as regular schedules. We will not delve into such models as the "flag", "wedge". "Head and shoulders", and the like. There are, however, two very important trend lines and Stith mention about them.
116
45 degree trend line is constantly used plotter tic-tac-toe to determine potential levels of support and resistance. The upward trend line is drawn at an angle of 45 degrees from left to right, bottom to top, from the bottom of the lowest column of Os. Downward trend line is drawn down and to the right at an angle of 45 degrees from the peak of the highest column crosses. The parallel channel line drawn through a series of peaks and troughs can also be very useful when trading charts tic-tac-toe. They are used to support and resistance to a large extent as a line with a slope of 45 degrees. Many experienced traders using the method of tic-tac-toe, consider a breakthrough parallel channel line more significant event than the emergence of a new peak or trough in the graph. (See Figure 2-75.)
117
"Calculate" your income Some builders charts tic-tac-toe argue that price targets can be predicted by one of the two popular methods of calculation. They are called respectively "payment horizontal" and "vertical calculation." Horizontal calculation is based on the same principle as the old graphic method - by measuring the width of the consolidation and then projecting the same distance upward or downward to obtain the target price. In order to use the calculation, we define the horizontal width of the shape of consolidation and then count the number of columns within it. Then we count equivalent to the number of cells from the bottom up consolidation point for the first target. For the second goal, we should make an upward projection from the peak of consolidation. Vertical calculation is very similar. For the purpose of long position we need to count the number of crosses in the first column after Telling her bottom, and multiply the number of crosses on the reversal elements (usually three). Then you reckon up from the bottom of the first column of X's and By enabling our very modest price target. More optimistic goal may be defined by making payment from the breakout point or peak consolidation. I must say that this method is very subjective, and we are always very suspicious methods, trying to predict specific price. We found an example of the market where the calculations have worked very well, but to be honest, we had to look it up. I could cite many more examples where calculations do not fit even close to real prices. (See Figure 2-76.)
Smoothing false breakouts One of the most successful and unusual techniques tic-tac-toe, with which we are faced, is based on an idea submitted by our subscriber, who is a very experienced trader, 118
using the tic-tac-toe. He found that if the graph of tic-tac-toe breaks through a double or triple top and can not be passed on to three cells or more, you can often get a quick profit, following a turn in the opposite direction. (See Figure 2-77.)
Synopsis Charts tic-tac-toe work quite well and can be the basis of many profitable independent trading systems. Charts are still popular in the halls of exchange and are used by some professional advisers. Most of today's technical analysts prefer
use them in conjunction with other technical indicators, such as stochastics and relative strength. Some traders argue that the graphics tic-tac-toe can be extremely useful in the analysis of Elliott Waves, where graphics can be set to filter out many unwanted or extraneous waves. (Since we do not use wave analysis, we can not comment on it.) Charts tic-tac-toe is usually considered to be simple to operate and extremely difficult to set up. Now, when computers can do the job, and graphs can be created and modified immediately, we think, the main drawbacks are overcome. We have offered our readers find
119
time for further training and to set the page for each market, which would contain the schedule of tic-tac-toe, based on weekly data, and one or more schedules with daily data. Experiment with the size of the cells and the reversal elements. We think that you will become a better trader through such an experience, and you will enjoy a completely new perspective on markets, given by graphs of tic-tac-toe. Most traders make the computer go into the small movements of markets and tightened too little change in prices. Charts tictac-toe can be a nice way to step back and look at the broader perspective.
Relative Strength Index (RSI - Relative Strength Index) Relative Strength Index (RSI - relative strength index) is probably the most popular of all counter trend oscillators. The index gives reliable signals repurchase and resale in most market conditions. RSI also produces excellent long term pattern of divergence, which can be used to identify the major peaks and valleys. RSI can be used both as a means of income, and fine-tune the instrument of market entries, derived from the signals of other methods. RSI formula was invented by J. Welles Wilder Jr. and fully explained in 1978 in his book "New Concepts in Technical Trading Systems". RSI calculates the ratio of the upper to the lower closures closures in the chosen time period and reflects the result as an oscillator with a scale from 0 to 100. The formula is as follows: RSI = 100 - (100/1 + RS), where RS = average top closure for the last n days divided by the average closing lower 120
for the last n days. The value near 0 indicates the resale market, and a value of about 100 - about his takeover. Wilder recommended the use of the 14-day time period, which he understood as a half cycle in most markets. (See Figure 2-78.)
At 14 days using as the default value can be expected that the market peaks and valleys will occur some time after RSI 70 rises above or falls below 30. We do not recommend to buy or sell exactly at these values, because the presence of friction, yes RSI often "sticks" to one end of the range for days or even weeks, giving false testimony about the peak or trough. Wilder and others advocated the use of some standard graphics techniques with RSI, alleging that certain figures predict a similar index figure of downstream data. The following are some examples of signals RSI, which we consider useful, guided by our own research and experience.
Spurious oscillations (Failure Swings) The first of these formations is false oscillation, which is easier to watch on a study of the actual RSI, rather than on an underlying graph. False oscillation comprises cleat formed RSI, projecting the mark 70 followed by a new stud with a lower peak than the first. The real sell signal is produced when overcome the low point between the studs RSI. Buy signal is the inverse model with two spikes pointing downwards and then overcoming the highest point between them in the upward direction. (See Figure 2-79.)
121
As you can see, the false oscillation can be a powerful signal. Remember, the best signals occur when the first spike goes far beyond the 30 down or away for 70 up. You can not afford to ignore these events. They usually celebrate significant interim changes in market direction. Beware of spurious vibrations, which have such a small number of deviations, which require a long time for their definitions. Our experience shows that the best spurious oscillations occur rather quickly and are easy to detect.
RSI divergence model Weekly charts We believe that the most significant and powerful signals arise in the form of RSI divergence between the index structure and the structure of an underlying graph. We found this divergence is especially useful in detecting major long-term peaks and troughs on the weekly charts. For example, take a look at the weekly chart S & P, leading to a peak in August 1987. (See Figure 2-80.)
Daily charts. We recommend using the 10-day and 14-day RSI detection patterns daily divergence. Examples abound. The first graph shows a divergence with the 14-day RSI. Note that the RSI divergence confirmed the inability to reach a new depression, indicating that the market is technically strong. Make sure that your entry into the purchase comes at a time of the day after the rise that marked the bottom of the second pin, rather than at an earlier time. (See Figure 2-81.) The following chart shows the divergence of the 10-day RSI, which predicted a breakthrough on Treasuries in early August 1989. Entry for sale - either at the close of August 2, or at the opening on August 3. (See Figure 2-82.)
122
Although it is difficult to formulate certain rule, we have found that the divergence in which peaks are separated by only a few days or more than 10 weeks, usually do not give good signals.
Filter entries RSi One of the most common problems faced by trend-following system is entering the market after a strong reversal. Entry will never fall exactly on the turn of the market, but occurs after a significant price move in a new direction. Most short-term movement, the trend unfolds, the market does either overbought or oversold, making it vulnerable to short-term correction. Almost all encountered this problem after receiving a signal following the trend, caused by a powerful change of direction. Should you dare to risk entering into the trade after three days of heavy lifting? Or five days of recovery? Will you wait? If you wait, then what are you waiting for?
123
Filter RSI offers the perfect solution to this common problem. If the value of RSI above 75 (if you buy) or below 25 (if you sell), then set aside your entry. Come only when the RSI will go back to a level between 75 and 25. Will necessarily be a minor market correction, and your entry will not occur in the levels of repurchase or resale. (See Figure 2-83.)
124
Re-entry with RSI Let's assume that your trade was stopped, and the trend is still continuing. That in this situation need is an accurate way to set the time of re-entering to your initial loss was minimal. Use a short-term (eg, 3-day) RSI and wait for it to turn around and go only in the direction of the trend. To illustrate, let's assume that your indicators suggest a downward movement of the market, and you need a point of re-entry. Further assume that the RSI dropped and is now below 50. Try to wait until the RSI back at the level higher than 50, then when it unfold down immediately sell. Waiting for easy upward movement is very sensitive short-term RSI has the effect of reducing any intermediate conditions repurchase or resale, allowing you to re-enter during a minor correction of the trend. (See Figure 2-84.)
Fixing income with the aid of RSI One of the most valuable applications RSI - use it to secure the revenue. Always nice to sit back and allow income to flow, but the use of relatively slow research will inevitably lead to the loss of income to generate a signal at the output. Need a way out, quickly recognizing when the market is at its peak, in conjunction with the method of tracking stops, which allows income to flow until the market continues to move. To fix the revenue try to use short-term RSI, 10 - or 14-day RSI is usually not sensitive enough. Latch signal finishes the formation of income when the RSI reaches 75 and above (25 and below, if you're in a short position), and then returns to 10 or more points. For example, RSI has risen to 87, and then went back to the 65. At this point the market is slowing down, and the need to take steps to protect your income. Install the nearest stop or a depression in the last n days, or at a predetermined dollar value depending on that are closer. (See Figure 2-85.)
125
We found that n-day depression is very useful as a point of tracking. Quite often, the market will move back without starting your stop, and you can keep track of her for a long time. Dollar tracking stops also works quite well (see "Testing the System", Chapter 3).
126
The slow stochastic oscillators (Slow Stochastics) We believe that in addition to the trend-following strategies applied by most professional traders, consistency in performance can be improved by means of a technical indicator for use on non-trending markets. The stochastic oscillator is one of the best tools. If you care about is to stay on the side of the trend, it can be used in trending markets. Stochastic oscillators have been popularized by George Lane, who used them in their educational courses on investment since the early 50's. We attended many lectures and George each time learning about something new stochastic oscillators. He perfected the use of stochastic oscillators for many years of his trading activities and can find innovative ways of how to make them work well in almost any situation. We owe much of our knowledge in the field of stochastic oscillators. Much of the following information was derived from an article by George stochastic oscillators, written for Technical Traders Bulletin. The basic formula of the stochastic oscillator is as follows:% K = today's close minus the trough of the last n days, divided by the peak of the last n days, minus the trough of the last n days. % D ~ a three-day moving average of% K. % K. % D and produce the socalled fast stochastic oscillator, which is rarely used because of its extreme sensitivity. (See Figure 2-86.)
127
Fast% K and% D, smoothed over a three-day moving average, given the slow stochastic oscillator, which is used more often. (See Figure 2-87.) * While not stated otherwise, in our subsequent discussion, we will talk about a slow, smooth version of the stochastic oscillator. The formula expresses the stochastic relationship between today's close and the range between the peak and the trough of the last n days. For example, when today's closing is equal to 30 and the range of the last 10 days, 20 to 50, the fast A = 30% - 20/50 - 20 = 0.33 is relatively small value. If today's close is 40, which is closer to the top of the range, the faster% K is equal O.bb. % K and% D can not be less than 0 or greater than 100. With the accumulation of days,% K and% D are depicted as lines, varying from 0 to 100. Values close to 0, in theory indicate oversold market. Values close to 100, theoretically indicate an overbought market.
128
The main stochastic oscillator signals are the intersections of the lines% K and% D in combination with the level of the% K and% D, testifying to outbid-raising or oversold. Usually oversold shows the value of% D below 30, and is overbought but ld - above 70. The values of 80 and 20 are also often used. We also saw traders are indifferent to the% K and tracking of when the level reaches the% D repurchase or resale.
Time Periods * Usually recommended for slow stochastic oscillator time period is 18, but George Lane uses a wide range of values, finding what he understands by the dominant cycle of the market in which trade is conducted, and then uses half of this number as the period for the stochastic oscillator. Technical analysts seem to have their own preferences. Our experience and tests suggest that the range between 9 and 12, the best compromise between speed signals (intersection% K and% D) or making and suitability to the logical completion of the signal produced by them, with a minimum number of false alarms. Like all the other technical studies, stochastic oscillators quicker reaction to market action with shorter time periods, and slower for a longer period. We will discuss some of the techniques used by other technical analysts to speed up the signals. We believe that these methods are not necessary. If you need faster signals, simply cut the time period. Do not forget that faster is not always better. You should look for the most reliable signals, but not the fastest.
When to use stochastic oscillators Stochastic oscillators work best on wide price ranges or soft with a slight trend up or down gradient. The worst market for the normal use of the stochastic oscillator is a market which is in a stable trend and affected only a minor correction. In such a market stochastic oscillator will produce a lot of concentration trtrendovyh entry points to be quickly drained of blood trend. If you continue to use the standard methods of trading with stochastic oscillators, you get a serious streak of losses. Remember, a trader who came up with the adage "the trend - is your friend" ("the trend is your friend"), did not use stochastic oscillators. (As an example of the loss when using stochastic oscillator see Figure 2-88).
129
How to identify and to quantify the shape of the market, which is in the "strong" trend? There are many ways, however, if during the "strong" trend is not obvious, try to measure the trend with ADX (see section DMI / ADX). You can trade with the stochastic oscillator on the trend, if you ignore the usual levels of repurchase and resale of 70/30 or 80/20, and to enter the market at the end of the signal resistance trend that given by the intersection of the stochastic oscillator at any level. However, there are better ways of following the trend, and we believe that our core value of stochastic oscillators become as indicators of peaks and troughs.
Divergences When the market creates a new peak or trough and the stochastic oscillator can not attest to its peak or trough, we have divergence. These may be simple divergence (again, see Figure 2-88) or "classical divergence signal" George Lane arriving at the triple peaks. (See Figures 2-89 and 2-90.) Note in Figure 2-90, the second peak (2) is lower than the first (1). The third peak (3) is higher than the second, but lower than the first. If the market does not behave as expected of him, * we may face a "model of secondary divergence" like the one shown in Figure 2-91. In this model prices also make three leap to the top, but the stochastic oscillator produces three descending peaks that create divergence. This "secondary" pattern, as opposed to classical point 2 point 3 below. We found that these models divergence give much better results than classical signals. This is true for the majority of the oscillator. We are pleased to have been more precise and specific as to the results of divergence, but they are extremely difficult to objectively test.
130
The left and right crossing Some technical analysts have speculated that the slow change in the direction to which responds% D, is somewhat more credible than a quick change of direction, which measures the% K. The fact that these traders are looking for is a model where the% D is beginning to change direction before crossing, which would mean that the% K crosses the right side of the peak or trough line% D. This sequence produces a "right cross" is 131
opposed to "the intersection of the left," where the% K crosses% D before the latter begins to change napravlenie.Ideya is that the right crossing produce better signals than the left intersection. We do not see any logic in that reasoning. When we are able to distinguish left from right-wing cross, which sometimes is not easy, we did not observe any correlation with the success of trade. (See Figure 2-92.)
Knees and shoulders When the% K crosses% D from the bottom up and then deflected downward by a few percentage points in the next period, but could not break through again% D in front of his new turn up, Lane calls it a "knee." This supposedly indicates the strength and stability of the increase in prices. If we have a similar pattern on the bottom side, is called "shoulder". The knee and the shoulder usually occurs when prices are making the shape of the stud which we described earlier. Rapid change in the trend forms a "left-hand" intersection in the stochastic oscillator. In models of the knee and shoulder% K gives a divergence on the thorns, which can not make the% D. (See Figure 2-93.)
Hooks and loops - warning model Some researchers have attempted to develop a stochastic oscillator methods that would have predicted the actual intersection and, therefore, would provide the ability to proactively start. Two examples of these models are predicting a loop and hook warning. Loop model is the simple observation that the% D line begins to curve, predicting the imminent change of direction to the intersection with the real% K. Warning hook - this observation is of extreme rotation or change the direction of the% K line to the intersection with the% D. We think that both of these models are too early signals instead of giving reliable signals. We recommend waiting to cross. Be patient and do not rush to the shadows. These predictive models can be pain-
132
Shui value if you will use them to secure income, not for entries. (See Figure 2-94.)
Bearish and bullish setup Bearish and bullish setup - this is yet another special tools George Lane. Bear installation occurs when prices rise and create a series of rising peaks and rising troughs. Stochastic oscillator gives divergence and creates a model of declining troughs at the time, as prices continue to rise. This setting indicates that the next upward spike can produce significant peak. (See Figure 2-95.) Bullish installation occurs when prices fall and produce a series of depressions and dropping dropping heights. Stochastic oscillator gives divergence and creates a schema of rising peaks at the time, as prices continue to fall. This setting indicates that prices will soon establish a major depression. (See Figure 2-96.) 133
Bearish and bullish divergences are the reverse setting (some call them convergences). Several of our subscribers sheet reported successful use of such models.
Fixing income Because of the nature of the counter trend stochastic income should be obtained quickly. Do not wait too long to the output signal from the stochastic oscillator, or you will find yourself in an awkward situation, turning a winning position into a losing. You need to develop a method to generate income until the market moves in a favorable direction. Well the method works price targets. Do not be greedy - use the success and conduct a tactical retreat. Sometimes trade with the stochastic oscillator can develop into a trend. If you are lucky enough and saved the position, you can change the strategy and allow your income flow. Such trades are rare, and if you've followed our recommendations for generating revenue, you have probably already come out of the trade when the trend is confirmed.
134
Volatility (Volatility) A large number of technical research and trading systems measures the volatility of the market. Virtually all studies Wilder (RSI, DMI, CSI, of parabolic, etc.) or by other means include the concept of volatility. Volatility is also part of various studies: trade lanes and envelopes (for example, Bollinger Bands), and is also a key component of the analysis of tic-tac-toe. Unfortunately, a simple calculation of volatility is not part of the standard features of most software packages and additional programs offering of volatility, belong to the class of "black boxes" where their methodology is not disclosed to the user in full. We use relatively inexpensive and fully open the program, which is called "Professional System Breakthrough" ("Professional Breakout System"), developed by Steve Noutisom and our testing the software System Writer Plus for most of the research related to the volatility. We prefer to use a volatility as an additional tool, not as a basis for the system. Most systems are based on the volatility of the need to rely on meaningless optimization that makes perform well on historical data. During periods when volatility is working well, the results are quite impressive, including cases of purchase exactly trough of the market when prices break through, and selling at the top, when the prices fall down. For volatile markets such systems achieve high-end results of this particular trade item for a short period of time. Rarely, however, can be found based on the volatility of the trading system is well-proven on a diverse portfolio for a long time. Not surprisingly, the volatility is the basis of a number of trading systems, which were sold from the early 70's at the price goes up to $ 10,000. All these systems used substantially the same method. Most of them are direct descendants of earlier similar systems with minor changes, which in many cases have been added only in order to avoid violations of intellectual property rights. They say that many of these systems are based on volatility, were highly profitable.
Measuring volatility All based on the volatility of the trading systems use the concept to determine the range of the recent market movement. The simplest definition of the range - the distance between the peak and the trough of the timealternating periods. Usually taken daily, but it can also be a week or a month, or even intraday period, measured in minutes. This simple definition of a range of mostly works fine, but it does not take into account the days with extreme price movements. Limited days, for example, may have a very narrow range, but the market is obviously highly volatile and volatility increases. Similarly, a day with a gap at the opening, in which trading takes place outside the range of the previous day, is an example of the increasing volatility, even if the actual range of the day is less than the previous day. Wilder saw this problem and introduced the concept of "true range" (TR-true range) as the higher of the following: 1. The distance from today's peak-to-trough today. 2. The distance from yesterday's close to today's peak. 3. The distance from yesterday's close to today's depression. (See Figure 2-97.)
135
By itself, the true range - it is still an isolated number. To make it meaningful, we need to take a certain number of previous days and find the value that gives us the average true range (ATR - average true range). This is a direct measurement of market volatility. If ATR increases, the market becomes more volatile. If ATR is reduced, the market becomes less volatile. It is difficult to answer the question of how many days are needed to obtain the "best" ATR. Author's formula volatility Wilder used 14 days, but modern systems vendors have optimized this variable and found that any number from 2 to 9 days works better.
How do the systems of volatility All the popular trading systems based on the volatility of the work based on the principle that a breakthrough or a price spike that goes beyond the recent range or average true range is a significant event, and it should be used as a point of entry into the market. For example, ATR in futures NYSE Composite over the past five days was 1.00 point. We would be interested in the price movement, expressed as a percentage of ATR, say 150 per cent, from the previous day's closing price. This means that we have either bought or sold if prices advanced to 150 * 1.00 percent, or 1.50 points. If the previous day's close was at 190.00, we will buy or sell at 191.50 to 188.50. Two variables of the system are: (1) the number of days used for ATR, (2) the percentage of movement in relation to the closing price the previous day, which is needed breakthrough. These variables are useful for optimization and calculation of the volatility can be easily customized to meet the past data of any particular market. Today, most of the available software packages ^ relies on optimization to determine the exact values used for each variable. As you might guess, a breakthrough system based on the volatility, the current systems are always present in the market. Each day, after the close of Compute ATR and then multiply it by the percentage of motion required to trigger trade. Add the result to the closing price, and you get the point at which the next day will start buying. Subtract the 136
result of the closing price, and you get the point, which will be launched for sale. Set both the order of the day and you are now ready to go. (See Figure 2-98.)
Comments and variations One of the highlights of the system described above is that there is no neutral zone, since you are either a long or a short position-risk on each trade is the difference between the point of entry and the point of reversal. If they both were launched on the same day or close in time to each other, then you obviously get twitching. The risk on trade depends entirely on the recent market volatility, which may be consistent or not consistent with the thickness of the purse trader or his technique of cash management. Another interesting aspect of the system is the volatility that is the entry point and the pivot point will move away from each other with increasing short-term volatility. Very easy to see how this can happen: the market moves, the range increases, and stops are farther and farther apart. This would reduce the twitching, but this situation may also increase the risk of early trade after entering it. All of this can be confusing and potentially damaging to the strict control circuit means, which plans to risk in the amount of a certain fixed amount of dollars on the trade, and then it turns out that this number has increased at a time when trade is already running. It is also possible that the pivot point will move away almost indefinitely. Let's say that Treasuries are at around 100, the system is in a long position and the reversal percentage is 150 percent two-day ATR. If ATR is not changing, the movement required to include a short position, too, will remain the same. If Treasuries slowly fall every day with day range large enough to hold in place ATR, and the short position is still not included, theoretically turning point may never work. It will just continue to pull away. This apparently rare, nevertheless possible, and such sequence could cause large losses (and test sequences are).
The disadvantages of systems based on volatility We believe that trading systems based on volatility are good for short-term use, but are limited in the long-term work. Their performance results often show real prospects for short
137
spurts, but they also have a tendency to lose your winnings over time and long-term work may not be better than break-even system. There are a few things that cause us concern. First, all vendors conducted their extensive optimization to find the "best" values for key system variables - the average true range, and the percentage of movement necessary for the inclusion of trade. Probably, suppliers have concluded that once were found magic (optimized) number, which give spectacular hypothetical results, then the system will be profitable in the future. Any variations of the systems based on the in-volatility, are extremely small and can be reduced to the two variables. For example, it may change slightly the definition of the average true range, or may be replaced by a simple daily range. Or supplier prefers to calculate the percentage of traffic from the opening price the next day instead of the closing price the previous day in order to include a large factor in night break and reduce twitching. These minor changes do not prevent large losses in performance results of the system. From our point of view, the problem of loss is the result of two factors: over-optimization, and possibly wrong conclusion that volatility works just as well when you set the outputs as inputs for the job. Now most of our readers alerted our negative feelings regarding the optimization and recycling systems. We believe that optimization is targeted adjustment to the curve giving useless and overly exaggerated illusion of potential profitability. However, a properly conducted test and subsequent anticipatory testing, followed by a real-time tracking, it may be worthwhile and valuable exercise. But let's look and think: if a simple optimization really worked, then to date, few geeks already many times have captured or destroyed all markets.
Recommendations Despite the problems that we consider typical of the approach based on volatility, we still feel that these systems are operating potential. We think "that the movements of volatility occur in the trend. The real difficulty is common to all approaches following the trend is to frequent antsy when the markets are not in the trend and have a low volatility. During the period of the markets will be either sluggish or dynamic, and most of the time they will stay in the flaccid state. As the system moving average of volatility, adjusted for trending markets will not work well for periods of sideways price movement. You can substantially reduce the initial risk for each trade by creating a neutral zone between the points of entry into the short and long positions. The easiest way to do this is to stop the risk of setting the interest that will be less interest ATR, including the entry. For example, in our example, we had the ATR at 100 points NYSE Composite, and we'd bought at the level of 150 per cent of this amount, or 150 points. Once we got into the trade, closer stop can be set by subtracting the smaller percentage of the ATR from the point of entry. Logically, all that is less than 100 percent of the ATR, will be seen as too close, exposure to almost random twitching, but the use of such numbers as 125 percent still requires the occurrence of extraordinary events and gives a closer stop level than the standard pivot point. If this happens the risk-off point, the system is in the neutral position until the trigger signal to buy or to sell at a higher volatility. Another possible improvement would be the lack of trade when the market is behaving sluggishly, especially in periods when volatility is at an unusually low level. It is possible to use a range of optimum profitability ATR for each market in which it is located between the acceptable boundaries are not too low and not too high. (See Figure 2-99.) It can be concluded that the sluggish market with a relatively small range will result in losing trades,
138
while the more volatile the market will tend to be more profitable. Usually there is a desire overoptimize system when markets become sluggish, but long-term it can use
be advantageous to completely stop all the action at times of stagnation and wait for the ATR will be more fit to what your intended system for the successful operation. (Again, refer to Figure 2-99.) A third possibility would be to add an external filter to determine the conditions that must be met before it can be adopted breakthrough. There are at least two ways to implement this approach through the use of ready-made technical research: DMI / ADX and CCI. We have repeatedly mentioned that the upward direction indicates Wilder ADX trending markets. Try trading breakouts volatility of only when lifting 18-day ADX. (See Figure 2-100). Similarly, the 20-day RSI CCI, based on monthly or weekly data also show you the extent to which the market is trending in a long period of time. Look for acceleration CCI with its zero level, and if this condition is met, the market is likely to move quickly enough to make a profitable trade based on volatility.
139
Quick outs Probably the best way to make it work based on volatilnos-minute system is the use of methods yield faster and more sensitive than the methods of entry. Using a smaller percentage vodatilnosti for signal output (as we described earlier) should reduce the loss of income, a reversal of the conventional system based on volatility. It also gives the time when the system does not trade. In theory, you need to avoid market action after the peak of the trend when prices go against the trend and volatility of the breakthrough, which will replace the position has not yet come. Other possible methods of outputs, which we recommend be Parabolic Wilder and simple dollar stop, or stop tracking percent (by percentage we understand the percentage of an underlying contract). Faster output will mean the loss of time from the time of the trend, but it's better than losing a significant part of each profitable trade.
Volume and open interest (Volume and Open Interest) It can be assumed that the trading system based on the volume and open interest will be operational. The one who drew graphs of volume or open interest, notice that they are often the same or ahead of price action. Volume often increases when the market breaks out in either direction. Falling open interest in a bull market sometimes tells you the market peak. Logically, the buying or selling pressure, reflecting the volume and open interest, there must be at least as effective as the techniques based solely on price. However, the value of volume and open interest as indicators of a large degree of importance is an illusion. We agree that there are rare instances when examining the relationship between price, volume and open interest can be helpful. Volume and open interest are the reinforcement of our beliefs about the strength of the market breakthrough and warn us of the approach of the market peak or trough. Despite the widespread promotion of volume 140
and open interest, we do not believe that they can serve as the basis of the trading system. At best, they can partially be used as support or preventive tools. Note: the future and our previous comments apply only to the terms of a futures and not to the volume of the stock market. Volumes in these two types of markets are quite different.
Trade with the volume The analysis of the futures market has always had a lot of difficulties in the application. One of them is the fact that the value of its formation does not end until nearly the end of the trading day following that, where there were real transactions. This means that a trader trying to make a timely analysis of the information is delayed by at least put on. This may turn into a two-day delay, if you do not get the information in real time and have to wait tomorrow morning bulletins. Another obvious problem mazakltochaetsya the full amount of the distortion created in the limited days. The volume in the limited days is often negligible, however, this occurs when the market is in a strong or weak price action. Here we have a case of the obvious predominance of buyers or sellers, but we can not measure it. (See Figure 2-101). Traditional volume analysis usually begins with the assumption that the volume will increase with the continuation of the market trend and decreased when the adjustment of the market. In markets with the lateral movement of the amount will be small. The volume will produce thorns breakout of the market (in any direction) and will grow
if the trend on the market. We noticed that the largest volume breakout will occur in areas of the rising trend of the market. Since this behavior is considered normal volume, any deviation from this behavior may give rise to suspicions that the trend is not as strong as it seems, and that perhaps it is not necessary for him to follow. If, for example, a break occurs in a small volume, many merchants will not go to the market, as long as the amount does not increase, thereby confirming breakthrough. Later, the trend mogutzahotet other traders close positions when the volume started to fall, while prices continue to rise in the direction of the trend. The 141
common assumption is that the amount of predicted price movement (very debatable assumption) and that the price action must be supported by the change in volume. We see serious flaws in the methodology. First and foremost, we are trading price, not volume. If the market is able to trend and we get the revenue, what we see to it that it does not support the volume? Indications of, from our point of view, even about not reliable enough to serve as a basis to exit a profitable position. Prices continue to move - that's what's important. Reasonable control of risk and carefully selected lead you stop watching the market more readily than it will make volume analysis. As already mentioned, we do not believe in the truth of the basic assumptions that the volume anticipates prices. It's as logical as to conclude that the price action predict or attract volume. Nothing attracts the attention of the public so as a breakthrough in the upward direction. And remember: this is futures trading, not the stock market. Each buyer must always be found by the seller. There is not a breakthrough in the volume up or down direction, as it happens in the stock market, and there are no experts who are trying to maintain order in the market, risking their own capital. In the futures markets, each buyer thinks that he is right, just think and sellers. The volume itself can not push the market, because on one side of the market can not be larger volume than the other. They are always exactly equal. We also noticed that the volume analysis largely involved judgment retroactively. (Because his message is always delayed, it can hardly be considered as a leading indicator.) If the volume produces a spike during the break, and the market continues to trend, everyone will say that the volume ahead of the price and volume spike that confirmed the beginning of a trend. If the trend does not continue, everyone will say that a breakthrough has been an explosion of consumer interest, confirmed by volume jump. The judgment of hindsight is always knocks 20 out of 20, and the volume is always ready to confirm any outcome. (See Figure 2-102).
Open interest
As is well known to our readers, open interest - the total number of open contracts on the market by the end of the trading day. Because for every buyer there is a seller and vice versa, the total open interest tells you about the number of both long and short contracts on the market. These numbers should always be precisely equal. Open interest is usually represented by a line above the graph of volume. The numerical values of open interest is calculated clearing house in the same way as the values of the volume and, therefore, too late. Public interest in a more general sense than the volume measuring the amount of money invested by or extracted from a futures market. If the two sides trade new to the market, open interest increase. If one side of a new and
142
the other side closes the position, open interest will not change. If both sides are closed, open interest will decrease. If this were the only possible variables, the interpretation of open interest would be much easier than actually is. Open interest in many markets, especially in agricultural commodity markets, changes with the seasons. Active hedgers increases during the harvest season, raising the public interest. Lifting open interest may carry sense only if it differs significantly from the normal seasonal trends. Graphic Products Research Bureau (CRB-Commodity Research Bureau) represent the five-year average open interest is to simplify these seasonal comparisons. Open interest will vary with the movement of speculative interest from one contract to another on the same market. If you follow the open Besides the interest only upon completing the contract, it may give false values for several reasons. For example, the drop in open interest in the growing market that comes close to delivery or expiration of the contract, may not mean that speculators throw the market and that the top is near. Falling open interest may simply mean that they believe in the power of the market and want to get out of the ending of the contract and enter into a contract that will allow them to be on the market longer. (We recommend to watch the open interest and follow him instead of learning it from month to month.) The duration may also affect the public interest under specific contracts, so be careful in bold interpretations of the data. In our opinion, the ideal schedule for the open interest will be an endless continuous graph showing the cumulative total open interest plus a five-year average.
The interaction volume and open interest
The easiest way to demonstrate a link between the volume, open interest and price action is to build a table like the one shown below.
143
As the table shows, the traditional interpretation of the open interest consists of four possibilities. 1. Prices are rising and open interest is growing. This means that new money pouring into the market and there is pressure from buyers. (Do not make the erroneous conclusion that there are more buyers than sellers. Growth rates indicates that buyers are willing to pay more and, of course, the sellers are willing to cooperate.) This situation is seen as bullish. 2. Prices are rising and open interest is falling. On the market are relatively few new buyers, and funneling money from it. The rise is likely due to the fact that holders of short positions close them, thus leaving the market. This is often bullish trend for a short time, because being in a short position is usually willing to pay any price to get out, and because they can not afford to stay and endure further losses. This action is extremely bearish. Without the receipt of new money into the market upturn will stop as soon as the positions are in short stop to close. But the closing of short positions may be renewed and last longer than anyone could have expected. According to George Lane, who is a very experienced trader iyarym advocate analysis of volume and open interest as confirming indicators, the total open interest (assuming the open interest of all contracts in the aggregate) always starts falling for five to eight days to reach the last vertex. 3. Prices are falling and rising open interest. The new money is coming into the market, and there is a selling pressure. This behavior is considered bearish. 4. Prices are falling and open interest is falling. It is the opposite of the situation 2. Now sellers in short positions * make money and can afford to stay on the market. Most of crushed by falling prices caused by the bulls, closing their positions. On the market receives little money. It was initially a bearish situation, but the market is ready for consolidation posletogo as the elimination of the broken ends of long positions. Short sellers are known for their lack of patience, and is likely to begin to close as soon as the down time will begin to subside. We do not see anything wrong with these interpretations, but there is a question of their value. If you want to verify the correctness of their assumptions, the interpretation of these may be useful as a method to confirm. We do not think that trade only on the basis of the analysis of open interest will be fruitful. Having said that changes in open interest does not precede price movements, we should mention an important observation concerning public interest, noted by many years ago. In markets with weak lateral or downward movement of an unexpected drop in open interest is often followed by the consolidation of the market. (See Figure 2-103.) Very often large market participants bid up the price in anticipation of market consolidation. There is a feeling that "insider-ry" know which way the market will go, and set their positions against the alleged movement. The theory of "insiders" can be seen in studies of volume and open interest. It is difficult to believe that anyone ever really "knows" what path the futures market. In fact, contrary to popular predstavuniyu, large commercial firms trading futures because they do not know which way the market will go. If they or anyone else knew anything with any degree of
144
certainty, the futures market would cease to exist in a matter of months. (Again, see Figure 2-103). A more likely explanation for the drop in open interest before the consolidation is a lack of certainty about the direction of the market in some
traders. Are in short positions are not interested in the market because of low prices. Are in a long position in the market is not interested in continuing a downward trend. However, consolidation is usually always start with a very negative attitude to the market, which is already low prices. In such a situation, such as a mood indicator "Bullish Consensus" Hededi will probably be better than the volume or open interest.
Studies of volume and open interest
The most famous and accessible technical research, using the volume balance is the turnover (OBV - on-balance volume), promulgated in the early 1960s by Joseph Granville. The calculations are simple. Scope of the day each assigned a plus or minus depending on whether closure of the day in the upper or lower direction. The resulting values are added to the accumulated sum, becoming a line OBV. The current value of OBV line is not really important, but figures it produced, can make a difference. Traditional logic volume / price is applied to the line of OBV, especially with regard to the divergence between OBV line and prices. We have doubts as to the correct application of this tool to the stock market futures markets. It makes no sense to attribute one side of the entire amount of one day, as the volume in futures markets is always divided 50/50. There are additional more subtle approaches to the question of volume, among them are particularly noteworthy oscillator accumulation turnover (VA - volume accumulation), developed by Marc Chaikin, and the Index of Demand (Demand Index), invented by James Sibbetom. From our point of view, the market shares of these indicators are not well suited to trade futures, but they can be effective in the securities market. (Trading of shares was lucky - they did not have to deal with an open interest.) Our conclusion, if you have not already guessed, is to ensure that the volume analysis works well in the stock market, but not on the futures. Public interest, at best, can be used as a confirmatory indicator. 145
Recommended reading
DMI and ADX
Babcock, Bruce, Jr. The Dow Jones-Irwin Guide to Trading Systems. Home-wood, 111.: Dow Jones-Irwin, 1989. Colby, Robert W., and Thomas A. Meyers. The Encyclopedia of Technical Market Indicators.Homewood, III.; Dow Jones-irwin, 1988. Hochheimer, Frank L. Computerized Trading Techniques. 1982. New York: Merrill Lynch Commodities, 1982. Wilder, J. Wetles, Jr. New Concepts in Technical Trading Syslems. Greensboro, N.C.: Trend Research, 1978.
Bands (Bands), Envelopes (Envelopes) and channels (Channels)
Babcock, Bruce, Jr. The Dow Jones-Irwin Guide to Trading Systems. Home-wood, III.: Dow Jonesirwin, 1989. Gallacher, William R. Winner Takes All: A Privateer's Guide to Commodity Trading. Toronto: Midway Publications, 1983. Hochheimer, Frank L. «Channels and Crossovers: An Explanation and Computerized Testing of Commodity Trading Techniques.» In Technical Analysis in Commodities, P. J. Kaufman, ed. New York: John Wiley & Sons, 1980. Irwin, Scott H., and J-William Uhrig. «Do Technical Analysts Have Holes in Their Shoes?» Review of Research in Futures Markets ^, no. 3 (1984), pp. 264-81. Lukac, Louis P.; B. Wade Brorsen; and Scott H. Irwin. A Comparison of Twelve Technical Trading Syslems.
The Commodity Channel Index (CCI - Commodity Channel Index)
f Colby, Robert W., and Thomas A. Meyers. The Encyclopedia of Technical Market Indicators. Homewood, III.: Dow Jones-Irwin, 1988. Lambert, Donald R. «Commodity Channel Index: Tool for Trading Cyclic Trends.» Commodities, October 1980.
Moment (Momentum) and rate of change (Rate of Change) Babcock, Bruce, Jr. The Dow Jones-irwin Guide to Trading Syslems. Home-wood, III.: Dow Jones-irwin, 1989. Colby, Robert W., and Thomas A. Meyers. The Encyclopedia of Technical Market Indicators. Homewood, III.: Dow Jones-Irwin, 1988. Eng, William F. The Technical Analysis of Stocks, Options and Futures. Chicago: Probus Publishing, 1988. Kaufman, Perry J. The New Commodity Trading Systems and Methods. NewYork: John Wiley & Sons, 1987. Murphy, John J. Technical Analysis of the Futures Markets. New York: NewYork Institute of Finance, 1986. Pring, Martin J. Technical Analysis Explained. New York: McGraw-HilI, 1985.
Moving averages (Moving Averages)
Maxwell, J. R., Sr. Commodity Futures Trading with Moving Averages. Red Bluff, Calif.: Speer Books, 1976. «TTB Interviews Joe DiNapoli.» Technical Traders Bulletin, February 1990, pp. 1-7.
Trading method MACD (MACD Trading Method)
Appel, Gerald. The Moving Average Convergence / Divergence Trading Method-Advanced Version. Toronto, Ontario: Scientific Investment System, 1985. Babcock, Bruce, Jr. The Dow Jones-irwin Guide to Trading Systems. Home-wood, III.: Dow Jonesirwin, 1989. 146
Parabolic SAR (Parabolic)
Kaufman, Perry J. The New Commodity Trading Systems and Methods. NewYork: John Wiley & Sons, 1987. Wilder, J. Welles, Jr. New Concepts in Technical Trading Systems. Greenboro, N.C.: Trend Research, 1978. Percentage of R (Percent R) Williams, Larry R. How I Made One Million Dollars Last Year Trading Commodities. Carmel Valley, Calif.: Conceptual Management, 1973.
Tic-tac-toe (Point and Figure)
Cohen, A. W. llow to Use the Three-Point Method of Point-and-Figure StockMarkel Trading. Larchmont, N.Y.: Chartcraft, 1972. DeVilliers, Victor. The Point-ana-Figure Method of Anticipating Stock Price Movements. New York: Trader Press, 1966. Kaufman, Perry J. The New Commodity Trading Systems and Methods. NewYork: John Wiley & Sons, 1987. Whelan, Alexander. Study Helps in Point-and-Figur »Technique. New York: Whelan, Alexander. Study Helps in Point-and-Figure Technique. NewYork: Morgan, Roberts and Roberts, 1962. Wyckoff, Richard D. Stock Market Technique Number One. New York: Wyckoff Associates, 1933. Zieg, Kermit C., and Perry J. ICaufman. Point-and-Figure Commodity Trading Techniques. Larchmont.N.Y.: Investors Intelligence, 1975.
Relative Strength Index (RSI - Relative Strength Index)
Wilder, J. Welles, Jr. New Concepts in Technical Trading Systems. Greensboro, N.C.: Trend Research, 1978.
The slow stochastic oscillators (Slow Stochastics)
Elder, Alexander. «Using Stochastics to Catch Early Trend Reversals.» Futures, June 1987, pp. 6870. Kaufman, Perry J. The New Commodity Trading Systems and Methods. NewYork: John Wiley & Sons, 1987. Lane, George C. «Lane's Stochastics.» Technical Analysis of Stocks and Commodities. Investment Techniques 2, pp. 87-90. Murphy, John J. Technical Analysis of the Futures Markets. New York: NewYork Institute of Finance, 1986. Schirding, Harry. «Stochastics Oscillator.» Technical Analysis of Stocks and Commodities 2, pp. 94-97. Stein, Jon. «Learning to Swing to Momentum with Stochastics Signals.» Fu-tures. May 1989. pp. 36-37.
Volatility (Volatility)
Kaufman, Perry J. The New Commodity Trading Systems and Methods. NewYork: John Wiley & Sons, 1987. Notis, Steven. User's Manual for the Professional Breakout System. Nesconset, N.Y.: Byte Research and Trading, 1989. Wilder, J. Welles, Jr. New Concepts in Technical Trading Systems. Greensboro, N.C.: Trend Research, 1978.
147
Chapter 3: Testing the system Basics Why should I test? We are often asked: "Why should test the trading system? After all, you only get a hypothetical data. How did you learn how the system will work in real time?" The real answer to the last question is that you will never know for sure whether your system to work in the future, but there are only two ways to find out if it has at least some potential. The first way is to trade on the system in real time, and the second - its testing. Since the exorbitant cost and long-term testing of the new trading system in real time are not valid, then the computer test gives a simple opportunity to see how the design of your system would work on historical data. You know its positive and negative features, and if testing is correct, you will learn what to expect when trading in real time. In addition, you achieve two things. First, to test the system, you should do it mechanically, and the only element that requires your intervention, the question will be: "I will go I'm in the following trades or not?" We think that the mechanical trading systems are best suited to the vast majority of traders. Second, you find out whether your system that Ralph Vines in his book "Portfolio Management Formulas" (see the readings at the end of this chapter) called positive expectation. Others call it the "benefit of the trader." This may sound simplistic, but if the h-estiro vanii your system is not profitable, it will not be profitable in real time. If you trusts only competent management of funds, there may be an illusion that a mediocre trading system can be transformed into a winning bids by various management and cash. This is not true. No options of cash management will not turn a losing system into a winning. You can use any strategy gaming, and if you do not have positive expectations for an extended period of time, the management of funds will not affect your results. You must possess give an advantage initially. The only way to ensure that you have the advantage of this, is to test your system. Do not expect to collect comprehensive results, saying that this or that trading system or study surpasses all others. We also hope that you will not be overly disappointed if, in spite of following all the rules and proper testing, will not succeed. Each of the tests ever conducted has weight only in the context of the internal structure and in relation to data included in the test. If we test the trading system on 20 different markets since 1980 to date, the results do not say • about the year 1975. More importantly, they are limited in that they can be said about the future. We do not "prove" and should not be expected to prove it. The best thing we can do is to be accurate and careful in our testing. There are no simple answers. In futures trading nothing is certain.
Software 148
Computer testing of mechanical trading systems was in the air for 20 years, but has gained popularity in the last few years, when they were presented with software packages PCs that have made this process possible without the mandatory presence of extensive programming skills. Almost any can now create a "profitable" trading system, showing fantastic hypothetical results based on historical data. Most trading systems sold for big money in the last 20-some years, have been created in this way. Unfortunately, as can testify to thousands of disappointed investors, commercial system - the "black boxes" are rarely, if ever, used to work as expected of them. As far as we know, the same can be said of the vast majority of trading systems ^ created by the testing software. The reasons for these failures are not obvious in the testing software, and methods for testing and evaluation. For our testing, we use the System Writer Plus and Computrac / SNAP. Both products are excellent programs that we can recommend without hesitation. None of them shows the error or lack of flexibility which would apply force researchers inadequate testing methods. The problem, as we have emphasized, is to test methods, not the software. We understand that not everyone has the software for testing. It is expensive and complicated, requires a long acquaintance. To create a viable trading system does not necessarily have such a program. The elements of the trading system will be similar regardless of whether you have expensive software, or you program yourself. You can even test the trading system without using expensive software if you just sit in front of a computer monitor and objectively to trade on your system by past data. In fact, this approach has some advantages. One of them is the fact that, without extensive optimization opportunities inherent in software packages, your system is less likely to prove to be overly tailored to fit the curve of prices. However, the benefits of software packages greatly outweigh their disadvantages. We strongly recommend that you either purchase the software for system testing, or write your own if you have the necessary skills and a strong belief that commercial systems do not meet your specific needs.
The elements of the trading system Perhaps we repeat, but I still would like to list the attributes that should have a trading system that successfully tested. Some (or all) of these items may seem obvious, but we are in the process of publishing our leaflet and writing this book for the last two years, talked with literally thousands of traders and can say with some degree of confidence to say that many people do not fully understand what should and Why should not make the trading system. The first necessary quality of the trading system to conduct a serious test to be her absolute automatism. The only element that requires intervention should be the decision whether to trade entirely on this system or not. All other solutions should be built into the system. We understand that many, if not most of the speculators are now traded on systems that are at least partly require manual intervention. We also know that most of the truly successful traders consider the personal intervention as a necessary component of trade. We're not going to challenge this position, but such a system is quite impossible to test, because it is hopelessly subjective. Elements of the personal intervention of the trader has no place in a mechanical trading system, which is prepared for testing.
Expect the worst 149
The system you are designing a test, try to take into account all the surprises. We often hear about traders, justifying the absence of certain elements in their favorite shopping engines. A typical response is: "For me this has never happened, so why to prepare for it? Why should I use stop if my system is always catches the peaks and troughs, and there was no loss that I could not afford it?" It is not just naive, but dangerous. Always assume that what can happen, will happen. The system should always expect the worst and be prepared deal with it. You should always strive to achieve full control of risks. Do not think that if it had not happened in the past, that will not happen in the future. Here is a typical example. Many traders prefer ispoluvat stop only at the close "to avoid loss antsy when stopping the intra-day trade. This can be explained by the fact that the intra-day price movement is inconsistent, and only close the matter. (In fact, stopping only at the close of commonly used by traders who were stopped at a potentially profitable trades, and believes that this will not happen again.) Can you imagine such a fool, who was on a long position in the S & P 16 or 19 October 1987, with stops only at the closing? Just the possibility of such situations should convince anyone that it is better to have regular stops and the strategy of re-entry. Expect the worst, then you will not be caught off guard.
Dangers optimization Unfortunately, in some way misrepresented the purpose of testing the many projects is the development of trading strategies, squeezes the maximum possible revenue from the historical data. The assumption is that they work better in the past, the better they will function in the future. If this were true, then everyone who has developed a historical model that works well in the past (meaning almost everyone who has ever used a software system testing) would be by now very rich. Obviously, this did not happen with the majority of traders. Typically, the test is as follows. The trader buys the most modern software and the necessary data. He puts together a few favorite technical studies or graphic shapes that have proved effective in the past, and riding the computer hour after hour in search of the exact values for each parameter, which would give the highest return. Impressed by the fantastic results greed wins, and he or she will immediately begin to trade, the trader will inevitably falls into the band losses and decides that something is wrong with his trading methods. The most obvious solution is to optimize the system again and get rid of the losing components. Trader optimizes again and was pleased to notice that the changes have removed most of the losing trades. Again sure despite previous losses it renews trade only to obtain a new series of unexpected losses. Many traders continue such efforts as long as they do not run out of either money or patience. Usually sum money, leading them to believe that they certainly would have done better if only they could afford overoptimize system and enter into the trade once more. Other traders conclude that the error lies in the use of automated trading systems, and they pereklgoprepared on subjective methods of trading, which can not be tested. The loss, of course, continue, but now they are simply the result of bad luck, poor execution of orders, triggering stops, manipulation insiders or lack of attention broker.
150
What's really wrong? It is useful to analyze the above process and see what went wrong and why it happened. The first and most obvious point is the questionable value of optimizations in almost any form. Any set indicator or indicators show huge income being optimized for the best combinations of parameters even when using a random data set. Computer analyzes millions of combinations, so there is a very high probability that some of them, at least in hindsight, will make money. When faced with temptation is almost instantaneous enrichment, as evidenced by the encouraging results of optimization, the temptation to immediately start trading becomes irresistible. Faith in the optimization process is so strong that traders will be optimized again and again, even though the state of their accounts would have to tell them that they are doing something wrong. This happened with our trader in the previous example. You can hear a trader says, "Just one more optimization, and I'll do it." Unfortunately, another optimization will never solve the problem.
Optimize or not to optimize Anyone who believes that the full optimization works just as well, as advocated by some system vendors, would not hurt to read "The Usefulness of Historical Data in System Parameters for Technical Trading Systems" Louis B. Lucas and B. Wade Brorsena. Their work is systematic and complete. They tested the trend-following system, a breakthrough channel and Wilder's directional movement system using 20-year data. The only variable that are optimized to be the number of days used at each calculation. This parameter is passed through the time period from 2 to 60 days by 5 days. They compared three different optimization schemes with random test which used the random values of the parameter set from 5 to 60 days. The most significant discovery was that the re-optimization strategy gave nothing in terms of performance. Each of the optimization methods give results that differ significantly from the results of a random test. With or without the use of optimization, revenue was in the region of 50 to 60 percent for the channel breakout systems and 30 to 54 percent for the system of directional movement Wilder. They are formulated as follows: "The results of all tests Govoryat that the predictive optimization capabilities are limited. Optimization was not able to predict a set of parameters that would give the best return on the portfolio than the strategy of random selection. " Let us emphasize that this was strictly a formal test, conducted with great attention to detail. Anyone who claims that the full optimization works better than a single-blind simulation to face directly opposite results, which have just been demonstrated.
How to avoid the adjustment to the curve Some adjustment to the curve is inevitable. It would be difficult and undesirable to develop technical study without it. When a trader is drilling through the eyes of the schedule and sees that the 9-day RSI, seems better suited to this particular market than the standard 14-day, he adapts to the curve. Since it seems simple and effective, there is only one step to test each parameter RSI. Once this process begins to profitable results, permutations are almost endless: "We'd better add some more technical research, to be sure we did not 151
miss anything. While we use this system, let's optimize it for the correct initial risk and better tracking stops to make it as complete. " The end product is a system that encompasses all the best motives and fitted under the curve in the p-noi degree. Despite the fact that it looks good on paper, the odds are against the fact that it will work in the future, become astronomical. The optimization results are just the opposite of what would seem obvious. The better the system looks and the more complete and complex is, the less likely it will succeed. There is a strong explanation for why optimization and curve adjustment to give poor results. Frankly, it's that simple concept that we can not understand why many traders do not pay more attention to it. Each statistic known concept of loss of freedom. In layman's terms this means that each of the parameters added to the trading system, is a loss of control over the degree of impact the final testing procedures. Than more technical research or trading rules you enter, the less healthy and reliable the results. The more you try to improve the system, the less likely it will work as well as in testing. You should have two to five variables. The fewer variables, the more reliable the results. An interesting consequence of this approach is that it allows you to look at your own work and to quickly understand whether it is fitting to the curve. The probability that a system will be tailored to fit the curve depends on the number of variables used in the test. The greater the number of technical studies and rules (especially the exceptions to the rule), the more the model is adapted to the curve. Beware of systems that are so complex that they require a computer to work with them. Another way to avoid the adjustment to the curve - refusing to create systems that are configured for specific markets. This is a trap that just to get into, and it is also the basic principle of adjustment to the curve. A good system is not required to operate historically in all markets, to be successful, but it should work in most markets with a small number of changes from market to market. If you need to change the system in order to adapt it to each market, then there is a serious flaw in the basic system. We are well aware of the argument that each market has its own unique character, but we also remember the times when currency futures were little volatile, and times when they showed fluctuations in the value of contracts for thousands of dollars a day. Markets change, and the best way to achieve the confidence that your system will go with them in the leg, it will be tested in the same form in the widest possible range of diverse markets. Before we leave this subject, we note one more subtle form of optimization. We're talking about practice sweep of historical data through the computer to find the "seasonality". There are a handful of famous traders / authors that provide test data demonstrating that, if you were buying a specific product on a specific day each year and sold it to another particular day, you would have increased your income x times. It's just nonsense that has absolutely no statistical meaning or use in trade. If we want to, think of your computer will allow us to optimize the data instead of the system. These are considered very small segments for exact dates, which are best suited to the system. Instead of fitting a curve of the system, we can fit under the curve data. Of course, there are many obvious logical and sometimes suitable for long-term use of the season (for example, the annual fall in prices during the harvest), but beware of the season to bring the following to the point of absurdity. Any recommendation for the seasonal trade, more specific than specifying the desired month to trade is to be treated with great suspicion.
Selection of the test period
152
Another important and often under-appreciated area is the choice of the period of the test data. At least during the test period should be long enough for at least 30 on each trading market. Getting less than 30 trading violates one of the basic rules of sampling theory, which states that there must be at least 30 data points to data set corresponding to the normal distribution. Note that this applies not days, weeks or months of data, and trading occurred. Any number less than 30 will produce statistically unreliable results. The more successfully, the better. It is equally important to market periods that you are testing, included as many examples of all kinds of market conditions. The up, down and sideways are the simplest (albeit subjective) examples of possible market conditions. The test period you should include as much as possible pain-1 above examples. Our goal is to simulate the possible conditions of the future to include the maximum number of market conditions of the past. If the test period is represented by only a few years of data, this can cause problems. For example, if the stock market had a serious period of falling prices, and thus on the data presented in futures on stock indices also had serious falls, the testing of such data will favor a bullish bias systems. Throughout its existence the stock market indices do not provide enough data to meet future market conditions. The oil market, on the other hand, we showed diversity in a much greater degree and can be expected that its data can be made more robust and healthy trading system. Let's clarify it another way: the results of a short period of testing to the crude oil market data can provide more reliable results than a longer period of testing on stock indexes, because the data while stock indexes contain a clear upward slope. The system, based on purchases in the market stock index is likely to give better results than the system only sales. However, as Yogi Berra once said: "The future is not repeated." Browse consequence is that the system should never have a slope of the sides of the market. Obviously, with a few exceptions, attractions, a large portion of the proceeds in the stock index will fall on the long side of the market. This does not mean that the trading system should give priority to this side. The system should not have an opinion or bias in which whatever side of the market. If this seems obvious, remember that in the 70s most of the income in the commodity markets has been obtained on the long side. Many trading systems developed during this period were essentially systems of the bull market. The easiest way to improve your results in this period was to reduce or eliminate short positions. We suspect that this was a bullish bias principal cause of the poor performance of many advisers in the commodity markets in the early '80s. Our conclusion: there is no strict definition of what kind of data should include test. If we assume that the average trend following system trades about once a month in each market, then at least three years should be taken as the minimum period for testing In order to test the primary produced at least 30 bidding. Then add two or more years for the advanced test (we will explain this later), and you get five years, which is usually acceptable minimum. Add more time if the market was not diversified (falling, rising, lateral) in the target period. You must include in your research as much as possible a variety of market conditions. We prefer to use a large amount of data and to test for different time periods. As long as you you do not make it yourself, you will never fully appreciate just how illusory can be profitable trading system, and how the test results depend on the choice of time period. We are very wary of systems that have not been tested by time periods, corresponding to a representative sample of market conditions. Note in Table 3.1 on the results affected by the change of time intervals, especially when it relates to losses. The return reacts similarly, that brings us to a curious thought. In carrying
153
out all the procedures of optimization / testing, we have focused on the aggregate returns as the only
criteria for choosing the optimal parameter for use in subsequent tests or in real-time trading. In our simple example returns roughly correspond to each other. The losses, however, are significantly different. How many traders want to nayayuchit to a loss of $ 10,000 in the trade contract with an average margin of about $ 2,500? The example in Figure 3-1 illustrates one rarely mentioned hazards testing in general, and in particular the optimization. When you test for improvement sheniya only one result (usually co-benefits), you are ignoring other important data. We recommend testing for a series of parameters, not just for one. We understand that this complicates the process and makes it largely subjective, but testing only to improve the co-benefits often leads to the right path and can be dangerous to your financial health.
Selection of data for testing To our knowledge, there is no commercial test software, which includes the opportunity to conduct trade from one contract month to another without having to cause a tear in the values of any technical study which it calculates at this time. The gap makes the defective study and therefore the test itself. Of course, it is theoretically possible to provide a series of computer contract months on this product, test each month separately and then combine the results, but we can not imagine a more tedious and error-prone process. The solution lies in the location of your data continuously without breaks, which would enable continuous testing. We will not delve into the details of how to ensure a smooth transition from one contract to another, but we are content with this, provided that the results are obtained in a reasonably accurate. We have data from two sources: FutureSource and Technical Tools. Our data Technical Tools came with the software to create our own long (or otherwise) contracts. If you have a number of different analytical software packages, Quote Butler of Technical Tools is a great way of throwing the data from one package to another without the need to buy data separately for each application. We have not talked to other data providers, but there are few reliable sources of inexpensive and reliable data.
154
Slippage and commissions Do not trust the test results that do not include some freedom for slippage and commissions. These components are very significantly alter your results. A lot of trading systems provide a small steady income when testing with no tolerance for slippage and commissions and to losses on when the transaction costs. This is especially true for shortterm day trading systems. The more often sells the system, the more important is the cost of transactions. A striking example is the recent publication in the press. The article described the indicator, which was to respond to the twists and turns intraday futures stock index. Despite the fact that trading volume was large, had to make allowances for the cost of transactions. We calculated that, taking into account a very small fee, and only rare slippage, the system will break even at best, and at worst - constantly losing money. Everyone has their favorite numbers for transaction costs. We allot $ 75 and $ 50 slippage and commission per revolution of the contract, for a total of $ 125 on trade. These numbers may seem high, but we prefer to err on the worst. When we test of the trading system, we can deliberately blind eye to the slippage and commission to simplify operations, but we will make sure that they are involved in the final test.
Types of testing * We will explain some of the most common schemes optimization and testing.
Simple optimization It is as simple as it sounds. You create a trading system and then optimize it for full parameter values as long as he does not find a set that gives the best return. From our point of view, it is the least efficient method of testing the system. This adjustment to the curve at its worst.
The cumulative anticipatory testing This is also called "run" of the optimization. The cumulative anticipatory testing requires that you have optimized the system to run at the beginning of your data, and then tested the results on the relatively small follow-up station. Then you have to overoptimize for the period including both sets of data, and repeat the cycle. For example, if you have 10-year data on T-Bills, you could optimize for the first 3 years, and then test them in the next year. If the results are still good, then you need to optimize for all four years and tested in the fifth year, and so on. This is a form of optimization, which has been tested by Lucas and Brorsenom, and they found no better than a simple optimization (see above).
Simple testing operezhayushee *
155
This method is also called "blind modeling" or "test sample outside." You are developing your system for initial data (say, the first 5 years of the 10-year data set) and then testing without changing what you consider to be your best combination of parameters and rules on a more recent time period. If the results are not satisfactory, the process is repeated. Optimize or not in the first phase of testing, it is not nearly as important as the retention of the number of variables on a small level. The most important thing that any trading system which is subjected to a simple testing or optimization without advance testing is likely to be doomed to failure. Forward-testing is the most elegant testing system. If your system has not proven to be profitable for the advanced testing procedure, discard it.
Measuring Performance The obvious purpose of the trading system is profitable. How much money does your model? Another way to calculate this factor is the percentage of return, which is the annual returns, based on the amount of funds required for a trading account. Rate returns to be found throughout the testing period, and then break it down into small segments to isolate the negative periods. Keep in mind that the percentage of return - it's just a function of the amount of capital used. You can double the percentage of return if you start trading with only half the capital, but in this system you do not improve. You can improve the system, starting with a great deal of capital, but the percentage of returns will be less. Frivolous and mindless shopping competitions are won by getting a big return for small amounts of seed capital. These revalued trade events result in non-viable or play around with the historical record, which becomes apparent when we see poor performance of commodity mutual funds managed by some famous winners of such competitions.
Sharpe ratio (The Sharpe Ratio) A popular measure of performance derived from the percentage of return is the Sharpe ratio, developed by William Sharpe, and is defined as the annual yield (a measure of profitability) less impervious to risk ratio of return divided by the annualized standard deviation of returns (a measure of volatility). Some remove impervious to risk ratio of return, so check it out before comparing results. The higher the Sharpe ratio, the higher the returns and lower volatility. Typically, commodity trading advisors calculate the Sharpe ratio for monthly data. We understand that the Sharpe ratio has its limitations (for example, an increase in volatility in the upward direction will decrease the Sharpe ratio), but, nevertheless, it is the most common index of this kind. Maybe you will be useful to compare the results from one system to another and compare your results with professional advisors. The best system is not the one that makes the most money, and the one that gives a higher Sharpe ratio.
The ratio of Stirling (The Sterling Ratio)
156
Because the Sharpe ratio has certain shortcomings were developed other statistical techniques for impartial evaluation of performance. The most popular of those is the ratio of loss to the size of the impact. The ratio of Stirling was created Dianna Jones of Jones Commodities. The formula is: The main disadvantage of Stirling relationship is that it is usually calculated annually and hence too slow to respond to changes in performance.
The ratio of Kalmar (The Calmar Ratio) The ratio of Kalmar, invented by Terry Young of CMA Reports, is the size of return over the past 36 months, divided by the maximum loss in the same period. It is calculated on a monthly basis, making it more sensitive than the ratio of Stirling
Geometric mean Perhaps the most mathematically precise measurement of the capacity of the trading system is the geometric mean of Ralph Vines. The geometric mean measures the growth factor of your trading system. The higher the geometric mean, the more likely that your system will yield a greater return for additional investment. For any system with a geometric mean of more than 1, you can increase the return on your account up to the maximum by calculating the optimal f, the optimal fixed part of your greatest loss, for use as a bet on each trade. We do not have room for the geometric mean and the optimal f, we also can not explain everything with the same elegance as the Vines. We believe that this book is one of the most significant achievements in the area of cash management in the futures markets. Note that your system may be profitable in most markets and losing money on some of them. One of our friend in the commodity trading advisor conducts operations in all markets, which he tested it (winning and losing), and argues that the curve changes its account becomes smoother from this diversity. He deliberately looking for a negative correlation between the products in the portfolio, and finds that profitable periods in his losing markets usually coincide with periods as lost his winning markets. The trading system will not be profitable in all the markets constantly. If you do it correctly developed, the losses on the losing markets will be minimal, and, in addition, these markets will, from time to time to give profitable periods. Be careful when testing a large number of markets, and then design portfolio alone winning contracts. This is the usual method of vendors, when the results are pure fantasy, despite the fact that it can produce an impressive record of your historical performance. This, obviously, is another form of adjustment to the curve.
Testing for specific results Trading system should be developed from scratch to attain a certain performance criteria values. Probably the most important of the possible targets are the winning percentage 157
and the ratio of average win to average loss. These criteria can be used to calculate the probability of failure (FOR - probability of ruin), which gives you some idea of the reliability of your system. Much of the software for testing gives other useful data. Below is a listing of their comments.
Comprehensive income (Net Profit) Comprehensive income is revalued measure of success for several reasons. First, make sure that a few large trades have not rejected your results. You do not want to use the system, the success of which depends on the non-recurring events (such as Hunt's corner the silver market and the fall in sugar futures at 63 cents). One should not think that, in real time your system will play anything like the aggregate income of your tests. Wrong plan future results based on past - we know that they do not recur. You can only make sure that your system is ready to cope with most of the predictable future market conditions.
The number of trades on the test sample (Number of Trades in the Test Sample) The total should exceed 30 for confidence in the statistical significance of the results. Even if you've tested the 25 years of data and have not received at least 30 trades, your results are inconclusive. We once listened to a lecture about the health of the stock market indicator, which produced one trade every 40 years. We would like to see the test results for 1200 years for this method to impress us. The more you have got trade, the better.
The highest winning and losing the largest trade (Largest Winning and Largest Losing Trade) The greatest benefit is important if they reject comprehensive income to the unreasonable amount. Many conservative traders will throw out the highest win on each market and recalculate the results. The greatest loss is especially important if it exceeds your normal measure of risk control. Perhaps there is some problem or unforeseen expenses that you have not noticed. Be careful with the removal of the largest losses. This is where most traders stumble upon adjustment to the curve. Do not develop special rules to circumvent the big losses, just review your sequence stops. Controlling the maximum loss can be especially important if you are trading on the basis of aggressive reinvestment using the pyramid in any form or by more subtle methods, such as the optimal f. You should avoid surprises that could devalue your entire strategy.
The maximum number of consecutive wins and losses (Maximum Consecutive Winners and Losers) The maximum number of consecutive defeats can be useful. This gives you an idea of the emotional experiences that you will face in the course of trade on your system. The prediction of this number can prevent panic when such an event does occur.
158
Losses from peak to trough (Peak-to-Valley Drawdown) Very important and, despite this, often deprive the attention quantity is the loss of a percentage of the invoice, measured from peak to trough. The system, which produces an annual percentage of return of 100 percent in five years - it is a system that will be hard to follow if it admits losses from peak to trough of 50 percent a few times within five years. You need a lot of courage and deep pockets to trade on such a system with confidence. In our experience, a smooth curve is much more preferable and more difficult to achieve than a high annual yield. This curve is very important as a measure of how your system will be practical when on the line will be put real money. More often than systems that generate the most comprehensive income have the highest losses. Combine the heavy losses from a losing streak and you will understand the reason why most traders prematurely abandoning a potentially good trading systems. We emphasize again, the system should be developed taking into account the stress barrier for each individual trader. To a large extent, as the expectation of consecutive losses, the expectation of potential losses that we have to endure, can provide a vital element of trust that will allow us to survive the inevitable losing periods. For professional money managers, there is another reason to calculate the maximum loss. Consultants merchandise trade say that the public is getting smarter (eventually) and more interested in those rare consultants, historical records which show a steady growth and a small loss than a high-flying birds that exhibit large short-term gains with big losses from peak to trough. Those of you who are going to become a commodity trading advisors (CTAs-commodity trading advisor) and manage the public funds, will do well to developing a portfolio system with the largest loss from peak to trough (with daily measurement) lying below 20 percent. This requires a combination of good money management (including the correct capitalization) and stable trading system with controlled risk. Some commercial software considers the loss as the total size of the account minus the open part of the bill to lose. The rational explanation of this approach lies in the fact that the open revenues will eventually turn into a private income, and therefore the risk of profitable positions will otsutstvsyat. This is not true. Futures positions are adjusted daily by the market. Rarely, if ever, lucrative trade is closed at the peak of profitability. Part of the bill, which is returned to the market in generating income and which is refundable upon receipt of losses deducted from your trading account in the same way. The most accurate way to calculate the maximum loss is the difference between the peak value and the corresponding total Padin day bills. These estimates reflect what really could happen to your trading account. Any other methods of calculating damages only misleading. If your software does not provide the appropriate calculations, print the data on daytime aggregate score and perform subtraction alone. If you can do this operation for each market, and then on the total portfolio, if you se tested. The results can be quite revelatory. Do not throw away your test results. They provide an early warning system that will alert you if your system will self-destruct in real life. Any results approaching the maximum loss, or the maximum number of consecutive losses, should be carefully studied, along with any drop in winning percentage or ratio of wins to the loss. Here are some basic guides to help you choose the purpose of testing.
Winning percentage (Percent Winners) The most successful traders use trend-following method is obtained from 35 to 45 percent of the winnings. It is difficult to get more than 55 percent, you can be sure that we will 159
conduct this testing. Be especially attentive to the impact that stop by the percentage of winnings.
The ratio of average win to average loss (Ratio of Average Win to Average Loss) This value must be greater than 1:1 (break-even). Obviously, the value of 3, 1, or 4-1 looks good, but if you get a modest percentage of wins 2:1 or even less, it will also bring you enough income.
Obshaya efficiency and maximum loss (Total Return and Maximum Drawdown) These values are inherent in the contract or the portfolio, expressed in dollars. For example, the total return on the S & P contracts should be compared only with the maximum loss on contracts S & P. Total return and the maximum loss is theare a wonderful expression of the risk / reward. Loss of both of them is more important. Both values can be expressed as a percentage of the margin, but the margin - this is a moving target, which is associated with the contract for months and can often change suddenly, so it can not always give an accurate measure.
Volatility and the likelihood of failure (Volatility and Probability of Ruin) The calculation of these two key variables will give you an idea of how reliable is your trading system in real time. The first value, which must be calculated is the standard deviation of your trading results. The greater the standard deviation, the more volatile will be your trading results. The smaller the standard deviation, the results will be less volatile. Other things being equal, select a system with less volatility (standard deviation of the minimum) in terms of individual trading results. The second key value - the probability of failure (POR - probability of ruin). POR trader gives the percentage likelihood that a balance of the account will sink to a certain point before a certain rise to a higher point. In the calculation included six variables: the percentage of wins, the average lucrative trade in U.S. dollars, the average loss in the U.S., the initial account balance, the level at which it can be said that the bill failed, and the level at which it can be said that the state of the account successfully. POR is based on the idea that any trading system events occur regularly and may experience abnormal events, but they all obey the theory of probability. For example, a coin toss is an infinite amount of time will be related fallout heads / tails 1:1, but about one in 1024 cases orlel will drop 10 times in a row. Any trading system, so live with the possibility that regardless of changes in market conditions, it is to some extent selfdestroying. POR - Is the probability of such self-destruction. The level to which we can control the percentage of wins and winning attitude to the loss dictates the degree of control that we have over our trading system. We do not have the ability to monitor market conditions, but we will at least be sure that our trading system will not self-destruct on my own. There are several ways to calculate the probability of failure (or risk of failure). We 160
will use a simple formula that can be found in P. Griffin "TheTheory of Blackjack". The following table shows a representative set of values and attitudes winning percentage of average income to average loss. To simplify the calculations we made the initial account balance is equal to $ 25,000, the purpose of income to be $ 50,000, and the level of loss (failure) equal to $ 12,000. (See Figure 3-2.)
As you can see from the table in, FOR dramatically changes with the percentage of winnings and win / loss ratio. A small correction in the system, which results in a positive change of any of the relationships, can make a huge contribution to the future of the system. POR can be very revealing. For example, the average CTA, managing public shares today probably has a winning percentage of 35 to 40 percent, with most of them less than 40 per cent, 35 per cent share of wins requires a high ratio of average win to average loss for the system to be successful. It is perfectly feasible and just when the markets are in a state trend, but when they become restless, the ratio of income to the loss drops sharply, and FOR rises to frightening heights. Careful monitoring of these two statistical variables and possible change your trading system to account for the non-trending markets, it may be necessary to ensure survival.
Testing of entries, exits and stops Once you have picked up the items in your trading system, it is tempting to immediately test it as a whole. Among other things, the system composition and the definition is a set of interrelated parts, and it seems reasonable that it should be tested in the complex. The problem with this approach is that one element of the system can improve or degrade the results relative to the other. You can have a wonderful method of matches, however, if you have weak output, it will reject your entry along with the rest of the system, if the results of aggregate productivity does not meet your standards. Also good to know how often you use one element than the other. If your risk control consists of two types of stops, the simple dollar stops and a recent peak or trough, it is in the process of development and testing will help you knowing how often included each of these stops. For these reasons, we have developed methods that share of the trading system and let you test them independently. These techniques are not ideal, because the parts of the system are often inextricably linked, but we find them very useful. Note that any demonstrated by our test results are not statements about the superiority of one method 161
over the other. We are trying to show a process, not an arbiter in the debate about which is better technical studies. We strongly recommend that you spend these procedures independently in the context of their own needs and come to your own conclusions.
Testing of occurrences Testing your favorite methods of occurrences can be very revelatory (and painful). We are all too often convinced that many of our cherished assumptions about the proper way to enter the market turned out to be, at best, mediocre. When you become an expert in the test. ing systems, you will probably find that the importance of entries is reduced and that the way you get out of the market, it becomes more important. All that you can claim from entering it so that it gave you more than a casual income potential. Once you got it, just from your exit strategy depends on whether you can catch as much income as possible while keeping losses to a reasonable level. One of the most important statistical parameters obtained during testing of the systems is the percentage of wins against the losses (% of wins). All things being equal a high percentage of wins, obviously preferable to a low percentage of winnings. Fortunately, if the ratio of the average income to average losses are installed correctly, it can bring on long term profits, even if the winning percentage has fallen to a very Malfoy value. Most long-term traders manage to survive, catching here and there a very large income, and come to a total of only 35 - 45 percent of the winning bidder. The problem is that, despite the small losses and big profits, the smaller the percentage of winnings, the more volatile will be trading results. To a certain point jump from the top of the account balance to the basin will become unbearable for all but the most traders with strong nerves. With even more of a challenge faced by day traders who need to develop a method that wins more than 50 percent of cases. These traders can not let their revenue stream, because they have to go before the market closes. Their attitude transaction costs to income from trading is usually very high and extremely difficult to maintain the ratio of average win to average loss rate of more than 1:1. No matter whether you are a short-term or long-term trader, you can not get a high percentage of wins without a proper entrance into the market. Despite the fact that the total scheme of the release of important occurrences (among other things, that ultimately determines the output return trade) is much easier to find good yields when the entry were made correctly.
Testing Methodology occurrences. The best way to effectively test every single element of the trading system - to isolate it as much as possible. However, the isolation of the elements of the trading system is much more complicated than it might seem because the trading system, by definition, consists of a set of interconnected components. Changing one component, even a small amount can dramatically change your trading results unexpected and unpredictable ways. We often do elaborate minor changes in our review system, which produced an unimaginable mess n'ashey carefully established strategy. Once this has happened numerous times outrageous, we decided to use a step by step approach that would allow us to isolate the occurrence of other elements of the trading system. We do not claim that this is the only or best way to test matches, but it seems logical, and it worked for us.
162
The method is simple. Customize your trading system and then remove your normal exits. Replace them with a method that automatically comes out of the market after a specified number of days after entry. In our testing, we usually looking for entry signals that would put us on the right side of medium-term trends. We have set out the test after 2,10,15 and 20 days. Selection of days gives some insight into the strength and direction of the market after entry. For example, if your 5-day output makes a small percentage of winning trades, while the 10-day yields give good results, we can conclude that we have some room for improvement at the set time that our direction seems right. If the five-day yields tend to perform better, we could have gotten a good method for occurrences of short-term trade, but not for our long-term objectives. You should set time periods for days out to suit your own style of trading. For example, you might find valuable test yields after only one day, or you might decide to go out after 30 days.
Analysis of test results We want to test showed us the win percentage, based as far as possible, on entering the correct direction and set the time of entering excluding other considerations. Remove any slippage and commission costs and do not use stops. It will still be ample opportunity to include factors related costs, when in the future we will collect comprehensive system for testing. Since you do not use a stop or any other real outputs, indicators of aggregate income received as a result of this type of test, in fact, do not make sense. Now we have to consider the income and loss of chance as the market. The same is true for the relationship of average income to average losses. When we are done collecting our integrated system, the current stop and lock outs income will determine the values of these variables. At this stage it is important to compare the value of occurrences is the percentage of winnings. If the method of entering standing, he should introduce you to the market in the right direction with a win percentage is much higher than random. We're not going to sink to the statistical test of "significance", but still give a rule of thumb, the entry is considered to be better than chance, if it is profitable to at least 55 percent of the time on a range of markets. Also, if you try to follow the trend, trading should show a rising percentage of winnings with increasing time interval. It is extremely important that any method you selected entries produced better results than a random occurrence, because the addition of stops and try to allow income to flow percentage of winnings will certainly be significantly reduced. The better your original winning percentage of occurrences, the closer can be your stop. If you prefer a relatively distant stops, you can achieve a higher percentage of wins at the expense of increasing the risk on trade. It is impossible to regulate the trading system by altering its basic elements, and see the results of the next iteration, but we found that it becomes a lot easier if you work, relying on precise knowledge of how effective your entry, and thus clearly see the result of your methods output. If you start with the correct method of entry in 75 percent of cases, and he slipped to 30 percent gains after adding all the necessary stops and exit strategies, you can continue to work, correcting only the imperfect elements of the system, instead of modifying entries. Most traders are looking at the system just described, would put the responsibility on the entry. Unfortunately, the occurrence of collecting praise and blame for the results of a system that is completely wrong. Our independent testing occurrences 163
allows us to understand how much praise or charges, they really deserve. It seems that most of the testing of technical studies carried out on the basis of the back when the same indicator is used for both entries, and for outputs. We will not bother demonstration of these useless test results. The study may be fine for some tasks, but we can never know the results because of the inability to isolate his ability to perform each function.
Testing Procedures All of the test results, as demonstrated in this section refer to the period from January 1986 to December 1990 (five years). Tests using daily data, namely the discovery, peak, trough and closing. Occurrence took place on the same day in the case of the channel break or receiving a signal from a system based on volatility. In all other cases, the entry was performed at the opening of the next day. The stop is not used. Slippage and commission have been taken to be zero. Output occurs at the close of a 5,10,15 or 20 days after the occurrence. We tested five markets: the German mark, gold, soybeans, Treasuries and crude oil. We limited our testing of these markets because they represent five different product groups and provide a basic understanding of the performance of occurrences. If we wanted to find the indicator, which gives extremely good results, we have expanded the survey to test a larger number of markets. Favorites us to test methods of entering are popular strategies which we hope you will find interesting.
Crossing the middle skolzyashih Crossing moving averages, probably still is the most common technique occurrences. However, the smoothing effect, which makes them attractive as occurrences is precisely what makes them relatively ineffective as outputs. We chose to test one of the standard works of moving averages: the intersection of 9-18. As with most studies, moving averages, long entry is generated when the 9-day moving average crosses above the 18day. Short occurrence generated when 9 18 intersects the top down. (See Figure 3-3.)
Channel Breakout We tested the main method that is on the market for a new daily peak or trough of the last n days. As an interim period, we chose 10 days. (See Figure 3-4.)
164
The intersection of the stochastic oscillator with borders * This is a standard counter trend entry: Buy when the% K crosses above% D after the oscillator has fallen below 25. Sell when the% K crosses% D down after climbing above 75. Intersections between 26 and 74 are ignored. (See Figure 3-5.)
The jump in the stochastic "Jump" stochastic oscillator has been described by George Lane. He is trying to use the property markets to follow the trend in the direction of the 14-period slow stochastic oscillator, which rises above the level of 75 or falls below 25, making this application stochastic trend-following method. (See figure 3-b.)
Relative Strength Index (RSI - Relative Strength Index) 165
Wilder's Relative Strength Index - is another popular method kontrtrendovzh. We used a common 14-day RSI, selling at overbought (above 75) and buying on the resale (below 25), (See Figure 3-7.)
The Commodity Channel Index (CC1 - Commodity Channel Index) Here we have used as an indicator of following CC1 zatrendom by buying or selling at the intersection of the zero line. The direction of entry determines the direction of the CC1. We used the 10-day RSI CC1. (See Figure 3-8.)
Moment '(Momentum) The study uses a simple computation time, which is the difference between today's closing price and the closing price n days ago. We used 10 days and were at the intersection of the zero line. (See Figure 3-9.)
Volatility (Volatility) 166
Occurrences narynok place throughout the day at the break in either direction on the p% of the n-day average true range (ATR - average true range). This type of occurrence has been the basis of many actively distributed trading systems. The first table (see Figures 310,3-11 and 3-12) shows the results of 150-percent breakout average true range over the last 10 days. The second table shows the results of 100-percent ATR breakthrough in the last 10 days. The third table shows the results of 100-percent breakthrough ATR last five days.
Random entry We could go on forever, but we think that at least one thing became clear: no need to use sophisticated analysis to understand that none of the above methods of entry can not significantly outperform the method of random occurrences. To prove this, here are the results of tests used a random number generator to make a decision about buying or
167
selling. The entries were made at the opening of the day, follow the release of ^ previous position. No technical studies have not been applied. (See Figure 3-13.)
The results were the kind that you would expect. Quite disappointing that none of the tested technical studies showed no results, substantially exceeding the random tests. A series of random occurrences confidently produce at times startlingly good results. These tests make you realize how careful you have to be careful, and the development and testing of the trading system.
The importance of the outputs We obviously do not recommend you to trade without stops, leaving the market through any number of days. Nevertheless, this exercise shows the complexity of finding the method of entry, that being separated from the exit strategy would give a non-random results. The task of entering, as we said, is to give us the probability of income greater than the probability of a coin on a certain side. Nuzhtyu you test your entries, and if your current method is not able to show the performance is better than just a random, then you'd better look for a new methodWhich is probably just as important, this study emphasizes the importance of the outputs. We argue that it is possible to create a profitable trading system using a combination of random occurrences and stops and exits income. It is impossible to construct a profitable system that uses a near-perfect entry and random stops and exits. Concentrate your efforts on what is particularly important is to control risk, good yields and cash management, instead of chasing the ideal method of entry. The next time you hear of someone receiving a great income, resulting from the latest high-tech method of entering, ask him about the strategies used by them to exit and write it down, because probably there lies the secret of success.
Testing outputs In the previous section we isolated and tested several methods of entry. As we have seen, test matches, isolated from the output, relatively simple, and the results are objective and easily assessable. Independent testing of the outputs is much more difficult. Since the entry and exits interact frequently unpredictable ways, a test designed to show us the relative merits of the various exit strategies will be influenced by the method of entry. We have tried to develop strength testing procedure that would give us some insight into the relative merits of various popular exit strategy. Despite the fact that we are not completely satisfied with the testing methodology, we believe that it horoshadlya enough to allow us to 168
objectively compare the different outputs under the same market conditions. The results were very curious.
Testing Methodology O To test can not isolate vyhodtak effectively as entry. The best method that we have managed to come up with is to test all the exit strategies, using the same simple method of entry. We picked up a method for entering, which gives reasonable results as the back of the system, and then tested each of the methods of entering the same data with identical occurrences. If each output method is tested using the same entries, we will be able to make valid comparisons of results. I must admit, the outputs of which work well with one system occurrences will not necessarily work as well with the other, but if the entry is a common characteristic as far as possible you'll at least get some idea of the relative effectiveness of different outputs. Similarly, testing of entries you can throw a fit by looking at the test results for your favorite entry strategies, and find that they are no better than a simple back of the system. We certainly were surprised by the many results of our testing. As a method of entry for our tests will fit almost any downside following system zatrendom. We selected the intersection of two simple moving averages (5 days/10 days, 5/20, 5/30 and 5/40) as the timer entries. The long entry will be signaled by a short-term moving average crossing from the bottom up over the long-term moving average and short entry will be generated in the opposite intersection. Outputs to benchmark performance will turns moving averages. At the opening of the trading day following the crossing of moving averages, we will close a trade and immediately initiate another, but in the opposite board. So we leave the point of entry as possible to directly compare the same outputs, we will trade only at intersections and never-between intersections. In our tests, we set the sensitivity of the methods set when the outputs so that the output will typically be generated before the next turn of the moving averages. If the output is not turned on, the trade will be closed crossing moving averages, and the new trade will be initiated on the same day and on the same opening price, as at the back of the system. Thus, the tests need to generate the same amount of the reference system trading market. Each output will compete with identical market conditions, including the same reversal of the master system. This allows us to compare one with the other outputs. Other things being equal will be easy to find an exit strategy with the best performance. One caveat: it's hard to find a solution that would always get us out of the market before crossing moving averages. The more trading is being actively used nerever outputs, the better for our test. We want as many of our exits were the result of the signal under test our method outputs, not reversals moving averages. Unfortunately, we can not get rid of the reference spreads, because then the number of trades and the date of occurrence will change dramatically from one test to the other outputs. Did we mention that it is not completely satisfied with our testing procedures. When we test entries, we can reasonably determine their effectiveness simply by measuring the percentage of winnings. In our testing outputs win percentage and other results are set almost at the same level common method of entering, as well as the fact that much of the output is indicated by the reference spreads, and not those outputs that are currently being tested. Since winning percentage has remained relatively constant for all tests (with one exception, which we'll talk), we need to spend an extra dimension to the performance outputs. The ratio of the average income to average losses will not be much point, since we are testing non-stop and without slips and and and commissions, statistical parameters that 169
actually shows the most substantial changes depending on the output is the total revenue. We have stressed that as a measure of the overall efficiency of the trading system of comprehensive income is not the best indicator. However, in this case, we'll use the aggregate income and the percentage of wins as values for comparison.
Reference system All the test results shown in this section refer to the period from January 1986 to December 1990 (five years). Tests used the opening day, the peaks and trough closure. All entries made on the opening day, following the moving average crossover. Despite the fact that the initial stop loss is a form of release, here they only complicate the process, so we did not use a stop control risks in our tests. Slippage and commission we set equal to zero. We tested the same five markets that used in the previous section: the German mark, gold, soybeans, Treasuries and crude oil. Those methods outputs, which we picked up for testing, are popular and logical strategies, and we assumed that they would be of interest * for most traders. As you can see, the results of our reference system is not very impressive, but it is suitable for our purposes. In general, the main vulnerability of moving averages is that it outputs. When prices change direction, we can expect to lose a large part of income before reversal is signaled. In theory, the outputs of which are more sensitive than the entry should catch most of every market move. Below is a description of exit strategies, which we tested.
Parabolic Stop and Reverse. Parabolic strategy we tested was calculated not quite in accordance with the original formula Wilder. We have used for testing System Writer Plus, a parabolic formula which there is a different from the original. Version System Writer allows the user to choose the number of days past to find the starting point of extreme peak or trough. In other respects, identical calculations. We conducted two tests, one with a 5-day parabolic strategy and another using a 10-day parabolic strategy.
Support / Resistance. One of the most common and clear us out - this is an outlet that tracks the peak or trough of the last n days. The stop is activated when the market correction comes to a stopping point. We conducted two tests, one with a peak or trough last 3 days, and the other with a peak or trough last 10 days. • Tracking RSI (Relative Strength Index). We are used to determine the Wilder's RSI is overbought and spine or re dedication of the market. When the 9-day RSI rises above 75 or drops below 25, and the market closes in the direction opposite to the direction of our trade, we go to the opening of the next day. The idea is to come up with still a strong market instead wait serious correction.
Key reversal. 170
Of different shapes to signal the turning points of market is probably the most widely known is the key reversal. The basic definition of the key reversal (to exit long position) is a new market peak, followed by a decrease in the closing price in relation to the previous day's close. For our test, we have identified a key reversal as a new peak in the last 10 days, followed by a close below the previous day's close. Key reversal to exit the short position is defined as a new trough the last 10 days, followed by a close above the previous day's close. (See Figure 3-14.)
Stop watching The risk for open trade may be defined in two ways. Source risk - the difference between the price when entering the market and stop control risks. Balance risk - the difference between the market price and the price of an open position dictated your exit strategy. Most trend-following topping outputs, such as the output of our reference system, are far away from the market, making the risk of open positions at large. Balance sheet risk is increased when the trade takes place on a portfolio of markets. This statement may be challenged by the fact that the balance sheet risk is a "nice problem" because it increases by lucrative positions. But there is a loss of loss, irrespective of whether they appear at the peak of your account balance or not. And in fact, and in another case, money disappear from your account. One of the easiest ways to control the risks of this kind is to place a stop behind the trade. Stopping may be calculated in several ways, but for this test we chose a fixed dollar stop, calculated from the highest or the lowest in the direction of closing the trade. We tested the tracking stops at $ 500 and $ 1,000. When you look at the results of these outputs, pay attention to the fact that the win percentage is much higher than that which can be seen in the remaining tests. Explanation: When the market corrected, sale closes before has an opportunity to become unprofitable. The price for this is that such a stop negatively affect the total income, as some stations will be closed positions are too far from the point of achieving their maximum potential income. Other things being equal, we like to use servo stops, especially in combination with other types of outputs. They offer a relatively stable way to protect revenues. Tracking stops will also lead to a decrease in volatility trading results, as measured by standard deviation. The smaller the standard deviation of trading, the smoother the curve bill.
171
Volatility Exit from the market is volatile at the break against the direction of the trend. We used to) 00 percent of the five-day average true range to determine our exit signals. From the basic theory that a trend change is often signaled significant, occurred in a single day price movement in the opposite direction.
The slow stochastic oscillator 172
Test stochastic we have simplified as we could. We come out of a long position when the% K crosses below% D, and out of a short position when the% K crosses above% D. We used the 14-day Slow Stochastics. This should lead us to the market when it starts to correct or change direction.
Revenue targets Many of the arguments in favor of the servo stops apply to the purposes of income. Generating revenue for a given amount of dollars will give the effect of smoothing the curve of the account. In theory, the big profits to be overlooked from time to time, compensated preserving some of the proceeds on each trade.
Random outputs When we tested the occurrence, we found that most of the popular methods of entering effectively nothing more than a random selection of points of entry. We thought it would be fun to try to test the random outputs. In this test, we selected a random output in two stages. First select a random number between 5 and 20 to determine the minimum number of days during which the trade will be open. At the end of the minimum number of days, the first closing against the trend will force us to get out of the trade Does opening the next day. (Despite the fact that the minimum number of days chosen at random, the strategy element is included following the Tren-house.) As you can see, the results are not too strikingly different from the results of other tests, however, if we had done a thousand random tests, we could come to horrifying results. (See Figure 3-14.)
Findings At the time, the main objective of this section was prodemonstrovat testing procedures, rather than a specific comparison of outputs, some observations are inevitable. At the time, the yields showed very different results, none of them improved the master system. Below are some preliminary conclusions: 1. Test a separate entry and exit strategy before integrate them into the system. If we have tested our entry on the moving average of the outputs using the stochastic oscillator, we could conclude that the moving averages are unprofitable. 2. Choosing an exit strategy can have a huge impact on the profitability of entry. 3. Output with the highest percentage of winning trades is not necessarily the most profitable. 4. The entries define the range of profitability, and the outputs are responsible for the final result. 5. Tracking stops provide a greater percentage of winnings. Remember that winning percentage and the ratio of average income to average losses are important elements in the formula of probability of failure. If we can reverse both of these criteria in their favor, then we are more likely to achieve our sales targets. We are confident that if you study the test data, then come to a more significant conclusions. We have touched only the surface of the problem, which deserves further study and consideration.
173
Testing stops We have seen that the independent testing market occurrences is relatively easy, while testing the outputs more complex process, perhaps the most significant result of the previous test outputs was that the tracking stops substantially increase the percentage of winnings The main difference between the outputs, we tested before, and stops the losses that we will test in this section is that the outputs of thought to generate revenue, and the thought of stopping in the main. To control risks. But, as you will notice in many cases, the boundaries between these concepts are blurred. The real purpose of both sets of tests is to see how effective the various stops and exits as constituting a simple trend-following system. Just like us, you will be surprised at some of our results,
Metolologiya Since stops are a form of output does not make sense to change our methods, and we will use the same test procedures that were used in the previous section. A reference system for our method we used two averages intersection (5/10, 5/20, 5/30, 5/40). The entries and exits are made only at the intersections. This is the reverse system: the opening, following a moving average crossover, we will close the existing position and immediately initiate the next trade in the opposite direction. One important reason for choosing such a system s back is that the point of entry are fixed and do not vary depending on the output. Thus, it will allow us to consistently compare how the different stops and outputs affect the results. Similar to the previous tests, some of our stops will not be included before reversing position caused by the intersection of the moving averages. Depending on the nature of the stop, it may be significantly or insignificantly, but in any case necessarily. The test period from January 1986 to December 1990. All entries made on the opening day, following the moving average crossover. Slipping and commissions were set to zero. Markets have remained the same as before: the German mark, gold, soybeans, Treasuries and crude oil. Analysis of test results stops subject to the same restrictions as the testing and analysis of the usual outlets. Winning percentage will change in some way, but most will be determined by the chosen strategy occurrences. Therefore, we can only come to conclusions about the limited impact of the stops on winning percentage. We are forced to rely on comprehensive income as the most impartial and accurate measure of the various methods of stopping. As we stressed earlier, the total income is far below the list of statistical parameters that need to keep track of when testing the whole system, but in this case it is the best of the measures available to us. We'll mention one more opportunity to compare the effectiveness of the stops. Stop control risks when used properly, due to lower losses. Behind this is a follow-up, if you like to carry out their own testing.
Stop the initial risk We have grouped the stops that were going to test in three main categories: initial risk stop, breakeven stop and stop watching. Stopping the initial risk can be defined as a stop 174
which limits the loss of some of that trade can accumulate from the date of entrance. When the trade immediately begins to move against us, stopping the initial risk is usually triggered before any other choice. This is your easiest "stop-loss". We tested three types of stops that can be classified as initial risk stop.
Dollar stops. The first initial stop, which we consider to be a simple dollar stop. When a trade goes against us by a specified dollar amount (measured at the point of entry), we immediately get out of the market. Stop acting until she turns or until the position is closed. We tested four dollar values are $ 500, $ 1,000, $ 1,500 and $ 2,000. Obviously, these numbers do not cover all possibilities, but they will give us an idea about the overall effectiveness of this method. In a sense, we have chosen the sum subjective. However, previous testing, we found that the stop nearer $ 500 significantly reduces the percentage of wins, and stopping over $ 2,000 result in large losses. Be careful with extra precision vyboradollarovyh stops. If $ 781.25 works best on Treasuries and $ 425 works best on soybeans, it is tempting to find the "optimal" a stop for each market. It is far better to find a round number with which you would feel comfortable with, and use the same dollar stop in all markets.
Support / Resistance. Many traders feel uncomfortable with a simple stop of the dollar, because they understand that it is not responding to a particular market or recent market changes. Probably the most common alternative is to place a stop at the peak or trough of the recent fluctuations, implying that these points will be important support and resistance levels. We used the peaks and troughs of the oscillations in the next 20 days before our occurrences.
Exit unprofitable. We tested this method because of the reviews that we often hear from experienced traders. They noticed that the best trades are profitable almost immediately. If this is true, then the obvious way to eliminate most of the losing trades will be testing the profitability of each position for a specified number of days after entry. In our tests we used 1, 5 and 10 days. If, after a specified number of days of no income, the trade will be closed at the opening of the next day. If the trade is profitable, it will be allowed to continue until the normal completion. In fact, it does not stop the initial risk, but it is best suited for this classification.
Breakeven stop Breakeven stop is defined as stops that are located close to the point of entering the trade when reached a certain amount of income. The obvious goal of such a stop is to prevent 175
the transition to a reasonable income losses. In modern literature, we have seen a few negative comments related to the break-even stops, mainly refers to the fact that they are prematurely removed from the potentially good positions. In our testing, we found the stop to breakeven income levels of $ 500, $ 1,000, $ 1,500 and $ 2,000.
Stop watching Stop watching - it stops that consistently translated after reaching certain logical price points. They may serve as a primary risk or stops output of income, or both.
Dollar from the close tracking It calculates the stopping point with the highest or the lowest close in the direction of trade. Stopping is not set until the position is profitable in the amount of stopping. We placed a stop of $ 500, $ 1,000, $ 1,500 and $ 2,000.
Dollar tracking the peak or trough It calculates the stopping point with the highest peak and the lowest trough reached in the course of trade. From the point of view of income when they reach the peak, some of the income should be protected placing stop. Any losses on trade will be limited to the difference between the peak or trough of trade and stop tracking. This stop watch starts immediately and is used as a stop-initial risk, as well as the servo stop. We placed a stop of $ 500, $ 1,000, $ 1,500 and $ 2,000. (See Figure 3-15.)
Findings 1. None of the tested stops benchmark results did not improve. 2. Close stopping lowered winning percentage, while the more distant stops increased percentage of winnings. 3. As a rule, stop with less dollar risk, which were placed too close or closed too early trade, were less successful than the stop, which allowed trading to develop fully. 4. Mediocre performance was observed with the approach of the stops, and performance improved during the expansion stops. When the stop becomes too large, better stop. 5. Once we have taken care of the initial risk, the simple dollar stops working as well as the so-called logical stop. This is probably also true for stops that are used as outputs of income. The highest values of total income were obtained using tracking stops, far removed from the current prices, but we think that in practice such a system to follow will be difficult. But many successful sales consultants are using this approach. 6. Break-even stop work extremely well. We have always argued that the outputs from the market is much more important occurrences. We have seen how a simple change in the yield trend-following systems can significantly change the performance results. Our tests have shown the impact of the various outputs, some of which reduce the risk of catching other income, and some do both, and more.
176
177
Create a simple trading system In this section, we are going to offer a chronology of the development and testing of a simple but effective trading system, starting with the selection of the appropriate size of the account, and the last choice for portfolio trading, risk control and setting proper construction of the system. We will explain the reasons for each decision, which will take in the course of this work. We believe in simple systems and simple procedures for testing, so we will not use sophisticated optimization techniques. We want to emphasize that our goal in this section is simply a demonstration of the build and test the system. System, resulting, it is not necessary is the system that we would recommend. We make no representation or anything about the performance of the system in the future.
178
Objectives of the trading system Our first requirement is that the system must be mechanical, simple in the sense of support and should not spoil the life of the trader. Under the support of simplicity we assume that the trader does not have to be glued to the quoted car all day and should be able to completely get rid of the real-time quotes, if he so wishes. This also means that all the necessary calculations may be made (using appropriate software) for a few minutes per day, and all orders can be administered at one time per day, preferably before opening. The level of activity of the system (what we call the "step") must be low enough so that the trader, part-time workers, felt safe and had a sense of control over the situation. Obviously, it is difficult to plan for the impact of relations of any investment scheme (this also applies to futures trading system), but the possibility of return should be commensurate with the level of risk. We'd like to see our testing shows annual recoil ratio at least between 20 to 30 percent. It is also important to have a winning percentage of at least 40 percent, and the ratio of average win / loss is not less than 2:1, which will give us the statistical probability of failure, which is close to zero. We would like to limit our losses from peak to trough by 40 percent or so. If such losses seem high, check the records of several successful commodity trading advisors. Despite the fact that they are professionals and work with much higher scores than the average trader, you'll find that very few of them were able to avoid losses of 40-45 percent. Given the small traders rating luck with parttime and limited the size of our bills, 40 percent sounds realistic.
Account size Sizing futures trading account, obviously depends on the thickness of the trader and his purse psychological resistance, but there are limits below which it becomes difficult to operate. In our experience, backed up by extensive research in this area, relatively small accounts means less than $ 25,000 and have a significantly lower probability of success than larger bills. With this in mind, we chose $ 25,000 as the size of our bill. Even with such a large amount of money damages will be difficult to avoid. Other things being equal, the greater the score, the easier it is to keep the percentage of losses on a small value.
Portfolio Portfolio selection for trade - it is always a subjective task. We would like to ensure that the portfolio is quite diverse and that the markets are not as endlessly volatile, the risk becomes impossible to control. We are creating a system with the idea that it will be possible to engage in 20 minutes a day, and this excludes the day trading. It is difficult, if not impossible, to trade stock indices on a $ 25,000 account and not be in a situation where one bad market day will erase most of our accounts. Also just for the safety net, we would like to limit the maximum level of 30 percent of risk accounts. We will trade on a conventional fixed portfolio of the five markets, each of which is a separate market group. In currencies, we will trade-mark in the precious metals, gold in 179
crops - soybeans and oil complex - crude oil. We will leave unattended food groups and textiles because of all these complex markets, in our view, the most likely (second only to the stock markets) are the major cause of unexpected losses.
Software and data For testing, we use two products: CompuTrac / SNAP and System Writer Plus by Omega Research. In this case, we used the System Writer Plus is very convenient because of the presence of the module, called the Portfolio Analyzer, written and sold by Tom Berry. Portfolio Analyzer takes individual files by using System Writer Plus, and combines them to show test results on a portfolio of markets, instead of just one market at a time. He also believes daily, monthly, quarterly and annual balance sheets, so you can see the result of the impact of changes to the system or a portfolio across the range of markets in which you trade. It allows you to test for portfolio losses and the ratio of return and do not bring there the results for each market for the aggregate report. It often happens that the market, which usually gives a total loss, significantly adds points portfolio, showing a negative correlation of returns relative to other markets. If the market (or markets) wins while the other loses, it softens and smoothes the total losses of the curve accounts. We, therefore, do not worry if one or two of our chosen market themselves collectively as lost as a result of our testing, as long as the balance of a smooth curve and the results of the overall impact are kept at the expected level. Will be used to test the six and a half years of daily data from January 1, 1984 and ending June 29, 1990. We can not go back even further and continue to test the full portfolio, since crude oil began to trade since the end of 1983. We pay special attention to the years 1986.1988 and 1989. It was particularly difficult years for those who follow the trend. If we do survive, it will give the reliability of our trading system.
Adjustment to the curve and optimization We have discussed this topic before, but, as the adjustment to the curve can not be avoided, do not hurt to mention it again. We will adjust to the curve in the technical research and the values of our stops to a reasonable extent. We will not be optimized, and will test our system, and if it does not work, we will modify. We are not going to test the endless combinations of studies and level of risk in order to find the optimal solutions. We will also use identical values research and stops for all markets, even more reliable thus ensuring that we spend so little adjustment to the curve as much as possible. If the system works for the majority of our chosen markets, and losing markets smooth the curve of our account, we conclude that sroyu job done satisfactorily. Remember, we have to deal srynkami and our wallets. We do not need to endlessly adapt to the curve to satisfy the insatiable need of a computer to improve.
Risk control
180
You can not have a successful trading system that can not adequately control the risk. For testing purposes, we divide the risk of two types: the initial risk, which is the difference between the point of entry and the protective stop and balance sheet risk, which is the difference between the inter-market balance and stop tracking. These two categories can be further divided into initial and balance sheet risk through a separate trade, and the starting and balance sheet risk in the portfolio. The initial risk is controlled stops. In our own trade we try not to risk more than 1.5 percent of the $ 100,000 bill, which means a stop at $ 1500.Tak as we trade in the same markets, which would trade at large, and our experience suggests that a closer stop in these markets increases the losing trades, we will focus on that amount. We can consider using other amounts if our win percentage is less than we expected. Remember to test outputs: as a rule, the closer station, the lower the percentage of winnings. Balance sheet risk is controlled by servo stops. There are several ways to do this, but we prefer the simple tracking of the dollar amount. This technique was unfairly brought into sharp focus the audience and the location, but we like it because it helps to quantify formerisk on the portfolio, which is found to be summeraznostey between the value of open positions and the servo stops. If the risk to the portfolio is too large jumps balance will be unbearable. This phenomenon is a sign of commodity trading advisors and their clients. Income is based on a group of items in a matter of months, and then markets unwrapped and unexpected sharp losses on the balance of the account. This occurs despite the fact that there was no losing trades. Client advisor Commodity trading can be in the absurd situation, paying incentive fees or taxes on income which are missing. The only way to control the risk of this kind are tracking station. We will track the $ 1,500 stop. If we had not used the tracking stopped, we would have used the break even stop, but, as their functions are relatively similar, we will only use a tracking stop. You can optimize the value of the stops on one or all together, but we will not do so. We want to get a stable trading system that will work tomorrow, not optimized, which worked yesterday.
Technical Studies - entry The most common indicator of trend following is a simple moving average. We used it for many years and have tested a lot of combinations. As a result, there are some preferences. Usually we prefer to approach double moving averages. Here are some of our favorite combinations: 3/12, 9/18, 10/18 and 10/20. You will remember that the combination of 5/10, 5/20, 5/30 and 5/40 have proved effective in the reference system, but usually we prefer something a little more sensitive. We will try to use the 3/12 moving averages exclusively for entries.
Technical Studies - logical outputs Parabolic Wilder is a common technique exit, and it was relatively impartially tested by us as an individual output in comparison with the reference system. We use it for our outputs, using the original values for the step Wilder points SAR. The starting point of a parabolic system will be different from the original, because the System Writer Plus (SWP) does not have the original formula Wilder. Wilder's formula uses a peak or trough to the parabolic trade as a starting point. The formula uses the SWP peak or trough of the last n days, 181
leaving you choose item relatively short period of n has the advantage of the location of the initial stop at a closer distance. When we tested parabolic systems previously selected number of days was particularly important as it is in the range from 4 to 15. We will use the 10 days, as we did in our testing outputs.
The first test For this test, we want to do everything as simple as possible, while leaving so little room for losses as soon as we can. We use the intersection of 3/12 moving averages for entries, 10-day parabolic systems for the exits, the initial risk of $ 1500 and stop watching. Our procedure will be testing all six and a half years of data to see how the outputs. If they work obviously bad, we split the testing period into two parts, the first part of which will last for four years, and the second - two and a half. Then, we'll test some other combination of our early data confirm we have received at the end of the combination on more recent data. (See Figure 3-16.) In fact, the results are quite encouraging. As we might expect, soybeans and gold in the loss, but look how well the system copes with Treasuries, German marks and crude oil. Perhaps with a little help we get a workable trading system. In the next step, skip through the results of Portfolio Analyzer. We see an average income of $ 3,282 a year (we have never allowed "pyramids"), which gives an annual income of 14.8 percent. We know a lot of traders who would be happy with this and that. But look at the damage: $ 22259! If it had happened at the beginning of our trade, we would have just thrown out of the market. Quick calculation shows that our win percentage of 35% and the ratio of average win to average loss of 2.06 result in a higher probability of failure. The possibility of reducing the portfolio to $ 15,000 (down 40 percent), before it rises to $ 50,000, is 45 percent.
As we can see, something happened that was predicted. We were able to develop a profitable, but unacceptable trading system. This exercise is a good example of why we should never use the total income as the sole criterion for judging the effectiveness of the trading system. As we have stated previously, test a specific set of criteria that would lead to better solutions.
182
The next step The next step is to admit, subjective and conducted from the standpoint of common sense. We're going to look at a list of individual trades in each market in order to see whether the entry was successful or not. This subjectively for two reasons. First, if you are analyzing a single item from a group of interdependent parts of the trading system, you will never know how much is really important to separate each element of the trade. Frequent small changes that you think is perfectly innocent, can break the chain of events, and a wave of change throughout the system. Second, does not really exist, and probably should not exist objective criteria to support your decision. Perhaps there is more experience will help. We view a list of trades to see how quickly became profitable trades and how great was this profitability. Obviously, the printout of the results of trading 36J is a waste of space, but when we consider the all five markets, identifies two common phenomenon. First, as you might expect, a good bargain quickly proved profitable. Second, the bad trades quickly and began to play within the group. When considering trading on the chart by one, the picture becomes clear. Our occurrence of moving averages work well when markets make wide swings, and large consecutive losses arise when markets often change direction.
ADX as a filter We often use Wilder's DMI and its derivative ADX. On most tests, where we used it as a filter to make an objective decision about whether a market is in a state trend, ADX shown to be effective. We found that good results are usually grouped in a range from 10 to 25 days depending on the application. We will use the 18-day ADX, because 18 is in the middle of the range. We have to be careful in choosing the manner in which test ADX, otherwise it will not act as a filter and as an independent method of entry. We program entry, so that we will buy or sell only on the rise and ADX crossing moving averages, resulting in a single day. If ADX strives down, the market is non-directional, and our observation of the bids revealed that this situation brings big losing periods. As you can see in the figure below, there have been radical changes. All markets are now profitable, however, we may wish to make more money on gold for six and a half years of trading. Surprisingly, treasury bills, and crude oil is not as profitable as before. Higher percentage of wins and the number of trades substantially downward. This illustrates the fact that our filter set to work. Let's run the Portfolio Analyzer. (See Figure 3-17 -) Total comprehensive income using the ADF was $ 61,333, more than double the output of the previous test. Annual return now reached 37.7 percent. The number of trades decreased to 151. What's even better, our win percentage rose to 44 percent, and the ratio of average win to average loss came to an optimistic value of 2.32. Our probability of failure is now only 2.9 percent. The maximum loss from peak to trough $ 9414 is well within our original norm. Monthly balance values are presented on the following pages. Note that we filmed income and compensate losses every year since, so each year with $ 25,000. (See Figures 3-18 and 3-19.)
183
Further testing If we want to bring the exercise to its logical conclusion, we should take a few more steps. Most importantly, we need to test each item to be changed on the whole range of its values in order to make sure we do not accidentally tune out the curve of our system. The results must be acceptable over the entire range of values. We could then optimize the system to obtain the "best" values for the first few years of data, and then test these values for subsequent years. In this case, it is difficult to make because of the insufficient number of trades, but if the system generates a lot of examples, we recommend this approach. Determination of the best values is not easy. The safest way is to choose the values that give results in the middle of the range. If the optimal value is at one end of the range, it would probably be better to ignore it and take the value from the middle of the distribution. For example, you tested the initial stop of risk in the range of $ 500 to $ 2,000 in increments of $ 100. The so-called best value had to stop at $ 900, and the results drop significantly when using a stop of less than $ 800. They continue to be normally distributed up to $ 2,000. It is better to use a value of between $ 800 and $ 2,000 instead of the best stops at $ 900.
184
185
Output Computer testing of trading systems is still in its infancy. Traders seem to be moving towards a more productive method of trading, but has yet been written on this subject a little bit. This is probably due to the fact that the testing does not exist as absolutes, as they do not exist in commerce. However, there is a misconception that testing something "proves" ^ this belief, we have tried to dispel. We hope that provoked some debate on this issue, and that's somewhere conducted research to improve the management of futures, and that it will be announced what some talented private trader.
Recommended reading Griffin, P. The Theory ofBlackjack. Las Vegas: Gambles Press, 1981. 186
Lucas, Louis W., and Wade Brorsen. "The Usefulness a / Historical Data in Selecting Parameters / or Technical Trading Systems. "The Journal of Futures Markets 9, no. 1 (1989), pp. 55-65. Vince, Ralph. Portfolio Management Formulas. New York: John Wiley & Sons, 1990. Young, Terry W. "Introducing the Calmar Ratio." Technical Traders Bulletin 3, no. 9 (September 1991), pp. 1-10.
Chapter 4 Trade for the day Introduction The fee for attendance in business A day trader enters and exits trades in the same market session, which is usually six hours. Very short term nature of trade during the day gives the trader advantages and disadvantages. The benefits include: lower level requirements and the lack of exposure to changes in market sentiment for the night. The disadvantages include the high demands on the time and effort trader, limited level of potential income and burdensome costs on frequent transactions. Transaction costs consist of slippage and commissions. A trader might get the idea that trade is conducted at prices issued at a computer monitor, but in fact he should buy at the offer price and sell at the bid price. The gap between supply and demand is essential for the presence of hidden charges in the business. Also realistic approach would be to wait continuing implementation of orders to stop at the price at which the stop is triggered. Commissions are larger and more obvious costs. At the same time, to compensate for these inevitable expenses a day trader is limited to a very small incomes. Even by the most optimistic scenario, the potential income a day trader are only part of the price range, occurring during the trading day. Let's assume that our day trader pays $ 20 per trade in commissions, and the spread between supply and demand is $ 10 for the purchase and $ 10 on sale. To complete the trader to trade with a total profit of $ 100, he should be smart enough to determine the movement of the price chart at $ 140. On the other hand, when the output will bring loss of only $ 140, the trader will lose $ 180. I do not need a Ph.D. to understand that this situation is not ideal for the business. In fact, even if the professionals in the halls of the exchange will be fine disciplined traders, it is enough just to survive. You do not even have any idea how many of these professionals are failing, despite the advantages given by the presence in the halls of the stock exchange and minimal costs on trade. Imagine how small must be the chances of success of a trader operating outside the exchange floor, and who is faced with the above costs. To be successful day trader has to fight for maximizing revenue from each transaction to overcome the huge transaction costs. Unfortunately, a day trader has very little effect on the earning potential, because the range of price movements during the day is very much limits the maximum income of the average trade. No trader can not reasonably be 187
expected to purchase exactly the hollows and selling exactly at the peak. Only a very good trader might expect to be able to catch a third of the price movement during the day. So, in order to get the $ 180, the total change in prices should exceed three times that amount, which is $ 540. How many futures markets provide daily price range of $ 540 to go over? Very few. How many futures markets can bring $ 180 total loss? Virtually all. Do not forget that a trader is smart enough to find markets with price swings of $ 540, and nimble enough to trade them so well to get $ 180, will only break even, until you get superior number of wins over the number of losses. In order to make money over a long period, a day trader must have a win percentage is much greater than 50 percent, or it must somehow figure out how to make more than $ 180 on the fluctuating prices of $ 540. (Or, better, and then a combination of both.) It also suggests that the trader is smart enough and disciplined in order to pacify their instincts and emotions and carefully limit the size of the losses.
Limited chances As you can see, a day trader is faced with an almost impossible task. We venture to make a very educated guess that less than one in a thousand a day trader makes money for any extended period of time. We recommend that you do not even try. Your time and energy can be used to improve your skills long-term trading. Even if you succeed in day trading, it will be difficult to re-invest their profits and continue to multiply. Day traders can operate effectively only to a small extent, so do not expect that you're lucky, at best survival inherited hard work. Despite our recent admonition, we know a lot of traders who would try to take a chance and become for a time day traders. Fortunately, the lessons learned can be applied later to a more serious and productive trade. We will try to tell you all that we can trade for the day, and make the learning process less costly. Obviously, we do not have all the answers, otherwise our opinion on the likelihood of success would not be so negative. We learned a lot about this subject for many years of trading, and the fact that we have decided not to play in these games, shows only our personal preferences relating to the distribution of our work time. We hope that the hard work we inherited information will be useful.
Selection of markets to trade during the day As we stressed earlier, very few markets offer a fairly wide price swings during the day that it makes them good candidates for day trading. Day traders generally prefer to focus on just one or two markets. The prices have to watch closely, and there are very few markets that are right for us, even on the assumption that we have the opportunity to watch many of them. At the moment, day traders prefer equity index liabilities, currency and energy markets. From time to time, and other markets may become interested in trading for the day due to the time periods of high volatility. We tested to determine what percentage of time the major markets have a common daily range of $ 500 or more between the peak and the trough of the day. Here are some of the results obtained on our data last 1000 days: the index S & P - 50 percent. New York Composite - 64 percent, British pounds - 53 percent, treasury bills - 50 percent, Swiss francs - 50 percent, Japanese yen, 38 percent treated with oil - 37 percent of the German mark - 35 percent, crude oil - 31 percent, soybeans Beans - 28 per cent, silver - 23
188
percent, gold - 21 percent and sugar - 13 percent. As you can see, only five markets had a range of $ 500 for 50 percent of the time.
Consider spreads In addition to finding a wide daily range, the factors considered when choosing a market to trade during the day, should also be a minimum liquidity and spread. Our sample costs included the cost between buying and selling at only S10. In the market S & P minimum spread is $ 25, while the market commitments - $ 31.25. If you are trading for the day with a commitment of $ 20 commission, you have to overcome the total cost of $ 82.5, added to the losses and gains of the deductible. Your average winning trade should exceed the average losing trade at $ 165, but that was not ubytkav. This suggests a spread of one tick, which is the best of all possible cases. The element of liquidity comes into play when determining the number of ticks in the spread between supply and demand. The gap in one tick - it's the best that you can count, but most markets have a wider spread. You can usually assume that the higher the average daily volume, the narrower the spread. For this reason, you may want to concentrate your day trading only in markets with very high volume. Otherwise, you can make good and timely decisions, but still lose money.
Maximizing revenues Day traders are constantly faced with the problem of capturing the greatest possible return on a relatively small range of prices. This situation naturally leads traders to the strategy of buying on dips and selling at elevations instead of following the trend. Most trend-following strategies are too slow to trade during the day. Counter trend strategies offer the ability to extract the most revenue from a small range of price changes. However, counter trend strategies are less reliable than the following strategy zatrendom as quick definition turning points is much more complex than a simple trade with the trend. We have noticed that the best day traders use elements of both. The most successful are trying to buy the hollows within an uptrend and sell on the peaks within the downtrend. A day trader who consistently makes money, how to be good to follow the trend, and to quickly find points of short-term reversal. Most traders lose money because they will never shine in any of these roles. When we consider some examples of possible strategies following the Tren-house, keep in mind the following two steps: first, find the intermediate trend and then find the point of a short-term reversal. In order to achieve a winning day trading both steps should be done quickly and accurately.
Our rebuttal Trading Methods for the day, which will be described below, are a few examples of the many methods by which shared with us over the last few years. We rarely try to trade during the day, so we have very little of its own experience with any of these methods. Various traders, who shared with us, claimed that their methods are very successful. We tried to pick the ones that seemed the most logical and kept casual verification on very limited data. The inclusion of these methods in a book 189
should not be construed as an endorsement or recommendation. In the best case, they should give the reader food for thought and a typical example of the many techniques and tools that can be used to trade for the day. Use at your own risk.
Method envelope 5-25 This method of day trading is based on a very unusual way of using moving averages envelope. Most systems based on envelopes, cause trade in the same direction as the breakthrough envelope. 5-25 method does exactly the opposite. We assume that the market will fluctuate between extremes of the envelope and consider out-of-boundary levels as repurchase or resale. After leaving abroad, we expect that the market will go down in the envelope and begins to move to the opposite side. Rules are as follows: 1. Use the 30-minute chart of futures on S & P. 2. Set the normal (no smoothing) envelope for five periods at a distance of 25 hundredths of one percent from the closing price. 3. Start search successfully, only when the boundaries of the envelope spaced apart by at least 150 points. When the closure of one of the bars 30-minute chart appears at least 5 points on the boundaries of the envelope, begin to trade in the opposite direction as soon as the next bar closes inside the envelope. 4. Use the initial stop loss placed at a point of extreme peak or trough in front of your entry. After the market will move 75 pips in a favorable direction for you, the stop loss should be changed at least to the level of your break-even point. 5. Take profit when the market reaches the opposite side of the envelope. If you want to simplify this process, use the boundary of the envelope at the time of your entry into the position as the target, otherwise you will have to reinstall your point out every half hour. (See Figure 4-1.) Several changes to the envelope, the system can be used for normal trade instead of trading for the day. We have had good results when trading the soybeans.
190
The system "Hi MOM" (High torque) We call this strategy day trading system "Hi MOM", because it signals the beginning of trading comes only at high torque. Here's how it works: 1. Use a 9-minute charts of futures on S & P. We chose a 9-minute interval, as the system has to be sensitive to small price movements, a 9-minute bars also divide the trading day at 45 equal time periods. Ten minute bars probably would have worked just as well, but we prefer to use the bars that represent the same time periods, not to get gypsy bar at the end of the trading day. 9-minute charts also allow us to start before traders using the more common intervals of 10,15,20 and 30 minutes. 2. Directly below the 9-minute chart S & P set 6-period time. Select the scope of the study so that you can easily determine when the time reaches the + / -150. 3. Look for divergence between the time schedule and S & P. The first spike of the divergence we are looking for, burst on our schedule since the level of + / -150. The second and third spikes divergence does not necessarily have to reach the level of 150. 4. After divergence Hi MOM enter the market immediately after the occurrence of the hook, terminating the figure of divergence. Place the initial stop loss so that it is performed by 20 points at the nearest peak or trough of schedule. (In point of divergence AB.) Use the servo stop using the peaks and valleys in the graph as the support and resistance levels. 5. Take profit when there is a divergence in the opposite direction, but do not expand trade. Trade only on the first day of divergence. The exception to the rules of a trading day is when ABC is set divergence with three spines instead of two. When we came in after the second peak and were not lucky enough to to be stopped before the third peak, initiate a second trade in the same direction unless the divergence continues. Close any positions remaining open at the end of the trading day. (See Figure 4-2.)
191
The system "Hi MOM" is a simple but very effective, because it brings together patient waiting periods of volatility (signaled by the level of the moment + / -150) with a beautiful clock entries, the proposed divergence.
Intermarket divergence, three-dimensional technique Ohama One of our subscribers in Los Angeles, Harry Sometimes, within a few years he worked side by side with Bill Ohama until his death in 1990. Harry is very successfully applied the well-known "3D" technology Bill in day trading. Here is an explanation: 1. Create your five-minute charts for two or three related products. For example, compare the five-minute chart S & P, NY Composite Index and the Main Market (Major Market Index). You could also compare the treasury bonds (T-bonds), Treasury notes (T-notes) and municipal obligations (muni bonds). There are other possible related groups, such as currency, energy futures or soy complex, but the best trades for the day are usually obtained on stock market indices or liabilities. 2. Carefully compare the five-minute charts for divergence when one market creates a new peak or trough, and one or more other markets in the group can not support the motion and create your peak or trough. (See Figure 4-3.) 3. When divergence is defined, trade should be conducted on the most traded (most liquid) market in the group. For example, in the case of stock indices, you ought to trade S & P, and no major market indices. 4. After entering the trade is recommended to use some method of tracking stops. For example, tracking stops at about 125 points on the S & P could be the starting point. It would be logical to use a remote stop during periods of volatility and closer stops, when the markets are calm. 5. If you got a quick profit of $ 500 for half an hour, just fix it. If the trade goes slowly, hold the position until it seems like it is moving in the right direction. Harry was waiting for the close of trade, and uses common sense to make a decision about when to take profits, and when the losses. This method is not a complete system because of the lack of special stops and more defined exit strategy. Your results may be better or worse depending on your skill in determining the output. We like it here a method of entering.
192
Hooks% K Kane Next day trading strategy on the S & P was sent to us by Steve Cain, who spoke with us at the Conference on Technical Analysis Fred Brown in Austin, Texas, in 1990. When we got back to the workshop, we began to observe the system and have been very encouraged by its performance on real data. Here's how it works: 1. Identify the trend, using the one-hour charts. Trade only when the hourly charts create the highest peak or trough of the highest in the last two hours. When the trend is up, look for buy signals only. When the trend is down, look only sell signals. 2. Occurrences: use the five-minute chart with 12-period slow stochastic oscillator. Buy when the% K (the faster line) is below 20 and will turn up. Sell when% K rises above 80 and turn down. (Do not forget to trade only in the direction of the time trend.)
193
3. Stop: use the initial stop at 100 points, or install a closer stop right outside the borders of the recent trading range. When the price will go ahead by 100 points, a good idea would be to raise the stop to breakeven. 4. Outputs: take profit when the% K creates a hook in the opposite direction from the 80 or 20. Another strategy would be to monitor the one-minute stochastic oscillator and the output in the event of any divergence with the current trend. (See Figure 4-4.) Becoming gave some valuable additional comments that follow. He said, when the signals entering the% K also creates a divergence with prices that move the result is extremely strong. It is also recommended in case of an unexpected jump in revenue by 100 points or more immediately take profits. Finally, he recommended: for the day when there are two consecutive loss, stop the trade and try again the next day. This is good advice for almost any method of trading for the day. We like the idea of this method is to buy the hollows in an uptrend and vice versa. We also like the idea of buying from hooks% K, instead of waiting the usual crossing signals. We think the strategy sets than S & P could also be used to trade for the day in other markets.
194
One-minute charts with stochastic oscillator This trading method was developed by Humphrey Chang, a former futures trader and broker in California. Humphrey explained that the method works best when day trading futures on the S & P, but said that sometimes uses it for day trading futures on the yen and the Swiss franc. Here is the method: 1. Use one-minute charts S & P and 21-period stochastic oscillator. 2. One of the most important rules is that the occurrence produced only during the first hour of trading. After the discovery wait until the two lines of the stochastic oscillator will reach extreme levels (above 80 or below 20) and then cross over. Come immediately after the intersection. Ignore all signals stochastic oscillator after the first hour. 3. Use a tracking station. Look for higher depression or if you're in a short position, dropping peaks. Humphrey warned against the use of stops exactly at the peaks or troughs of the day. He noticed that these are the points being monitored by traders in the stock room, and punch them as often as possible. 4. Trade remains in force as long as you stop tracking or stop the closure of the market at the end of the day. (See Figure 4-5.) Humphrey said that the method can be made more reliable, if only trade in the direction indicated by the 14 half-hour-period stochastic oscillator. For example, if the half-hour the% K is above the% D, then you should use a one-minute stochastic oscillator for buy signals. Points of support As we explained in our previous chapters, we do not believe in the common practice of predicting a specific price. However, if enough people use exactly the same methods and, as a result, considering the predicted same price, they will really get a predictable effect on trade. We suspect that the popularity of the points of support causes them to act as a selfrunning predictions. The formula for the calculation of reference points (or levels of support and resistance) was opened to us by one of our subscribers leaf Ueyntraubom Neal, author of "The Weintraub Daytrader" and the exchange floor trader in Chicago, which holds some unique training seminars for traders called them Commodity Boot Camp . Neil explained to us that the formula is very widely used by many traders especially the exchange floor that mark reference points in order to enter the market at the trough of the next day. Neil suggested that familiarity with these calculations can be particularly useful for traders S & P. We begin to calculate the points of support by adding the peak and trough the previous day. We then divide the sum by three to obtain the average price.
195
Example: Yesterday's peak = 365.30 Yesterday's trough = 361.30 Yesterday's closing = 364.40 Total = 1091.00 divided by 3 = 363.66 (average price) Now to find the peak of the current reference point (or resistance level), we simply take the average price of the previous day, multiply it by two and then subtract the cavity the previous day. Example: 363.66 (yesterday's average price) * 2 = 727.32 727.32 - 361.30 (yesterday's trough) = 366.02 (expected reference point is the peak) To find the reference point depression (or support level) we simply take yesterday's average price, multiply it by two and subtract the peak of the previous day. Example: 363.66 (yesterday's average price) * 2 = 727.32 727.32-365.30 (yesterday's peak) = 362.02 (expected reference point depression) These numbers are approximate support and resistance levels, which are widely used throughout mnogihlet day traders and traders of the exchange floor. Since they are not the point of the graph, and the computed values, the creators of the charts is probably already noticed them as a fait accompli, while the exchange floor traders in advance to celebrate these numbers in their sales notes. Neil believes that we can advance computing one step further, if we want to calculate the "highest peak" (extreme point of resistance), as well as "the deepest depression" (or extreme level of support). To calculate the highest peak we take yesterday's average price of 363.66, subtract the expected reference point depression 362.02 and add the expected peak of the reference point 366.02. Our answer to 367.66 would be a good target for the rising side, if the resistance is broken through 366.02. It can also identify the next possible level of resistance in moving the market. To calculate the deepest depressions we take yesterday's average price of 363.66 and subtract the difference between the expected reference point depression 362.02 and 196
366.02 top reference point (the difference is 4.00.) Our answer is 359.66 likely target, or the lowest point on the way down, if our level of 362.02 support will not sustain. For our example, we used the real numbers, and we could not resist the temptation to test S & P after the closing, in order to see what we have done. Dent of the day was 362.10 against the expected reference point depression 362.02. Not bad. (The September S & P 6-21-90.) Again, we caution that these points are working solely because of their popularity and this popularity may be short-lived. If they briefly stop working because of a certain important factors that may never make again, and you'll be able to part forever with this method. On the other hand, if more people will follow these methods, over time they will work better than earlier. And while we think it's an interesting technique, which was worth mentioning.
Gaps prices discoveries At a meeting of technical analysts of Southern California in 1989, Bruce Babcock Jr., who was invited to speak, described on its Strategy Day Trading. Bruce is the publisher of Commodity Trader Consumer Report and the author of several books on commodity trading. His book, "The Dow Jones-Irwin Guide to Trading Systems", to which we are repeatedly referred to, contains a wealth of useful information, and we encourage our readers to his book, and a piece of CTCR. Here are the highlights of day trading strategy, which we have described Bruce: 1. When trading S & P futures for the next month expect the opening price, which creates a significant gap in relation to the previous day's close. Then look for trade in the direction of the break. If the break occurs at the opening in the upward direction, look for the entry on the side of buying. If the break occurs at the opening at the bottom edge, look for the entry on the sell-side. 2. Mark the level of prices on a few items above (stop shopping) or below (stop sale) on the range of the opening price and the open position when the market begins to trend in the direction of the gap and the purchase price exceeds the stop or stop the sale. 3. Use a stop loss at about $ 500 and get out of the position at the close. (See Figure 4-6.) Strategy breaks can be tailored to meet your individual trading style. Bruce made it clear that he has done considerable testing with various breaks and intermediate parameters. Generally, the larger the gap, and the more points you mark on the opening range, the more likely that you'll get a winning trade. However, the expectation of big breaks and a longer follow in their direction means a reduction in the number of trades. Remember, profitable day trading requires a high volatility, so we recommend keeping the options closer to the upper level than the lower. Thus, you will only trade after the market demonstrate some volatility. As a starting point, try to wait for the gap of 75 points and 25 points to note following the trend. If you want to increase the number of trades, reduce these numbers, and if you want to reduce the number of trades, take the numbers above.
197
RSI divergence Here is an idle day trading method that uses a short-term RSI for finding potential peaks and troughs of the market S & P. This is a logical approach that should work in any market, and can be modified for use as a long-term trading. This is one of our favorite day trading strategies designed for those who do not necessarily trade every day. This method may be idle for several days between trading signals, but when the signal occurs, it gives a win percentage greater than some of the more active strategies. Since he trades every day, it can be used as a complement to the more active system. This method works best when the trades in the direction of a longer trend. When there is no dominant trend, the signals may be taken in any direction. You can use the ADX as a tool to measure trends. When ADX is falling, the auction may take place in either direction. Be patient and not to prejudge the divergence. The following are specific rules: 1. Use the 30-minute chart of the S & P with shestiperiodnym RSI, based on the close. 2. Look for the model of divergence, in which the first spike RSI crossed levels of 80 or 20. The second spike RSI does not have to reach these levels. Buy or sell immediately after the divergence confirmed the 30-minute closing in the direction of the signal. 3. Use by entering a stop at 100 points S & P or level two ticks above or below the recent peak or trough, preferring some of them, that would be closer. 4. Get off at the bus stop or at the close of the day. 5. Do not open new positions in the last 45 minutes of trading on the market. (See Figure 4-7.) In markets other than the S & P, this method day trading mozhetbyt modified by replacing the 30-minute chart in the shorter term. If this step has been done, then the levels of RSI 80/20 can be set to a higher or lower level. 198
It should be noted that in periods when the market is less volatile, 70/30 works better than 80/20, as the RSI fails overbought or oversold. However, one of the important advantages of this system is that it requires a fairly volatile market to make RSI reach level 80/20, and the signals entering the trade not arrive until until sufficient volatility to make trade worthwhile. Do not neglect this valuable feature of the system, setting the levels below 70/30 only in order to get a more frequent trades.
The system of "Knife" Sibbeta We present a system of day trading NYSE Composite Index (commonly referred to as traders "Knife" ("Knife"), because according to her trade on the New York Futures Exchange-tion (New-York Futures Exchange (NYFE). It was told to us by Jim Sibbetom , our old friend, who taught us 25 years ago. He's probably best known for "Demand Index" and its sheets of paper with the news bulletins on silver and gold. Jim says he prefers to trade NYFE, but not S & P, because if we consider the return on every dollar invested, it can make more money on NYFE. He also believes that it is more orderly market with less risk. Here is an explanation of his strategy: 1. Use a 5 -, 10-or 16-minute charts nearest contract NYFE. Need only price data, not pricing models, so that the time slots play no special role. 2. Determine the point on the graph recent significant peak or trough. (For now, just to block out her eyes). If the point was the high point, then look for an opportunity to go into a short position once the market falls by 70 points from the peak. If depression was a significant point, look for the purchase, as soon as the market goes up by 70 points from the trough. As soon as the market moved to 70 points from the peak or trough, follow the current trend in the assumption that it will continue. You should use stop buying or stop the sale, which will automatically enter you into the trade in the direction of 70 points. 199
3. Once you start trading, protect yourself sequence is very close to the stop loss of 30 pips from your entry point. 4. If the trade goes favorably for you, then when you go for 30 points, move your stop to the point of entry. When you go for 70 points, again move the stop on the 50 items so that you stay on revenue 20 points. When you go for 90 points, use a stop tracking 70 points and be prepared to not only come out, but also expand trade in the opposite direction. (You can not turn around at the end of the trading day, when they are ready to hold the position until the next morning. Sib-bet to hold the position in this way only if the other indicators confirm the decision.) 5. If you're unlucky and you were stopped to the point where your stop removed by 70 points, you should try to re-enter the market in the same direction. When you return the reentry position as the only market to produce movement in 20 points in the direction of your previous trade. (The fact that the market has not turned by 70 points since the previous signal, indicating the continuation of a trend in the same direction in which you are trying to sell earlier.) Sibbet says he often got the opportunity of re-entering at a better price than the one where he went, and made a profit on the second attempt. (See Figure 4-8.) We do not really like the high activity and close monitoring of the market, which are required to follow the method Sibbeta. If you go out for coffee, you can miss two or three changes, causing a stop, and a couple of turns. You will also need a very patient and understanding broker who will put up with frequent changes of stops. However, there are some basic advantage of the system, and we thought it would be nice to mention it as an occasion for reflection. This system can be the basis for other more practical with the broad parameters of the system.
Divergence of the stochastic oscillator This method combines day trading ADX, a half-hour and three-minute stochastic stochastic oscillator. This system works best on S & P futures and currencies. 1. Use the 18-day ADX / DMI to measure the strength of the daily trend. If the ADX is rising, trades must be made at the direction of the trend. If ADX is falling, trades can be made in either direction. 2. Check the direction of short-term trend, using the half-hour chart the slow stochastic oscillator as an indicator of the trend. You can trade in the direction or against the trend of the stochastic oscillator, while falling 18-day ADX / DMI.
200
3. Use a three-minute chart of futures contracts. Set up a different schedule, using the three-minute bars for the 21-period slow stochastic oscillator. 4. Entering the auction takes place after the divergence between the three-minute schedule futures and three-minute chart of the stochastic oscillator. The first point of divergence must arise when either stochastic is above 80 or below 20. Look for the occasional three-point divergence. They are less common than the two-point divergence, but it's really good signals. Moreover, you can trade after the divergence of the three-point signal, regardless of the trend ADX day or stochastic oscillator. (See Figure 49.) 5. Place your protective stop. The starting point of the stop loss for S & P should be 20 points higher than the recent peak of short positions, or 20 points below the recent trough for long positions. The stop can be changed after each peak or trough of the stochastic oscillator, setting the level, spaced by 20 points from the new peak or trough in the graph S & P. 6. Take profit or go out on the close. If a three-minute stochastic oscillator signals the opposite of your position, being at a point below 20 or above 80, then this is the place where you should take profits.
Key reversal and stochastic oscillators This idea of day trading combines several technical elements: recognition of figures (what we did not say much), stochastic oscillators and divergence. We anticipate that most of our readers have some understanding of stochastic oscillators and divergence, as we have described them in detail in the previous chapter. However, the part devoted to the recognition of shapes, is a new element and requires a short explanation. The purpose of pattern recognition is to attempt to predict the turning point of the market by monitoring the sequence of price movements that occur regularly and have a predictable value. The figure described here is called a "key reversal". (See Figure 4-10). This pattern consists of three bars with key depression, which creates a new short cavity in which the third bar follows without generating new cavity and which is closed above the key closure. Rotation of this figure can also be used to determine market peaks. When we keep in mind this figure, the system is as follows: 201
1. Use the 30-minute chart of the futures for S & P and the nine-period slow stochastic oscillator. 2. Look for the key figures of short-term reversals, which we have described above. 3. Enter a trade when you spot a figure key reversal, which is accompanied by the value of the stochastic (% D) for a figure below 30 or above 70 purchase, if it is a figure of sale. 4. Place the protective stop at one or two ticks of the key peak or trough. If the stop is too remote to be admissible, use a closer dollar stop. 5. Come out to any shape key reversal in the opposite direction, or at the closing. (Again, refer to Figure 4-10.) Your chances of success will be higher if you live to see the divergence between the stochastic oscillator and the prices of futures. However, the divergence is not required.
App Formulas technical research We found that the formula of technical studies vary from vendor to vendor software, as well as different authors and traders. Some differences may be the result of programming errors or misunderstandings, and some may be honest attempts to improve the formula. A 202
situation where two standing side-by-side computer monitors show RSI from two suppliers with different values can lead to confusion. In our opinion, there is no "correct" formulas. While the use of the study is consistent, your results should be similar to those obtained using some great source. If a small change in the indicator gives a completely different performance results, you should ask if you are not too rigged study under the curve data instead put the blame on erroneous calculations. The following formulas are commonly used. They are not meant to define and the only possible and should not be treated this way.
Directional Movement Indicator (Directional Movement Indicator - DM1) and Average Directional Movement Index (Average Directional Movement Index - ADX) Refer to the corresponding section of the chapter "Technical Studies".
203
Weighted (Weighted)
The most common method of weighing moving average - multiplying the price of each day on the number of days that have passed since the appearance of this price. In the 10-day weighted moving average today's price will be consigned to 10 times more weight than the price of a 10-day prescription.
Divergence - the convergence of moving averages (Moving Average Convergence Divergence - MACD)
MACD consists of a single line, which is the difference between the exponential moving average, and the second the "signal" line, which is an exponential moving average of the first line. The first line is calculated as follows:
Usually MACD uses 12dneydlyaEMA1 and 26 days for EMA2. This approach gives the MACD, which Appel recommends to the sell side of the market share, but that most practitioners use both for long and short signals. Appel configuration for purchase to the 8day and 17-day EMA, respectively. 204
Parabolic SAR (Parabolic)
The first point Parabolic Stop and Reverse (Stop and Reverse - SAR) in a series of data this is an extreme price previous parabolic trade. Therefore, SAR ^ = EP. Subsequent SAR is calculated as follows:
205
206
207
208
