DV Indicators Manual v1 Free

2010

Indicator Manual

April 2010 | Version 1.0

DV Indicators | Copyright CSS Analytics Inc 2010 All Rights Reserved

www.dvindicators.com

Contents Introduction – Indicator Construction Principles

page 1

DV Indicators

Page 3 page

page

DV Bounded (DVB)

3

DV Stochastic (DVS)

28

DV Unbounded (DVU)

5

DV Bands (DVBU/DVBL/DVBM)

29

DV Oscillator (DVO)

6

DV Band Indicator and DV Band Percentile (DVBI/DVBP)

30

DV Super-Charged Bounded (DVSC)

7

DV Zones (DVZN)

32

DV Intermediate Oscillator (DVI)

9

DV Super-Charged Percent Exposure (DVSE)

33

DV Mean-Median Divergence (DVMM)

10

DV Bounded Percent Exposure (DVBE)

34

Rolling Exponential Moving Average (REMA)

13

Daily Historical Volatility (DHV)

35

Rolling Relative Strength Index (RRSI)

14

DV Percentile Rank Volatility (DVPV)

36

DV Fractal RSI (DVFR)

15

DV Composite Volatility (DVCV)

37

DV Adaptive RSI (DVAR)

17

DV Breakout, Composite Vol Plus, Composite Vol Minus (DVBR)

38

DV Aggregate-M (DVAM)

19

DV R-Squared Autocorrelation (DVRAC)

39

DV Adaptive Agg-M (DVAA)

21

DV Differential Autocorrelation (DVDA)

40

DV Intermediate Stretch (DVIS)

22

DV Composite Fractal Efficiency (DVFE)

41

DV Intermediate Magnitude (DVIM)

23

DV Smoothed Trend (DVST)

42

Trend Stochastic (DVTS)

25

DV Self-Adaptive Slow (DVAS)/ DV Self-Adaptive Fast (DVAF)

43

DV Trend-Minus Cycle (DVTO)

26

CSS Analytics Adaptor One (CSSA1)

45

DV Super-Smoothed Double Stochastic (DVDS)

27

Example Applications System One – DV Swing System

47

System Two – DV Combo System

49

Appendix 27 April 2010

Page 46

– Amibroker Code & Installation Instructions


Page 51 2

DV Bounded (DVB) SPY - Daily 12/03/2010 Close 115.46 115.46

Type: Short-Term Mean Reversion Platforms Amibroker, Tradestation, Excel Baseline Strategy Long <= 0.50, Short > 0.50 Suggested Usage …

September November SPY - DV Bounded (5, 252) = 0.51

December

2010

February

March 0.8 0.50996 0.2

Created with AmiBroker - adv anced charting and technical analy sis sof tware. http://www.amibroker.com

Description

“The RSI Killer”

The DV Bounded is also known otherwise as the “DV2” and was created by David Varadi to capture the normalized relative close. The “DV2” version of the DV Bounded represents the 2-period average of the relative close to the high-to-low range. The “bounded” portion represents the normalization using a percentile rank function to re-scale the indicator on the basis of the historical distribution of values.

27 April 2010


3

Example 1: DVB v RSI2 on SPY Amibroker Code in Appendix 1

DVB

30.40% CAGR Since 2000 Ticker: SPY Rules: Long: DVB <= 0.50, Short: DVB > 0.50 Period: 1-Jan-00 to 14-Apr-10 Metrics CAGR: 30.40% Sharpe: 1.23 Num Trades: 912 Average Trade: 0.33% % Winners: 62.4% DVR: 0.70 SPX Correlation: -30%

RSI2

16.20% CAGR Since 2000 Ticker: SPY Rules: Long: RSI2 <= 50, Short: RSI2 > 50 Period: 1-Jan-00 to 14-Apr-10 Metrics CAGR: 16.20% Sharpe: 0.50 Num Trades: 60 Average Trade: 0.20% % Winners: 66.7% DVR: 0.37 SPX Correlation: -29%

27 April 2010


4

DV Unbounded (DVU) SPY - Daily 18/02/2010 Close 110.91 115.46


September November SPY - DV Unbounded (5) = 0.01

December

2010

February

March 0.008 0.00099 0.00000 -0.008


Description

“The RSI Killer”

This is the unbounded variant of the DV2 that was originally introduced by Michael Stokes at MarketSci. Unlike the original DV2 - which is bounded by using a percentile rank function - the unbounded DV2 is simply a raw conversion of the average of the relative close to the high-to-low range.

27 April 2010


5

DV Oscillator (DVO) SPY - Daily 12/03/2010 Close 115.46 115.46


September November SPY - DV Oscillator = 0.65

December

2010

February

March 0.8 0.649249 0.2


Description

“The Mother Oscillator”

This is the “mother oscillator” framework from which the simpler DV Bounded variant was originally derived. It was conceived in its first form as a flexible blueprint for the selection of the optimal oscillator calculation. The DVO is partially adaptive and more accurate/reliable than the original version. As a consequence it requires more historical data, and the indicator values make smoother transitions to changing volatility. As of the current writing, no other public indicator that we are aware of shows raw performance as strong as the DVO on the major indices.

27 April 2010


6

DV Super-Charged Bounded (DVSC) SPY - Daily 12/03/2010 Close 115.46 115.46


September November December SPY - DV SuperChargedBounded (5, 252) = 0.55

2010

February

March 0.8 0.545817 0.2


Description

“The big brother of the DV2”

This is the smoothed and slightly altered version of the DV Bounded that delivers superior performance. The nature of the calculation is very similar, but the differences make it more responsive with less whipsaws than the original.

27 April 2010


7

Example 2: DVU,DVB,DVO & DVSC Baseline Strategy

Comparison of Baseline Strategies on SPY 3,000,000

DVO DVU Long: Enter DVU <= 0.0, Exit DVU > 0.0 Short: Enter DVU > 0.0, Exit DVU <= 0.0

2,500,000

DVSC

Equity

2,000,000

DVB Long: Enter DVB <= 0.50, Exit DVB > 0.50 Short: Enter DVB > 0.50, Exit DVB <= 0.50

DVB

1,500,000

DVU

DVO Long: Enter DVO <= 0.50, Exit DVO > 0.50 Short: Enter DVO > 0.50, Exit DVO <= 0.50

1,000,000

DVSC Long: Enter DVSC <= 0.50, Exit DVSC > 0.50 Short: Enter DVSC > 0.50, Exit DVSC <= 0.50

500,000

Jan-00

Jan-01

Jan-02

Jan-03

Jan-04

Jan-05

Jan-06

Jan-07

Jan-08

Jan-09

NOTE: Back-test assumes 100,000 starting equity with compounding of profits

Jan-10

Amibroker Code in Appendix 1

System Metrics No.

Net % Profit

CAR

Max. Sys % Drawdown

Sharpe Ratio

Avg Profit/Loss %

% of Winners

DVR Ratio

SPX Correlation

DVU

1184%

28.17

-24.4

1.13

0.31

62.31

0.65

-0.32

DVB

1433%

30.4

-25.79

1.23

0.33

62.39

0.69

-0.3

DVO

2529%

37.42

-19.58

1.33

0.41

64.02

0.76

-0.33

DVSC

2152%

35.37

-19.76

1.29

0.4

63.89

0.82

-0.32

27 April 2010


8

DV Intermediate Oscillator (DVI) SPY - Daily 14/04/2010 Close 121.19

121.19

Type: Intermediate Mean Reversion Platforms Amibroker, Tradestation, Excel Baseline Strategy Long <= 0.50, Short > 0.50 Suggested Usage …

Sep Oct Nov Dec SPY - DV Intermediate(M+S) (5, 10, 252) = 0.68

2010

Feb

Mar

Apr 0.8 0.677291 0.2


Description

“A Smoothed Intermediate Oscillator”

This is the smoothed intermediate oscillator framework designed to complement the DVO. It was also designed to be flexible for use in adaptive systems. The current DVI is a combination of two different indicators: 1) DVIM- DVI Magnitude which measures the distance price has travelled over multiple intermediate time frames and 2) DVI Stretch which measures the net up or down days over multiple time frames. Both indicators are complementary and combine to produce a superior composite- the DVI- in most cases. The DVI line is very smooth and is designed to help identify areas of higher or lower value rather than peaks or valleys as it stays in oversold or overbought territory for longer periods. This makes it ideal for combination with intermediate trend systems, or for filtering DV2, DVSC or DVO trades.

27 April 2010


9

DV Mean-Median Divergence (DVMM) SPY - Daily 17/08/2005 Close 122.2

121.19

Type: Intermediate Trend Platforms Amibroker, Tradestation, Excel Baseline Strategy Long: Rising and/or >0.0, Short: Falling and/or <0.0 Suggested Usage …

December 2010 February SPY - DV Mean-Median Divergence(26,12,9) = 0.06

March

April 1.41192 0.8 0.2


Description

“The DV MACD”

This is the DV equivalent of the MACD, substituting the median for the average to make it more responsive to picking up acceleration in the trend. It is one of the few indicators that are successful on noisy intraday data. It is one of the best intermediate term trend indicators we have tested, and can be used numerous ways: 1) as a trend filter 2) as a trend indicator 3) an oscillator 4) it can be re-scaled to create a volatility measure.

27 April 2010


10

Example 3: DVMM on WYNN,BIDU WYNN

44.98% CAGR Since 2002 Ticker: WYNN Rules: Long: Enter DVMM > 0.0, Exit DVMM <= 0.0 Short: None. Period: 1-Jan-02 to 14-Apr-10 Metrics CAGR: 44.98% Sharpe: 1.26 Num Trades: 41 Average Trade: 8.66% % Winners: 56.10% DVR: 1.00 SPX Correlation: 4%

BIDU

63.00% CAGR Since 2000 Ticker: BIDU Rules: Long: Enter DVMM > 0.0, Exit DVMM <= 0.0 Short: None. Period: 1-Jan-00 to 14-Apr-10 Metrics CAGR: 63.00% Sharpe: 1.25 Num Trades: 27 Average Trade: 11.69% % Winners: 51.85% DVR: 1.06 SPX Correlation: -1%

27 April 2010


11

Example 3: DVMM on WYNN,BIDU WYNN

Example Trades

WYNN

WYNN - Daily 25/10/2002 Open 13, Hi 13.15, Lo 12.85, Close 13.01 (0.0%)

70.0

63.68 60.0

50.0

40.0

30.0

20.0

Oct Dec 2009 Feb Mar Apr WYNN - DV Mean-Median Divergence(26,12,9) = {EMPTY}

May

Jun

Jul

Aug

Sep

Oct

Nov

Dec

2010

Feb 8.0 4.0 0.00000 0.0 -1.25154 -4.0 -8.0 -12.0


27 April 2010


12

Rolling Exponential Moving Average (REMA)

Type: Trend Indicator

SPY - Daily April 2010 REMA 20

Platforms Excel

SMA 20

Baseline Strategy Long when rising, Short when falling Suggested Usage …

Sep 2009

Oct 2009

Description

Nov 2009

Dec 2009

Jan 2010

Feb 2010

Mar 2010

Apr 2010

“The EMA Alternative”

Jeff Pietsch, of Market Rewind fame, is the brain child behind this EMA alternative1. This is an exponential moving average that creates a “rolling memory” as new price data is added to make it slightly less weighted than a standard EMA. It can be used to smooth indicators or as a replacement for a moving average in a trend strategy. Our preferred use is to apply the REMA to a momentum or ROC series to enhance responsiveness for relative strength applications. The DVST uses the REMA in this format and can be applied also as a trend indicator.

1. See http://marketrewind.blogspot.com/2009/02/solving-ema-backtest-dilemma.html 27 April 2010


13

Rolling Relative Strength Index (RRSI) Type: Mean Reversion

SPY - Daily April 2010

Platforms Excel Baseline Strategy Long <= 30, Short >= 70 Suggested Usage …

Feb 2010

Mar 2010

Apr 2010

SPY - Rolling Relative Strength Index

80 50 20

Description

“A More Responsive RSI”

The RRSI uses the REMA1 in the place of the conventional average in the RSI calculation. This makes it a more responsive price oscillator and superior to the RSI in many respects. All other features are identical to the RSI. Users can select their preferred time period, such as a 2-period.

1. See http://marketrewind.blogspot.com/2009/02/solving-ema-backtest-dilemma.html 27 April 2010


14

DV Fractal RSI (DVFR) Type: Intermediate Mean Reversion


Platforms Excel Baseline Strategy Long >= 50, Short < 50 Suggested Usage …

Feb 2010

Mar 2010

Apr 2010

SPY - DV Fractal RSI

80 50 20

Description

Level 1 Adaptive

“The Level 1 Adaptive RSI”

The Fractal RSI is a “Level 1” adaptive indicator that dynamically shifts weight in the RRSI between 2 and 30 days by default. This shifting is done to accommodate changes in volatility and fracticality in the market price data. However users may select any two periods in between these bounds. The longer term bound (30 days by default) is always set to have a trend influence on the shorter-term bound. Some good parameter choices you may wish to explore include 2/30, 2/15, 3/30,3/20, 10/30, 14/30.

27 April 2010


15

Example 4: DVFR on SPY,IWM SPY

26.17% CAGR Since 2000 Ticker: SPY

Equity Curve (Compounded) 1,000%

Rules: Long: Enter DVFR > 50, Exit DVFR < 50 Short: DVFR < 50, Exit DVFR > 50

800% 600%

Period: 18-Apr-00 to 23-Apr-10

400%

Metrics CAGR: 26.17% Sharpe: 1.20 DVR: 1.06

200% 0% 2000

IWM

2001

2002

2003

2004

2005

2006

2007

2008

2009

2010

20.60% CAGR Since 2001 Equity Curve (Compounded)

Ticker: IWM

500%

Rules: Long: Enter DVFR > 50, Exit DVFR < 50 Short: DVFR < 50, Exit DVFR > 50

400% 300%

Period: 4-Jun-01 to 23-Apr-10

200% Metrics CAGR: 20.60% Sharpe: 0.79 DVR: 0.70

100% 0% 2001 -100%

27 April 2010

2002

2003

2004

2005

2006

2007

2008

2009


16

DV Adaptive RSI (DVAR) Type: Intermediate Mean Reversion

^GSPC - Daily April 2010

Platforms Excel Baseline Strategy Long: Falling or <= 50, Short: Rising or > 50 Suggested Usage … Feb 2010

Mar 2010

Apr 2010

^GSPC - DV Adaptive RSI

50

Description

Level 2 Adaptive

“The Level 2 Adaptive RSI”

The Adaptive RSI is a “Level 2” adaptive indicator uses the same bounds as the Fractal RSI, but determines weight based on relative profitability within the bound spectrum instead of optimization. The Adaptive RSI performs well across a broad array of instruments and does a very good job of adapting considering its limited input. It requires a lot of data to calculate and as such is only available in the Excel Plug-In.

27 April 2010


17

Example 5: DVAR on ^GSPC ^GSPC

22.64% CAGR Since 1952

Ticker: ^GSPC Rules: Long: Enter DVAR < .50, Exit DVAR > .50 Short: DVAR > .50, Exit DVAR < .50 Period: 15-Jan-52 to 23-Apr-10 Metrics CAGR: 22.64% Sharpe: 1.49 DVR: .87

27 April 2010


18

DV Aggregate-M (DVAM) Type: Intermediate/Long Blended


Platforms Excel Baseline Strategy Long >= 0.50, Short < 0.50 Suggested Usage …

Feb 2010

Mar 2010

Apr 2010

SPY - DV Aggregate M

0.8 0.5 0.2

Description

“The Composite Trend / Mean-Reversion Indicator”

The DV Aggregate M is a composite trend/mean-reversion indicator that is based on integrating the short-term and long-term price distributions. The AggM assumes that prices are noisier and tend to mean-revert in the short to intermediate term, but tend to trend over longer time periods. It has been shown to be effective across a wide array of markets and stocks and is one of the more general/robust DV Indicators.

27 April 2010


19

Example 6: AggM on SPY, EWZ SPY

14.43% CAGR Since 2000 Ticker: SPY Rules: Long: Enter AggM > 50, Exit AggM <= 50 Short: Enter AggM <= 50, Exit AggM > 50 Period: 1-Jan-00 to 14-Apr-10 Metrics CAGR: 14.43% Sharpe: 0.57 Num Trades: 171 Average Trade: 0.89% % Winners: 66.08% DVR: 0.47 SPX Correlation: -38%

EWZ

26.50% CAGR Since 2001 Ticker: EWZ Rules: Long: Enter AggM > 50, Exit AggM <= 50 Short: Enter AggM <= 50, Exit AggM > 50 Period: 1-Jan-01 to 14-Apr-10 Metrics CAGR: 26.50% Sharpe: 0.73 Num Trades: 136 Average Trade: 2.11% % Winners: 67.65% DVR: 0.56 SPX Correlation: -9%

27 April 2010


20

DV Adaptive Agg-M (DVAA) Type: Intermediate/Long Blended

^GSPC - Daily April 2010

Platforms Excel Baseline Strategy Long >= 0.50, Short < 0.50 Suggested Usage …

Feb 2010

Mar 2010

Apr 2010

^GSPC - DV Adaptive Aggregate M 0.5

Description

Level 2 Adaptive

“The Level 2 Adaptive Aggregate M”

The Adaptive Aggregate M is a “Level 2” adaptive indicator like the Adaptive RSI that determines weight based on relative profitability within the bound spectrum but constrains the long-term setting to avoid changing the nature of the original indicator. It can be combined with the original Aggregate M to give more accurate signals.

27 April 2010


21

DV Intermediate Stretch (DVIS) SPY - Daily 01/04/2010 Close 117.8

121.19


December 2010 February SPY - DV Intermediate Stretch(10, 252) = 0.24

March

April 0.932271 0.8

0.2 Created with AmiBroker - adv anced charting and technical analy sis sof tware. http://www.amibroker.com

Description

“The Stretch Isolator”

This is one component of the DVI that isolates the stretch in terms of net days up or down over a series of periods and is re-scaled to create an accurate and consistent measure. The stretch is not like RSI because it does not consider the magnitude of up or down days but rather the net difference. It also looks at the short, intermediate and long time frames with a net weighting on the intermediate. This is a very accurate and useful indicator for identify mean-reversion areas of value rather than actual turning points. It can also be used as a trend measurement in conjunction with the ADX (Average Directional Movement).

27 April 2010


22

DV Intermediate Magnitude (DVIM) SPY - Daily 29/03/2010 Close 117.32

121.19


Oct Dec 2010 SPY - DV Intermediate Magnitude (5, 252) = 0.40

Feb

Mar

Apr 0.8 0.613546 0.2


Description

“The Magnitude Isolator”

This is the other component of the DVI that looks at the magnitude of up and down movements and is smoothed to reduce noise and also re-scaled to create an oscillator that appears to move in a “wave” format. The DVIM is a great complement to the DVIS because it measures the net percentage move over a variety of look backs weighted primarily on the intermediate time-frame. Like the DVIM, it does not attempt to identify turning points, and is instead a measure of relative value that does not change as frequently as other oscillators. The “wave” movement of the indicator is ideal for “hook” type mean-reversion strategies that buy or sell following a transition from overbought or oversold levels. It can also be lengthened to create an excellent trend or relative strength indicator.

27 April 2010


23

Example 7: DVIS/DVIM on SPY DVIS: SPY

18.65% CAGR Since 2001 Ticker: SPY Rules: Long: Enter DVIS < 0.50, Exit DVIS >= 0.50 Short: Enter DVIS >= 0.50, Exit DVIS < 0.50 Period: 1-Jan-01 to 14-Apr-10 Metrics CAGR: 18.65% Sharpe: 0.78 Num Trades: 245 Average Trade: 0.71% % Winners: 68.57% DVR: 0.54 SPX Correlation: -18%

DVIM: SPY

13.92% CAGR Since 2001 Ticker: SPY Rules: Long: Enter DVIM< 0.50, Exit DVIM >= 0.50 Short: Enter DVIM >= 0.50, Exit DVIM < 0.50 Period: 1-Jan-01 to 14-Apr-10 Metrics CAGR: 13.92% Sharpe: 0.54 Num Trades: 271 Average Trade: 0.51% % Winners: 68.63% DVR: 0.48 SPX Correlation: 8%

27 April 2010


24

DV Trend Stochastic (DVTS) SPY - Daily 20/01/2010 Close 113.89

Type: Intermediate Trend Platforms Amibroker, Tradestation, Excel 121.19

Baseline Strategy Long: Falling and/or >= 0.50, Short: Rising and/or < 0.50 Suggested Usage Short when it hits 0.20 from above, and long when it hits 0.80 from below

A S O N D 2008 M A M J J A S O N D 2009 SPY - DV Trend Stochastic(10) = 0.80

A M J J A S O N D 2010

A

0.83337 0.8 0.5 0.2


Description

“A Super-Smoother Stochastic”

The Trend Stochastic is a super-smoothed 10-period stochastic that moves gradually relative to the standard indicator developed by George Lane. It borrows from concepts used to smooth noisy data such as momentum. As a consequence it can be used as a trend indicator or a filter for meanreversion trades. It is a number scaled between 0 and 1, where .5 is not necessarily the median point. It can be traded long anywhere above .4 to .6, and short below. Or it can be traded by observing whether the stochastic is rising or falling, and potentially a combination of both the direction of movement and relative position.

27 April 2010


25

DV Trend Minus Cycle (DVTO) SPY - Daily 01/04/2010 Close 117.8

121.19

Type: Short/Intermediate Trend Platforms Amibroker, Tradestation, Excel Baseline Strategy Long >= 0.50, Short < 0.50 Suggested Usage …

May Jul Aug Sep Oct SPY - DV Trend Cycle Oscillator(3, 252) = 0.29

Nov

Dec

2010

Feb

Mar

Apr 0.59761 0.5


Description

“The De-Cycled Trend Indicator”

The Trend-Minus Cycle oscillator is a transform that nets the difference between the trend stochastic and a calculation related to the DVDS or supersmoothed double stochastic. The DVTO is useful for markets where an intermediate trend signal is hidden or obscured by a cyclic component. This trend may occur at medium frequencies such as the S&P500 or at extremes such as the case with Oil. If the DVTO is highly unprofitable on a given market that is usually because it is mean-reverting at the intermediate level—this occurs in markets such as natural gas or gold stocks. The benefit of the DVTO is that you can use it to filter intermediate trend signals such as in the MACD or DVMM, or you can use it to combine with longer term trend indicators to create a more accurate multiple time frame system.

27 April 2010


26

DV Super-Smoothed Double Stochastic (DVDS) SPY - Daily 01/04/2010 Close 117.8

121.19

Type: Short Term/Intermediate Mean Reversion Platforms Amibroker, Tradestation, Excel Baseline Strategy Long <= 0.20, Short >= 0.80 Suggested Usage …

Jul Aug Sep Oct Nov SPY - DV Double Stochastic(10) = 0.41

Dec

2010

Feb

Mar

Apr

0.9 0.891468


Description

“A Super-Smooth Double Stochastic”

The super-smoothed double stochastic is the preferred mean-reversion variant to the conventional stochastic by George Lane. The DVDS normalizes the 10-period stochastic position within the channel to increase peak/valley classification accuracy and is smoothed twice to make it less prone to whipsaws. It does not use a percentile rank classification which tends to increase median accuracy versus extreme accuracy.As a consequence is a good compliment to shorter –term oscillators like the DV2, DV Stochastic, or RSI2, or even intermediate oscillators like the DVI as it helps to increase the odds of finding a temporary bottom or top. DVDS levels below 10 and above 90 often coincide with peaks/valleys within a few days. With the broader trend, levels below 20 and above 80 tend to be the appropriate signal levels.

27 April 2010


27

DV Stochastic (DVS) SPY - Daily 01/04/2010 Close 117.8

121.19

Type: Sort Term/Intermediate Mean Reversion Platforms Amibroker, Tradestation, Excel Baseline Strategy Long <= 0.50, Short > 0.50 Suggested Usage …

Sep Nov Dec SPY - DV Stochastic (2, 252) = 0.45

2010

Feb

Mar

Apr 0.8 0.710344 0.2


Description

“The Short-Term Stochastic”

This is the short-term cousin of the DVDS and resembles a stochastic version of the dv2. Like a conventional stochastic it permits highly profitable entries from oversold/overbought levels when the indicator is rising/falling, and this is a much lower risk entry than classic rsi2 and dv2 variants. It isn’t designed as much for binary use, although the binary version responds better to adaptation than DV2 or RSI2 variants because it is more predictable in its oscillation. Itdoes use the percentile rank to permit a consistent number of entries.

27 April 2010


28

DV Bands (DVBU/DVBL/DVBM) 121.19 120.825 119.289

SPY - Daily 01/04/2010 Close 117.8 DV Bands(20,252,0.975, 0.5, 0.025) = 117.38

115.191

November 2010 SPY - DV Band Indicator(20,252) = 0.70

February

March

April

Type: Charting Indicator Platforms Amibroker, Tradestation, Excel Suggested Usage …

0.992032 0.9 0.5


Description

“The Rescaled Bollinger Band”

The DV Lower Band (DVBL) price is similar to the lower band of a standard Bollinger Band but is re-scaled to the annual lower band level frequency. The DV Middle Band (DVBM) price is similar to the middle band of a standard Bollinger Band--but is not the average price but rather the re-scaled average price. The DV Upper Band (DVBU) price is similar to the upper band of a standard Bollinger Band but is re-scaled to the annual upper band level frequency. DV bands ‘contain’ price more effectively because they utilise the actual distribution of historical prices (unlike Bollinger Bands which assume a normal distribution).

27 April 2010


29

DV Band Indicator and Percentile (DVBI/DVBP)

SPY - Daily 01/04/2010 Close 117.8 DV Bands(20,252,0.975, 0.5, 0.025) = 117.38

121.19 120.825 119.289

Type: Intermediate Mean Reversion

115.191

Baseline Strategy Long <= 0.10, Short > 0.90

Platforms Amibroker, Tradestation, Excel

Suggested Usage …

November 2010 SPY - DV Band Indicator(20,252) = 0.70

February

March

April

0.992032 0.9 0.5


Description

“The Rescaled Bollinger Band”

This is the DV Band Indicator for DV Bands. For the Excel Plug-In this is the percentile rank of the H,L,C version of the z-score. The DVBP returns the price associated with a user-defined percentile and H,L,C range. Thus if you wanted to know what the price of the S&P500 would be at the 95th percentile based on the last 30 days of prices you would highlight the H,L,C range over the past “n” days and type in “95%.” In the case, with price data in a spreadsheet in descending order, you would type in as follows: =DVBP(C3:E32,95%)=$119.13 Both the DVBP and DVBI can be used as probability-based indicators, where the likelihood of exceeding the absolute extremes the next day is very low. In our testing the upper and lower DV Bands contained up to 95% of closes out of sample—that is, the chances of exceeding the upper or lower band or 97.5th and 2.5th percentiles were roughly equivalent to their promised probability (2.5+100-97.5=5% of values outside the range). This makes DV Bands a good tool for mean-reversion strategies and/or filtering for abnormal market conditions.

27 April 2010


30

Example 8: DV Bands Price Containment

Study The following table empirically tests the DV bands ability to ‘contain’ price, as compared to the original Bollinger Bands ability to contain price. It is clear from these results that DV Bands do in fact contain significantly more of a stock’s price movement than Bollinger Bands. For example, the original Bollinger Bands contain 82.4% of SPY prices (i.e. the close is within the bands), whereas DV Bands contain 89.6% of SPY prices. Bollinger Band DV Bands Ticker Bars Containment Containment Improvement SPY XLB XLE XLF XLI XLK XLP XLU XLV XLY

27 April 2010

4333 2843 2843 2843 2842 2843 2843 2843 2843 2843

82.4% 83.0% 82.7% 82.9% 81.9% 81.7% 83.1% 82.1% 82.3% 82.4%

89.6% 89.8% 90.1% 90.3% 89.8% 90.1% 89.6% 90.4% 89.9% 89.7%

7.2% 6.8% 7.4% 7.4% 8.0% 8.4% 6.5% 8.3% 7.6% 7.3%


31

DV Zones (DVZN) for Excel Plug-in Only Type: Environmental Filter Platforms Excel Only

Description

“Define Market Regimes Empirically”

This is one method of identifying regimes, in this case DV Zones are constructed using DVRAC- a trend filter that uses multiple smoothed regression slopes that are smoothed with a 30-day measurement periods, and DVPV which is the percentile rank of 30-day historical volatility. The following table shows the corresponding number codes and indicator values used to de-lineate the zones:

Regime up trend low volatility up trend high volatility no trend low volatility no trend high volatility down trend low volatility down trend high volatility

Zone Number 1 2 3 4 5 6

Trend (DVRAC) > 0.2 > 0.2 > -0.2, < 0.2 > -0.2, < 0.2 < - 0.2 < - 0.2

Volatility (DVPV) <.5 >.5 <.5 >.5 <.5 >.5

DV Zones are excellent for developing comprehensive trading systems. They are static regimes that capture a broad spectrum of market conditions. You also may wish to shift a portfolio allocation to various systems dynamically based on the current zone position. The most common use of zones is to understand the performance of a given indicator within each zone, since they will be very different. It is instructive in some cases to consider longs separately from shorts. For example, in up trends with low volatility, shorting using a short-term indicator like the RSI2 or DV2 might not be desirable. The same may apply with going long in down trends with low volatility. As always taking a smaller position size is the least risky alternative if you have a good system to avoid missing out. The zone concept can be applied using other indicators as well such as the 200 day moving average. DV Zones allow the user to create trading systems for each market regime, as opposed to one system for all market regimes.

27 April 2010


32

DV Super-Charged Percent Exposure (DVSE) SPY - Daily 14/04/2010 Close 121.19

121.19

Type: Short Term Mean Reversion Positing Sizing Platforms Amibroker, Tradestation, Excel Baseline Strategy Position Size = Normal Position Size * DVSE Exposure % Suggested Usage …

November December 2010 February SPY - DV Super Charged Pct Exposure(1,2,100,22) = 1.27

March

April 1.26666 1.1


Description

“Determine exposure using the DVSC”

This is the DV Super-Charged DV2 percent exposure model that varies between 150% and -150% based on a simple algorithm. The DVSE increases exposure on both the long and short side at extremes, and normalizes its exposure positioning to keep size proportionate.

27 April 2010


33

DV Bounded Percent Exposure (DVSE) SPY - Daily 14/04/2010 Close 121.19

121.19

Type: Short Term Mean Reversion Positing Sizing Platforms Amibroker, Tradestation, Excel Baseline Strategy Position Size = Normal Position Size * DVSE Exposure % Suggested Usage …

November December 2010 February SPY - DV Bounded Pct Exposure(0,2,100,22) = 1.29

March

April 1.2866 1.1


Description

“Determine exposure using the DVB”

This is the DV2 percent exposure model that varies between 150% and -150% based on a simple algorithm. The DVSE increases exposure on both the long and short side at extremes, and normalizes its exposure positioning to keep size proportionate.

27 April 2010


34

Daily Historical Volatility (DHV) SPY - Daily 14/04/2010 Close 121.19

Type: Environmental Filter Platforms Amibroker, Tradestation, Excel 121.19

2007 2008 SPY - Daily Historical Volatility = 8.59

2009

Suggested Usage …

2010


Description

“User Adjustable Historical Volatility”

This is the standard historical volatility calculation with a user defined measurement period. The default setting is 30 days. The calculation is the standard deviation of the natural logarithm of price changes over the period selected scaled to 1 year as defined by 252 trading days using the square root rule. It has been multiplied by 100 to make it comparable to the VIX. When the VIX is trading above this number it is bullish for the market and vice versa based on our historical research.

27 April 2010


35

DV Percentile Rank Volatility (DVPV) SPY - Daily 14/04/2010 Close 121.19

121.19

Type: Mean Reversion / Trend Filter Platforms Amibroker, Tradestation, Excel Baseline Strategy Mean Reversion environment when > 0.50, Trend environment when < 0.50 Suggested Usage …

N D 2009 M A M J J A SPY - DV Percent Rank Volatility(20, 252) = 5.98

S

O

N

D

2010

M

A


Description

“Normalised Historical Volatility”

This is the percentile rank of DVHV with a 30 day default setting and a 252 day look back for the normalization. The DVPV is an excellent tool for system filtering and testing as well as position sizing and portfolio allocation. Primarily it should be considered as a key measure to watch when deciding between mean-reversion and trend-strategies. It is also a key component of “zones” analysis (see DVZN).

27 April 2010


36

DV Composite Volatility (DVCV) SPY - Daily 14/04/2010 Close 121.19

Type: Mean Reversion / Trend Filter Platforms Amibroker, Tradestation, Excel 113.33

Baseline Strategy Mean Reversion environment when > 0.50, Trend environment when < 0.50 Suggested Usage …

J S O N D 2009 M SPY - DV CompositeVolatility(5, 252) = 0.62

A

M

J

J

A

S

O

N

D

2010

0.5 0.0398406 Created with AmiBroker - adv anced charting and technical analy sis sof tware. http://www.amibroker.com

Description

“A Composite Volatility Measure”

This is a composite volatility measure that includes ratio volatility , longer term historical volatility, and a measure of daily variation. It can be used as mean-reversion filter on its own or combined with DVPV. A rising DVCV over the past week or 5 days and a reading above the lowest quartile or .25 is more favorable for short-term mean-reversion on an absolute and risk-adjusted basis. Day-traders may also want to pay attention to the DVCV.

27 April 2010


37

DV Breakout, Composite Volatility Plus, Composite Volatility Minus (DVBR/DVCP/DVCM) SPY - Daily 14/04/2010 Close 121.19

121.19

Type: Intermediate Mean Reversion Platforms Amibroker, Tradestation, Excel Baseline Strategy Long: >= 0.0, Short: < 0.0 Suggested Usage …

May Jun Jul Aug Sep SPY - DVCV(P+M)(10, 252) = 0.06

Oct

Nov

Dec

2010

Feb

Mar

Apr 0.0629482 0.00000


Description

“The Anti-Thesis to the DV2”

All of these indicators are related- the DVCP and DVCM are the components of the DVBR which is a breakout indicator. Both the DVCP and DVCM measure short-term range expansions, with the DVCP (composite plus) making a positive range expansion from lower volatility and the DVCM (composite minus) making a downside or negative range expansion from lower volatility. The DVBR is a breakout indicator that is by default the 8day exponential moving average of the difference between DVCP and DVCM. The DVBR was highly profitable in the pre-mean reversion days on the S&P500 and recently made a resurgence during the rally. It is the anti-thesis to DV2 and many others, and tends to perform well on trendy stocks that have high LTR ratings such as in the Livermore index. It can also be used to screen/filter DV2 trades or any other mean-reversion trades or screen for trend entries along with DVRAC.

27 April 2010


38

DV R-Squared Autocorrelation (DVRAC) SPY - Daily 14/04/2010 Close 121.19

121.19

Type: Intermediate Trend Filter Platforms Amibroker, Tradestation, Excel Baseline Strategy Long: >= 0.20, Short: < -0.20 Suggested Usage …

Aug Oct Nov Dec SPY - DV R-Squared AutoCorrelation = 0.81

2010

Feb

Mar

Apr 0.807997 0.00000


Description

“The Premier Trend Filter”

This is the premier trend filter used within the 6 different regime “Zones” that reflected different combinations of trend and volatility. The DVRAC utilizes a slope-based R-squared correlation of prices that captures the Highs, Lows and Closes. Smoothing is applied to the indicator which minimizes false signals but also has the disadvantage of producing lag. DVRAC levels above .2 show a statistically significant positive trend which helps to provide a minimum criteria for placing buys to avoid false entries. Levels between -.2 and .2are indicative of temporary uncertainty in the trend condition, and this area is often called the “No-Trend” zone. Levels of the DVRAC below -.2 show a statistically significant intermediate down trend. This must be analyzed within the context of the long-term trend to consider going short.

27 April 2010


39

DV Differential Autocorrelation (DVDA) Type: Environmental Filter


Platforms Excel

Suggested Usage …

Feb 2010

Mar 2010

Apr 2010

SPY - DV Differential Autocorrelation 80 50 20

Description

“A Trend / Mean-Reversion Filter”

The differential autocorrelation indicator is a filter that detects shifts in the difference in autocorrelation between closing prices and high versus low prices. This is significant as a trend/mean-reversion filter since a strong trend will tend to have persistently higher closes but variation in highs and lows. Levels of DVDA well below .25 characterize short-term random drift in the market and indicator signals are less valuable in this area. In general, a falling DVDA over 1-5 days is a mean-reversion environmental signal. A rising DVDA signals favorable conditions for the prevailing trend.

27 April 2010


40

DV Composite Fractal Efficiency (DVFE) Type: Intermediate/Long Term Blend


Platforms Excel Baseline Strategy Long: >= 0.75, Short: < 0.25 Suggested Usage …

Feb 2010

Mar 2010

Apr 2010

SPY - DV Composite Fractal Efficiency 0.75

0.25

Description

“Measure Fractal Efficiency At Two Time Frames”

This is a composite mean-reversion and trend indicator that measures fractal efficiency at two different time frames. The DVFE assumes that meanreversion resides in the short-term and trends exist in the long term similar to the Aggregate M indicator. The signals are not highly correlated to the Agg M even though the default periods are nearly the same. This is because the DVFE is best traded at extremes of >.75 for longs and <.25 for shorts. The binary DVFE is not nearly as impressive as the Agg M, which naturally introduces the possibility of taking extreme DVFE signals and using binary Agg M to take lower level signals.

27 April 2010


41

DV Smoothed Trend (DVST) SPY - Daily 14/04/2010 Close 121.19

121.19

Type: Environmental Filter Platforms Amibroker, Tradestation, Excel Baseline Strategy Long: >= 0.0, Short: < 0.0 Suggested Usage …

Aug Oct SPY - DV Trend = 5.67

Nov

Dec

2010

Feb

Mar

Apr 5.66667 4 -4


Description

“A REMA Smoothed Momentum Measure”

The DV Smoothed Trend is a momentum measure that is smoothed twice using REMA or the rolling exponential moving average. The use of momentum/velocity assures that lag is minimized, and the use of double smoothing reduces the noise component. The DVST can be used as either a relative strength indicator, or a trend indicator.

27 April 2010


42

DV Self-Adaptive Slow, DV Self-Adaptive Fast ( DVAS / DVAF ) Type: Adaptive Buy/Sell Levels Platforms Excel Suggested Usage …

Level 1 Adaptive

Description

“A Level 1 Self-Adaptive Overlay For Existing Indicators”

Both DVAS and DVAF are Level 1 class self-adapters that resample the prevailing distribution of the underlying to match with indicator signals to determine whether the best direction to trade an indicator is mean-reversion “MR” or trend “TR.” They can be applied with an indicator chosen by the user, but are best suited to shorter-term or intermediate term indicators. The self-adapter also determines the best levels to buy and sell using a given indicator by quickly adapting to the new distribution of returns for the underlying. The DVAS is less sensitive and tends to change more slowly while the DVAF rapidly adjusts to current conditions and potentially noise. There is a trade-off to using both and the best way is to use one in conjunction with the other.

27 April 2010


43

DV Self-Adaptive Slow, DV Self-Adaptive Fast ( DVAS / DVAF ) Type: Adaptive Buy/Sell Levels Platforms Excel Suggested Usage …

Level 1 Adaptive

Description

“A Level 1 Self-Adaptive Overlay For Existing Indicators”

Both DVAS and DVAF are Level 1 class self-adapters that resample the prevailing distribution of the underlying to match with indicator signals to determine whether the best direction to trade an indicator is mean-reversion “MR” or trend “TR.” They can be applied with an indicator chosen by the user, but are best suited to shorter-term or intermediate term indicators. The self-adapter also determines the best levels to buy and sell using a given indicator by quickly adapting to the new distribution of returns for the underlying. The DVAS is less sensitive and tends to change more slowly while the DVAF rapidly adjusts to current conditions and potentially noise. There is a trade-off to using both and the best way is to use one in conjunction with the other.

27 April 2010


44

CSS Analytics Adaptor One (CSSA1) Type: Adaptive Buy/Sell Levels


Platforms Excel Baseline Strategy Long: >= 0.50, Short: < 0.50 Suggested Usage …

Feb 2010

Mar 2010

Apr 2010

SPY - CSS Analytics Adaptor One 0.5

Description

Level 1 Adaptive

“The First Institutional Class Adaptive DV Indicator”

This is the first institutional class entrant of the DV Indicators - it is a Level 1 Adaptive short-term mean-reversion indicator that looks at multiple time frames and different measures of volatility and fracticality. It dynamically self-adjusts to account for these factors to avoid making costly errors in buy/sell decisions. CSSA is more robust than the standard DV Indicators and works on a wider range of markets and stocks - especially entering at more extreme levels. It is a small preview of the future which will include Level 3 and Level 4 class that will have multiple layers of adaptation and self-adjustment. CSSA is the frame of a new class of indicators that will represent compact artificial intelligence machines.

27 April 2010


45

Example Applications

27 April 2010

System A

DV SPY Swing System

System B

DV Combo System


46

System A: DV SPY Swing System Description

The DV SPY Swing System combines

[......... Rules Provided with subscription ................]

Long

Entry

Exit

Zone 1

Zone 2

Rising SMA200 and DVDS10 < 20

Falling SMA200 and DVDS10 < 20 and DVO <= 20

DVDS10 >Rules 80 DVO > 50 [......... Provided with subscription Short ................]Zone 2 Zone 1

Entry

none

Falling SMA200 and DVDS10 < 20

Exit

none

DVDS10 > 80

27 April 2010


47

DV SPY Swing System (Continued) Amibroker Code // ======================================================================================== // DV Indicators by David Varadi http://cssanalytics.wordpress.com/ // Copyright 2009-2010 David Varadi/ CSS Analytics, All Rights Reserved // For Personal Use by Registered Users Only; NOT for Redistribution; No Implied Warranties // ======================================================================================== SetFormulaName("System A - DV SPY Swing System"); SetBarsRequired(10000,10000); SetOption("CommissionMode", 3); SetOption("CommissionAmount", 0.00); SetOption("InitialEquity", 100000); SetOption("MaxOpenPositions", 1); SetOption("PriceBoundChecking", 1); SetOption("AllowPositionShrinking",True); SetTradeDelays( 0, 0, 0, 0); PositionScore = 1/C; PositionSize = -100; BuyPrice = SellPrice = ShortPrice = CoverPrice = Close; UpTrend = MA(C,200) > Ref(MA(C,200),-1); DownTrend = MA(C,200) <= Ref(MA(C,200),-1); DVDS_10Period = DVDS(H,L,C,10); DVO_indicator = DVO(H,L,C);

[......... Code Provided with subscription ................]

// Long Signals Entry1 = UpTrend AND (DVDS_10Period < 0.20); Entry2 = DownTrend AND (DVDS_10Period < 0.10) AND (DVO_indicator <= 0.20); Buy = Entry1 OR Entry2; Exit1 = UpTrend AND (DVDS_10Period > 0.80); Exit2 = DownTrend AND (DVO_indicator > 0.50); Sell = Exit1 OR Exit2; // Short Signals Short = DownTrend AND (DVO_indicator > 0.50); Cover = DVO_indicator < 0.50; NOTE: Results differ marginally between Amibroker and Excel because of different trade management techniques 27 April 2010


48

System B: DV Combo System Description The DV Combo system combines short-term mean reversion, using the DVSC2, with a longer term composite trend and mean-reversion indicator, namely the AggZ. The benefit of this approach is that the system will stay long until both time-frames dictate that the position should be flipped – this reduces the number of trades thereby avoiding chop and the associated trading costs. This system is intended for the S&P500 spider, the “SPY”. For those not familiar with the AggZ, the calculation is dead simple: AggZ= (-1x( 10-day z-score)+(200-day z-score))/2, where z-score = (close-sma (closing prices over last n periods))/(standard deviation( closing prices over last n periods))

[......... Rules Provided with subscription ................]

Combo Score Calc > 0, add 1

AggZ < 0, minus 1 < 50, add 1

DVSC2[.........

Combo Score

Rules Provided with subscription ................] > 50, minus 1

System Rules Long

Short

Entry

Combo Score > 0

Combo Score <= 0

Exit

Combo Score <= 0

Combo Score > 0

27 April 2010


49

DV Combo System (Continued) Amibroker Code // ======================================================================================== // DV Indicators by David Varadi http://cssanalytics.wordpress.com/ // Copyright 2009-2010 David Varadi/ CSS Analytics, All Rights Reserved // For Personal Use by Registered Users Only; NOT for Redistribution; No Implied Warranties // ======================================================================================== SetFormulaName("System B - DV Combo System"); SetBarsRequired(10000,10000); SetOption("CommissionMode", 3); SetOption("CommissionAmount", 0.00); SetOption("InitialEquity", 100000); SetOption("MaxOpenPositions", 1); SetOption("PriceBoundChecking", 1); SetOption("UsePrevBarEquityForPosSizing", True); SetOption("AllowPositionShrinking",True); SetTradeDelays( 0, 0, 0, 0); PositionScore = 1/C; PositionSize = -100; BuyPrice = SellPrice = ShortPrice = CoverPrice = Close; function zscore(price,Length) { av = MA(price,Length); st = StDev(price,Length); zs = (price - av) / st; return zs; }


AggZ = (-zscore(C,10) + zscore(C,200) )/2; myDVSC = DVSC(H,L,C, 2, 252); ComboScore = IIf(AggZ>0,1,-1) + IIf(myDVSC<0.50,1,-1); // Long System Buy = ComboScore > 0; Sell = ComboScore <= 0; // Short System Short = ComboScore < 0; Cover = ComboScore >= 0; NOTE: Results differ marginally between Amibroker and Excel because of different trade management techniques 27 April 2010


50

Appendices

27 April 2010


51

Appendix 1: Amibroker Code Example 1: DV2 v RSI2 // ======================================================================================== // DV Indicators by David Varadi http://cssanalytics.wordpress.com/ // Copyright 2009-2010 David Varadi/ CSS Analytics, All Rights Reserved // For Personal Use by Registered Users Only; NOT for Redistribution; No Implied Warranties // ======================================================================================== SetFormulaName("Example 1 - DV2 v RSI2"); SetBarsRequired(10000,10000); SetOption("CommissionMode", 3); SetOption("CommissionAmount", 0.00); SetOption("InitialEquity", 100000); SetOption("MaxOpenPositions", 1); SetOption("PriceBoundChecking", 1); SetOption("AllowPositionShrinking",True); SetTradeDelays( 0, 0, 0, 0); PositionScore = 1/C; PositionSize = -100; BuyPrice = SellPrice = ShortPrice = CoverPrice = Close; Value1 = Param("SystemToRun",1,1,2,1); if(Value1 == 1) { // RSI2 SYSTEM Buy = RSI(2) <= 50; Sell = RSI(2) > 50; Short = RSI(2) > 50; Cover = RSI(2) <= 50; } else { // DVB SYSTEM myDVB = DVB(H,L,C,2,252); Buy = myDVB <= 0.50; Sell = myDVB > 0.50; Short = myDVB > 0.50; Cover = myDVB <= 0.50; }

27 April 2010



52

Appendix 1: Ami Code Example 2: DVU,DVB,DVO & DVSC // ======================================================================================== // DV Indicators by David Varadi http://cssanalytics.wordpress.com/ Copyright 2009-2010 David Varadi/ CSS Analytics, All Rights Reserved // For Personal Use by Registered Users Only; NOT for Redistribution; No Implied Warranties // ======================================================================================== SetFormulaName("Example 2 – DVU,DVB,DVO,DVSC"); SetBarsRequired(10000,10000); SetOption("CommissionMode", 3); SetOption("CommissionAmount", 0.00); SetOption("InitialEquity", 100000); SetOption("MaxOpenPositions", 1); SetOption("PriceBoundChecking", 1); SetOption("AllowPositionShrinking",True); SetTradeDelays( 0, 0, 0, 0); PositionScore = 1/C; PositionSize = -100; BuyPrice = SellPrice = ShortPrice = CoverPrice = Close; Value1 = Optimize("SystemToRun",1,1,4,1); if(Value1 == 1) { // DVU SYSTEM myDVU = DVU(H,L,C,2); Buy = myDVU <= 0; Sell = myDVU > 0; Short = myDVU > 0; Cover = myDVU <= 0; } else if (Value1 == 2) { // DVB SYSTEM myDVB = DVB(H,L,C,2,252); Buy = myDVB <= 0.50; Sell = myDVB > 0.50; Short = myDVB > 0.50; Cover = myDVB <= 0.50; } else if (Value1 == 3) { // DVO SYSTEM myDVO = DVO(H,L,C); Buy = myDVO <= 0.50; Sell = myDVO > 0.50; Short = myDVO > 0.50; Cover = myDVO <= 0.50; } else if (Value1 == 4) { // DVSC SYSTEM myDVSC = DVSC(H,L,C, 2, 252); Buy = myDVSC <= 0.50; Sell = myDVSC > 0.50; Short = myDVSC > 0.50; Cover = myDVSC <= 0.50; }


27 April 2010


53

Appendix 2: Amibroker Installation Description The Amibroker Plug-in package is easily installed. 1) 2)

Firstly close Amibroker, then simply double click on the provided installation file “DVPlugins.msi” and follow the installation prompts When the installation is complete, start up Amibroker, where you will be presented with the dialogue box below. This will only appear once, and should not be cause for concern.

3)

Within Amibroker, the AFL files for all the DV Indicators can be found under the Charts pane > Custom Folder > DV Indicators Folder

27 April 2010


54

DV Indicators Manual v1 Free

Recommend Documents