Business Whitepaper
Page |1
A Hedge Fund of Digital Assets backed by Artificial Intelligence
Business Whitepaper
Authors: Ayan Dasgupta, CEO and CIO, Guna Fund
April 17, 2018 Abstract The popularity of digital assets is ever-growing. While this space offers immense potential of growth and returns to an investor, it does not come without risks and a high requirement of mind share. The Guna Fund is a hedge fund of digital assets backed by artificial intelligence (AI) aimed at providing a low risk, high and stable return to its investors by relieving them of the hassles of tracking the market/assets, ensuring security and preventing them from committing individual trading mistakes. The use of AI is intended to optimize the performance of the fund.
Business Whitepaper
Page |2
1. Need for a Fund in Digital Assets The digital assets space has come of age in the past year or so. While Bitcoin was the first cryptocurrency to be developed in early 2009 by the pseudonymous Satoshi Nakamoto, today, over 30,000 digital assets exist. Each of these assets belong to different categories that offer a different use case and they can be broadly classified as under: Cryptocurrencies, like Bitcoin or Litecoin , which primarily serve the purposes of being a store of value, of value, a means of exchange, and a unit of account. Platforms, like Ethereum or NEO , which create an infrastructure that can be leveraged to build on top of. Ethereum, for example, allows developers to create smart contracts (code that can execute transactions without a third party or intermediary) on top of the platform the coin creates which can power decentralized applications. Protocols, like 0x or Kyber , are the set of rules that govern the network. Blockchain protocols usually include rules about consensus, transaction validation, and network participation. Decentralized Applications (DApps), like Augur or Cindicator , which represents the decentralized equivalent of an app or platform and are built on top of platforms like Ethereum.
The market capitalization of this space grew from under $20 billion in early 2017 to as high as $850 billion in January 2018, before retracing towards a consolidation level of around $250 billion (more than 1000% return). The potential of digital assets, backed by decentralized computing technologies like Blockchain, Hashgraph, etc. is immense and highly disruptive, and the market is expected to expand to over $1 trillion by 2020. 1.1 The Problem at Hand
While this offers a very lucrative opportunity for individuals and institutions to derive gains on their investments on these assets, the volatile nature of the market coupled with fraudulent projects in the space pose challenge to their profit realization. Many investors get swayed into investing in digital assets during the end of a hyper bull run only to witness their investments suffer loss, albeit notional, after a couple of months. They also see themselves at the wrong end of a scam ICO investment and their hard earned money is thus lost without being able to reap the long term benefits that this space has to offer. Even an investor who is well versed with this space faces issues of managing time to put into research of promising projects and track the markets, which are open 24/7, regularly to enter and exit at the ideal buying and selling opportunities and wishes for a solution to take care of such needs at times. 1.2 The Solution
To address the above mentioned problem, we present to you the Guna Fund, a hybrid hedge-cum-equity fund of digital assets backed by artificial intelligence. The fund’s management team will conduct due diligence and invest in only those projects and digital assets which pass their screening parameters. It will also try to maximize the fund’s value during bearish market cycles by taking hedge strategies that keep the fund’s position as neutral to market risks as possible. Investors of the Guna Fund will own a highly valuable portfolio of blockchain assets by investing in only one blockchain asset – the the Guna Fund Token (GNA). This solves the issue of actively choosing from a large number of blockchain assets and timely adjusting their portfolio for making gains.
2. The Guna Fund 2.1 Close Ended Fund The fund will be closed-ended since it is otherwise expected that its performance will attract substantial amounts of new capital. The fund ’s management team might be compelled to use that new capital to chase a relatively small group of assets, thereby creating a buying pressure which drives up the prices of those assets and thereby inorganically affects the market. This issue further affects the existing token holders by reducing the future profitability of the fund, since the fund ’s management will have to buy the assets at
Business Whitepaper
Page |3
a higher price than normal. A close ended fund also allows it to have a lower expense ratio courtesy the absence of a need to issue and distribute new tokens. This allows it to command less fees despite being an actively managed fund. 2.2 Actively Managed Fund
The entire fund will be actively managed. The fund management activities will mainly involve the following: Investment in altcoins Margin trading Investing in ICOs
The portion of the total fund to be allocated to each activity will vary depending on the market conditions. Participating in markets that offer returns on successful predictions on events pertaining to sports, films, elections, weather and other natural events is a plan for the future. 2.3 Use of Artificial Intelligence (AI)
To optimize the performance of the fund’s management, the team will deploy AI algorithms to obtain profitable trade signals. Details of the algorithms and their implementation are provided in the technical whitepaper. 2.4 Flow of Trading
The trading of assets held by the fund will be carried out by their fundamental, sentiment and technical analysis. 2.4.1 Fundamental Analysis
Fundamental analysis is a method of evaluating an asset in an attempt to measure its intrinsic value, by examining related economic, financial and other qualitative and quantitative factors. The end goal of fundamental analysis is to produce a quantitative value that an investor can compare with an asset ’s current price, thus indicating whether the asset is undervalued or overvalued. Rubrics will be developed by the fund management team and AI systems that are aimed to select the best projects. These may be copyrighted and published after due testing. Based on the rubrics used, a Net Fundamental Score (NFS) will be assigned to each asset. 2.4.2 Sentiment Analysis
Also known as opinion mining or emotion AI, sentiment analysis refers to the use of natural language processing, text analysis, computational linguistics, and biometrics to systematically identify, extract, quantify, and study affective states and subjective information. Generally speaking, sentiment analysis aims to determine the attitude of a speaker, writer, or other subject with respect to some topic or the overall contextual polarity or emotional reaction to a document, interaction, or event. The attitude may be a judgment or evaluation (see appraisal theory) theory), affective state (that is to say, the emotional state of the author or speaker), or the intended emotional communication (that is to say, the emotional effect intended by the author or interlocutor). In the parlance of trading blockchain assets, sentiment analysis involves analyzing tweets, subreddits, threads on forums, etc. about certain blockchain assets and forecasting the price movements from the posts. Details of the sentiment analysis approach are provided in the technical whitepaper. The AI component of the Guna Fund will perform sentiment analysis of the fund’s assets through the following flow:
Business Whitepaper
Page |4
2.4.2.1 Guna Retriever (GR)
Presently, cryptocurrency price movements mostly occur on the basis of news, rumors, and speculations. We call these events to be catalysts to the price movements. Some common catalysts are: a) b) c) d)
Listing on a new exchange Coin burn (reduction of supply) Partnerships with major existing businesses Updates and alpha, beta or stable releases, etc.
Whenever any project plans to exercise any such event, rumors inevitably surface through posts on social media such as Twitter, Reddit, Bitcointalk, Medium, and Steemit. The posts usually involve the use of the project’s cashtag, which is the ticker (symbol) of the currency prefixed by a dollar sign, e.g. $BTC for Bitcoin. The retriever component of the AI will retrieve rumors or news based on certain keywords corresponding to the previously mentioned catalysts. So, if we are looking to invest in projects which are in talks with big organizations for a potential partnership, rumors about such developments will have surfaced on social media, and posts related to such rumors will be captured by the GR. 2.4.2.2 Guna Validator (GV)
The authenticity of a rumor or news will be determined by the validator component of Guna ’s AI. Only those rumors that are backed with proof or are originating from sources with extended past credibility will be considered for further sentiment analysis. 2.4.2.3 Guna Sentiment Analyzer (GSA)
The posts returned by the GV will then be assessed by the sentiment analyzer. It will determine and filter the favorable rumors or news from those otherwise based on the aspects of news quantification detailed above. Post filtration, the projects will be ranked on the basis of their sentimental favorability and a Net Sentiment Score (NSS) will be assigned to each asset. The list of such projects will be forwarded to the Guna Technical Analyzer. 2.4.3 Technical Analysis
Technical analysis is a trading tool employed to evaluate assets and attempt to forecast their future movement by analyzing statistics gathered from trading activity, such as price movement and volume. Over the years, numerous technical indicators have been developed by analysts in attempts to accurately forecast future price movements. Some indicators are focused primarily on identifying the current market trend, including support and resistance areas, while others are focused on determining the strength of a trend and the likelihood of its continuation. The team will aim to develop their own technical analysis models assisted by the AI system, based on strategies like mean reversion or statistical arbitrage or devise a new technical paradigm, and these may be copyrighted and published after due testing. 2.4.3.1 Guna Technical Analyzer (GTA)
On receiving the recommendation from the GSA, the Technical Analyzer will evaluate the assets technically by assessing its charts, patterns, volume data, and momentum indicators to determine its ideal entry (buy) and exit (sell) prices. At the end of its analysis, it will assign a Net Technical Score (NTS) to each asset and also indicate the prices at which an asset should be bought or sold. 2.4.4 Trade Execution
Business Whitepaper
Page |5
The execution of trades will be performed by a mix of human and AI activity. The Net Asset Score (NAS) will be considered before buying or selling an asset. It will be computed as, NAS = NFS*NSS*NTS 2.4.4.1 Guna Trader (GT) In the event of an AI trade execution, the Guna Trader, trading component of the Guna AI system will place buy and sell orders on the exchange on which the particular asset trades.
All components of the Artificial Intelligent System will be trained by the Guna Fund ’s team of blockchain asset analysts.
3. Token Metrics In order to invest in the Guna Fund, one needs to purchase our fund’s utility token, GNA. Summary of the Token: a) b) c) d)
Token Type: Private Blockchain Token Maximum Supply: 50,000,000 GNA Available for Purchase: 50,000,000 GNA Minimum investment: NA
3.1 Revenue Model and Fund Utilization
A total of 50,000,000 GNA tokens will ever be in existence. At the time of the private sale launch, 1 GNA will be worth $1. An entry load of 2% will be levied , i.e., on an investment of 1 BTC being made, the contribution to the fund is 0.98 ETH. Profits made by the fund will be commissioned on the last working day of a calendar month at a rate of 20%. However, if the NAV of the fund is below the watermark (below the NAV at the time of last commission realization) of the previous month, commission will not be charged for that period.
Out of the total revenue collected by the Guna Fund, its utilization will be as follows: a) b) c) d) e) f)
15% for further research and development 10% for marketing and PR 5% for legal consultation 5% for advisors 15% for partners 50% for the core team
3.2 Token Offering
The Token Offering will begin from February 27, 2018 when the fund goes live. Individuals can send their investments to the fund in the form of Bitcoin (BTC), Ether (ETH) or Litecoin (LTC). The deposit addresses for each of the above mentioned cryptocurrencies are as follows: BTC: 1Ni2GnyTbt8yFGqRa3RgVYvH33aMeiDPuX ETH: 0xa72af427aaeb9c654559734e7b030a3d799a4d1d LTC: LhAPG6KwpnjZKfy7UTeUAvAd1aZPSdLuNw 3.3 Token Distribution and Exchange
Business Whitepaper
Page |6
On deposit of their funds in the form of BTC/ETH/LTC, the investors will receive equivalent amount of GNA tokens in their account wallet on the website within 24 hours. Funds need not be sent from ERC20 compatible wallets. GNA token holders can redeem their GNA tokens by exchanging them for BTC/ETH/LTC at the then current rate on the Guna Fund website. 3.4 Price Metrics of the Token
The fund will build and maintain its portfolio of blockchain assets and apply various techniques of investment and trading to maximize its Net Asset Value (NAV). From March 5 th 2018, the fund will go live and the price of the GNA token will be pegged to the NAV of the fund . The price of a GNA token will be the live value of the Guna Fund’s NAV. It will be computed by the following formula: =
Where, GNA = Price of the GNA Token (in a certain currency C) AUM = Total Assets under Management TS = Total number of tokens sold TS = MCS - ∑ =1
i
Where, MCS = Maximum Circulating Supply xi = Number of GNA tokens in the master GNA wallet of the i-th exchange it is listed in The above formula ensures that the price of the token is not affected by the buying and selling of it.
So, if the fund makes a profit of 30% from March 5 th th 2018 to April 30 th th 2018, the price of 1 GNA on April 30 th th 2018 will 2018 will be $1.30 . Likewise, if the fund incurs a loss of 15%, the price of 1 GNA will be $0.85. 3.5 Commission on Profits
As mentioned in the section on revenue model, the fund will charge 20% commission on profits on the last day of a calendar month. The commissions will be realized by means of airlift of GNA tokens from the investor’s wallet to the Guna Team’s GNA master wallet. Illustration 1:
Paul invests 100 GNA worth of BTC/ETH/LTC on March 5, 2018 at a GNA price of $1. He further invests another 50 GNA worth of cryptocurrencies on March 10, 2018 when each GNA token was trading at $1.08. He decides to withdraw 20 GNA worth of fund on March 17, 2018. Since at this time each GNA is worth higher than at the time of his purchase, he will be charged a commission of 20% of profits realized on his withdrawal, i.e., 0.35 GNA will be airlifted from his wallet. He further invests 30 GNA worth of capital on March 25, 2018 and the Guna Fund team realizes commissions on profit made till March 31, 2018. Date
GNA Bought
GNA Sold
GNA Price
GNA Airlifted
GNA Balance
March 5 th 2018
100
$1
100
March 10th 2018
50
$1.08
150
Business Whitepaper
Page |7
March 17th 2018
20
March 25th 2018
$1.12
30
0.35
129.65
$1.10
April 1st 2018
159.65
$1.15
3.43
156.22
Profit Commission Realized on March 17 th 2018: 20
1.12 − 1
∗ [0.2 ∗ {100 ∗ (
1.12 − 1.08
) + 30 ∗ (
150
)}]
1
1.08
Profit Commission Realized on March 31 st 2018: 100
1.15 − 1
150) ∗ 1 2 9 . 65 ∗ (
0.2∗[{(
50
1
)} + {(
1.15 − 1.08
150) ∗ 1 2 9 . 65 ∗ (
1.08
1.15 − 1.10
1. 10
)} + {30∗(
)}]
Illustration 2:
Paul invests 100 GNA worth of BTC/ETH/LTC on March 5, 2018 at a GNA price of $1. He further invests another 50 GNA worth of cryptocurrencies on March 10, 2018 when each GNA token was trading at $1.08. He decides to withdraw 20 GNA worth of fund on March 17, 2018. Since at this time each GNA is worth higher than at the time of his purchase, he will be charged a commission of 20% of profits realized on his withdrawal, i.e., 0.35 GNA will be airlifted from his wallet. He further invests 30 GNA worth of capital on March 25, 2018 and the Guna Fund team realizes commissions on profit made till March 31, 2018. Date
GNA Bought
GNA Sold
April 5th 2018 April 18th 2018
40
April 23rd 2018
35
April 30th 2018
GNA Price
GNA Airlifted
GNA Balance
$1.15
156.20
$1.14
206.20
$1.13
0.073
171.12
$1.10
171.12
Profit Commission Realized on April 23rd 2018: 35
1.13 − 1.12
∗ [0.2 ∗ {100 ∗ ( 206.22
1.13 − 1.1.2
) + 50 ∗ ( 1.12
1.13 − 1.10
) + 30 ∗ ( 1.12
)}] = 0.073 1.10
Profit Commission Realized on April 30th 2018: 0 (Since the GNA price of $1.10 on April 30th 2018 is lower than that of $1.15 at the time of the last commission booking on March 31st 2018) Illustration 3:
Paul carries forward the balance of 171.12 GNA into the month of May 2018. He invests another 60 GNA worth of capital into the fund on May 14, 2018 with each GNA being priced at $1.19. He decides to withdraw 25 GNA worth of capital from the fund on May 21, 2018 at which time each GNA token is worth $1.23. The commission charged by the Guna Fund team at that moment will be levied on the investments made at a time when each GNA was priced lower than $1.23. This comes to be 0.46 GNA which gets airlifted from his wallet.
Business Whitepaper
Page |8
After investing another 15 GNA worth of capital on May 28, 2018, he gets charged the monthly commission on profits made during the period between the previous realization of monthly commission on March 31, 2018 to the last day of May 2018 which comes to be 3.96 GNA, leaving him with 216.70 GNA to carry into June. GNA Bought
Date
GNA Sold
May 1st 2018 60
May 14th 2018
May 21st 2018
25
May 31st 2018
GNA Balance
$1.10
171.12
$1.19
231.12
$1.23
15
May 28th 2018
GNA Airlifted
GNA Price
0.46
205.66
$1.24
220.66
$1.25
216.70
Profit Commission Realized on May 21st 2018: 25
1.23 − 1.13
231.12
∗ [ 0 . 2 ∗ { 1 00 ∗ (
1.23 − 1.13
1.13
)+ 50∗(
1.13
1.23− 1.19
+ 60 ∗ (
1.23 − 1.13
)+ 3 0 ∗ (
1.23 − 1.14
1.13
)+ 40∗(
1.14
)
)}] = 0.46
1.19
Profit Commission Realized on May 31th 2018: 100
1.25 − 1.13
231.12)∗205.66∗(
0.2∗[{(
50
1.13
) ∗ 2 0 5 . 66 ∗ (
1.25− 1.24
∗ (
1.13
) ∗ 2 0 5 . 66 ∗ (
1.25− 1.13
231.12
+ {(
231.12
)} + {(
30
1.25 − 1.13
60
1.13
)} + {(
)}
1.25− 1.19
231.12) ∗ 2 0 5 . 66 ∗ (
1.19
)} + 15
)] = 3.96
124
3.6 Revenue Sharing
Service to our business partners (detailed in the following section under Partnerships ) will be monetized and 5% the revenue generated will be shared with the investors through airdrops. This is intended to be a gesture of gratitude to the community of investors for contributing towards a large corpus of which a portion will be utilized for conducting R&D. 3.7 Voting Rights
The GNA token will allow its holder the right to vote in all major decisions of the fund. This privilege will come to use when we would need to venture in directions that affect the investors of the Guna Fund. Based on feedback from the investors, polls would be conducted on the fund’s website. Examples of voting would be to decide whether to shift our ROI base from fiat to ETH or BTC take advantage of arbitrage positions nuclearize the funds based on their categories 3.8 Referral System Investors are entitled to receive referral bonus of 5% of the first investment made by their referral. This bonus will be borne by the fund management team from the commission on profits made by the referred investor.
Illustration:
Business Whitepaper
Page |9
Paul refers Anne who invests $100. After a month, Anne’s investment becomes $130. Commission made by the Guna team = 0.20*$30 = $6 out of which $5 will be credited to Paul in the form of GNA tokens.
4. Transparency and Reporting Transparency in the Guna Fund’s trading activities and current holdings will be ensured through the read only exchange APIs as well as through proof of reserve for non-exchange wallets (this includes both hot and cold wallets). The insights section of fund’s website will show the following live data: a) b) c) d)
Portfolio of the fund Price of the GNA token Value of Key Performance Indicators (KPIs) of the fund Historical data of i. GNA token price vis-à-vis fiat and major cryptocurrencies ii. Funds under management
A summary of the trades made and positions taken during a 30 day period, along with the strategies and rationale behind the trades, will be published on the last day of every calendar month via a b log post.
5. Security Not more than 25% of the fund will be held on exchanges for active trading. These assets used for active trading will be distributed over multiple accounts on exchanges in a manner such that not more than 5% of the actively traded fund resides in a single account. This would be done to reduce the impact of an account ’s compromise on the fund, if and when it happens. The assets meant to be held for mid to long term positions will be stored in hardware wallets like Ledger Nano S and Trezor. Assets not supported by these wallets will be stored in encrypted USB drives. These storage devices will be held in secure safe deposit boxes with backup paper-wallets stored at separate secure locations. Locations are undisclosed for security consideration. A public audit will confirm the presence and safe storage of the USB drives and paper-wallets to the investors of the fund.
6. Partnerships 6.1 Investment and Trading Firms
Blockchain projects are growing on traditional investment firms as an upcoming asset class. The Guna team intends to handle the outsourced blockchain assets investment activities of such large firms or provide them its expertise and research. 6.2 Other Industries
In addition to forecasting prices of blockchain assets, the models developed by research and utilization of hybrid intelligence will see extended utility in other domains such as weather forecasting, retail, healthcare, gambling and sports betting, agriculture and others. The team behind Guna aims to offer solutions and hybrid intelligent consulting to leading firms from these domains and monetize such offerings.
7. Roadmap •
•
• • • •
May 2017 – Fund Fund launched as a mutual fund of digital assets by the CIO in private September 2017 – Use Use of margin trading/hedge fund strategies adopted by the fund’s management March 2018 – Fund Fund opened to public April 2018 – Commencement Commencement of development of the Sentiment Analyzer components May 2018 – Commencement Commencement of development of algorithmic trade bots June 2018 – Migration Migration of investors’ funds to the private blockchain
Business Whitepaper
•
P a g e | 10
January 2018 – Diversification Diversification into other asset classes like equities, forex, commodities and other entities of alternate prediction markets
8. Team •
Ayan Dasgupta, Co-founder and Head of Execution and Investment Decisions
•
Nilesh Chakraborty , Artificial Intelligence and Deep Learning Expert
9. Legal Considerations Considerations To be filled.
10. Conclusion 11. References