This is the first episode in a series of Blog Posts about a suite of Technologies developed as a foundation layer for the new era of Decentralized Financial services. Today I will introduce the first ever self-auditing Autonomous Trading System: AlphaNexT.
AlphaNext is a quantitative Autonomous Trading System (“ATS”). It uses statistical methods to analyze the Bitcoin markets and make measured trades with controlled risk. Its performance is publicly auditable in real-time here.
A new protocol named Proof-of-Return-on-Investment (“PoROI”) was developed to create immutable and independently verifiable proofs of its performance. The PoROI system is Blockchain agnostic and currently using the Ethereum Blockchain to store the proves. To allow for proofs on high-frequency trading systems PoROI will also be implemented on the IOTA Tangle as soon a scalable permanode solution is released (more on IOTA in the next post).
There are currently no low-cost solutions to prove the performance of trading systems or funds. Currently, audit companies independently verify balance sheets and track record for hedge funds and large asset manager. However, the cost for these audits is high and is only performed periodically. Not only are they outside of the means of most smaller funds, trading systems or traders, but such audit reports rely on the integrity of the underlying data provided to the audit firm and are therefore prone to manipulation and errors.
In the age of thousands of Telegram, Discord an other pay-to-join channels from people claiming to have a strategies to beat the market, a lack of an easy and cheap way to prove the performance of such strategies leads to a lot of scams and fraud. You will surely remember the infamous BitConnect pyramid scheme of 2017, which claimed that the profits generated by a “trading bot” would allow the company to pay up to 1% daily interest to anyone locking up their Bitcoin on the platform. Nearly everyone lost their money.
The Ethereum blockchain or IOTA Tangle, amongst other DLT implementations, can be used to store data in an immutable and publicly verifiable way. We can leverage this to store an audit trail for the performance of traders or trading systems.
The public availability of this audit trail also poses a challenge: trading data is very valuable. When generating the audit trail the Intellectual Property of the strategy and the value of the Signals must be preserved and protected, meaning that we can not just publish every Entry or Exit Signals cannot simply be published on the chain as is.
The main design goals of PoROI are to create proofs that:
- are time-stamped on when the Entry and Exit Signals are generated;
- protect the information in the Signal until it loses its value;
- do not need any information about the underlying trading strategy;
- are as cheap as possible to create and free to validate;
- are Immutable;
- are machine readable.
These are the goals that guide the design of the process of generating proof with the PoROI protocol. In this initial version, we take as underlying assumption that the Strategy in question is Long-Only with a stop-loss on each position (AlphaNexT fits this description). The process to create a proof is as follows:
- Upon generation of the Signal, a random Key (or Salt) is created, this is required to be of reasonable length to avoid brute-force attacks.
- The Signal and the Key are hashed together using a secure hash function; the current implementation of PoROI uses SHA-256.
- The resulting hash message is stored on a DLT by sending a transaction to a dedicated address. All transactions need to be sent to the same address or be linked in some other way (to avoid someone later only showing the good ones).
- At random intervals “noise” Signals are sent using the same process, this is done to avoid someone listening to the transaction stream and guessing its content based on previous information.
- After a “reasonable” amount of time, once the signal has lost its value, the Signal and Salt can be published in clear-text (without being hashed), together with an identifier of the transaction containing the previously committed hash.
The Proofs can be easily submitted for verification by sharing the address to which the transactions are sent. The verifier simply look into all transactions on the address and check that the result of hashing the Signal with the Key corresponds to the same hash published at the time the Signal was created.
While the AlphaNexT strategy itself was developed and built with the full intention of being profitable (and as of now it is), it has acted as the testing ground for many of the Ideas Presented in this and the next blog posts. You can see its functionality by visiting our website website (www.alphanext.ch). Here a brief explanation of the current constraints linked to generating the Proofs.
First of all, AlphaNexT executes very simple trades, it can do only 3 things: Buy, Sell and set a Stop-Loss. More complex trading system require such things as laddered orders, conditional orders, arbitrage transactions, liquidation levels, trailing stop-loss and take-profits. While these things are not in any way impossible to create Proofs for, they require some more thought, especially when a constant, near real-time, update of the information is necessary. This brings us to the next aspect of AlphaNexT: it trades the 30min candles, and generates around 10 signals (5 trades) a week. This has the benefit that the amount of transactions that need to be sent over Ethereum is relatively small. If a strategy requires Signals to be sent very few minutes (to update a trailing stop-loss for example) this would greatly increase the network fees required.
One solution could be to move the Proofs from the Ethereum Blockchain to the IOTA Tangle, which does not have any transaction fees. The concept of a “digital twin”, which is pretty much what PoROI creates, is well understood in the IOTA Community and there are a lot of libraries and tools such as “Masked-Authenticated-Messaging” and “Chronicles” being developed to support the creation and validation of such digital counterparts.
The ability to create cheap, immutable, real-time and machine-readable Proofs of the performance of an Autonomous Trading System provides great added value in terms of increasing trust in funds, trading systems and individual traders.
While the currently available DLT solutions are not fully satisfactory when it comes to scaling and fees, there are strong projects such as IOTA becoming production ready over the next year.
Big thanks goes to Matthieu Gueissaz for proofreading the blogpost and becoming the first beta tester!
Source: Crypto New Media