The Enigma Catalyst white paper describes itself as an off-chain decentralized exchange protocol coupled with the first investment platform focused on enabling developers to build, test and execute micro crypto-funds.

Developed by the Enigma team at MIT, Catalyst aims to be the "glue" between cryptocurrency exchanges, while also extending open source algorithmic trading tools like Quantopian and Zipline for use in the cryptocurrencies ecosystem. These sorts of platforms and modules allow users to select from a leaderboard of trading algorithms and import already curated and formatted data, without having to do all the coding for the statistics of the strategy.

AI/ DON'T USE THIS ONE Newsweek is hosting an AI and Data Science in Capital Markets conference on December 6-7 in New York. Photo: NewsweekMediaGroup

Catalyst sees a gap for algorithmic trading on crypto-assets without a custodian. There's lots happening in the trading space, with the likes of ICONOMI offering a centralized cryptoinvestment platform; or Prism, backed by Shapeshift, which operates using a semi-centralized model. Meanwhile, decentralized on-chain investment solutions such as 0x, Melonport and Bancor may face a performance issue, and currently only support Ethereum ERC20 compatible tokens.

Catalyst proposes to circumvent the on-chain performance issue by routing orders through off-chain payment channels and using hashed timelock contracts (HTLCs), the same vein as the Lightning or proposed Raiden networks.

This design allows users to make fast, cross-chain transfers while maintaining full custody of their assets. Order books are maintained by a permission less network of liquidity providers, each of which spans multiple, individual payment networks. To begin trading, users open payment channels with a chosen liquidity provider, in the currencies they wish to trade. Orders are then submitted to the liquidity provider that a trader chooses, and matched with an online counterparty.

Guy Zyskind, co-founder of Enigma and Catalyst, said that when family and friends would ask him for tips on how to invest their money in crypto-markets, he realized the barriers to entry were very high. In order to do algorithmic trading today in crypto markets, as opposed to stock markets, you have to work as an engineer, curate data sets and build testing tools.


He said: "Catalyst is basically going to be a one-stop shop for developers and quants where they can, in just a few minutes, consume data that we in the community will provide around crypto markets. They can then use that to build trading strategies and test them in simulated way, so you don't risk any money. Then start actual live trading."

Any data-driven system is only as strong as the input data it receives, and Catalyst aims to provide the sort of big data required for modern artificial intelligence and machine learning tasks. The type of data suggested include historical time series market data, initial coin offering data and sentiment data from news, social media subreddits and crypto forums.

The platform will incentivize users to crawl and curate useful datasets that can be used by the community as it evolves. Incentivization would be directed by the market: those datasets that are accessed more frequently would result in more incentives than those that are not as popular.

Zyskind said: "I think one of the biggest things in the platform that we are building is a data market place. We came up with a few datasets that we think are going to be interesting that we are going to curate – one of them was blockchain data, so all the metrics for Ethereum, Bitcoin and all the others. Developers will start curating datasets and putting them on our platform for others to use and then the sky is the limit."

In addition, a web-based leaderboard ranking of all strategies deployed by developers will include standard return and risk metrics, such as ROI, Sharpe ratio, alpha and beta and max drawdown. To prevent short-lived strategies from being over-represented, Catalyst will favor strategies that have been robust in different market conditions and have built a longer track record over time.

"The leaderboard will make it easy for people to invest in the highest performing strategies. The quants behind these algorithms will be able to set the management and performance fees themselves, creating a truly open-market."