Alt Data Enables Crypto Arbitrage
Swiss hedge fund Cryptonomics Capital has spent the past 15 months backtesting its data-driven cryptocurrency trading strategy, which it has seasoned with a healthy dose of alternative data.
“The cryptocurrency market is still new enough so that you can apply methodologies and topics that do not work in the traditional financial markets since you’ll need much lower latency and much more power,” Tom Debus, a founder and managing partner at Cryptonomics, told Markets Media.
Although the global crypto market remains highly unregulated, the fund follows the standard regulatory policies of Switzerland and Lichtenstein, according to Debus.
“We stick to crypto-to-crypto transactions only,” he said. “We do not trade between cryptocurrency and fiat currency, which allows everything to be clear-cut.”
The fund arbitrages one of four cryptocurrencies— bitcoin, ethereum, bitcoin cash, and litecoin— against the remaining three currencies using a conservative stop-loss strategy fueled by market data as well as news and social media sentiment.
“You have so much volatility in that asset class itself that you do not need to apply too many freaky experiments on the algorithms,” said Debus.
Cryptonomics algorithms track between 17 and 21 indicators, which include the fee structure of the three current exchanges on which the fund trades.
“Once they all fall into the right sequence and boundaries, we open a position,” he said.
The five-person fund initially planned to roll its algorithms by hand but chose to roll out AlgoTrader’s end-to-end algorithm development and execution platform.
The process starts with manual fundamental analysis, according to Debus.
“Then we commit to AlgoTrader for all of the technical strategies and use our own AI stack and data-science platform for doing sentiment analysis on news, social media, and other things,” he added. “That is an additional five or six signals, which we only include with the growing AUM that we are expecting over the next few months.”
AlgoTrader introduced cryptocurrency support in mid-2017, noted Richard Chmiel, COO at AlgoTrader. “Cryptonomics is one of about 25 firms using the platform for trading cryptocurrency.”
Cryptonomics also is investigating trading on three additional exchanges, which would entail tweaking its algorithms.
“An exchange’s fee structure is one of those 17 indicators that we are looking at regarding how they might lower a trade’s profit,” Debus explained. “That is part of the strategy.”
As long as cryptocurrency investors do not get too excited and start using to much leverage in their trading, he expects the current arbitrage-friendly environment to last for the next 18 to 24 months.
“We obviously will use that lead time to exploit different venues and strategies, especially with the news sentiment analysis,” he added.
Tailoring equity-trading strategies to client specifications is a work in progress.
Cantor Fitzgerald cites execution quality, customization, and access to liquidity as trading differentiators.
Market 'electronification' sets the stage for auto-ex, MTS says.
Growing use in developing regions and other asset classes.
On the sell side, ‘Execution as a Service’ is in, off-the-rack algos are out.