Bitcoin and crypto ETFs explained
Can algo trading beat the market?
An examination of one crypto quant fund, their algorithmic trading strategies and how they seek to outperform the markets.
With digital assets reaching historic highs in recent months — driven, in part, by the strong performance of Bitcoin — it is unsurprising that new crypto funds are emerging at a rapid rate.
While many of these new entrants have adopted a discretionary trading strategy, research conducted by PwC indicates the quantitative strategies have outperformed other approaches in recent years — even in periods when market conditions were less favorable.
As a relatively immature and dynamic market with highly fragmented liquidity, digital assets pose traders with unique opportunities to generate alpha — as well as some unique infrastructural challenges. This is particularly true of quantitative and high-frequency trading, which rely on having automated, fast and reliable connectivity to a wide array of liquidity providers and sources of market data.
But what do crypto quant funds actually do? What strategies have they adopted to outperform the market?
In order to gain a better understanding of what’s happening on the ground, I spoke with the algorithmic trading team at CRYFIN — a Czech Republic-based asset management company that practices quantitative trading of digital assets. The firm has used AlgoTrader’s digital asset trading platform as a hub to coordinate its systematic trading strategies. The team gave insights into some of the approaches CRYFIN is using, and the infrastructure required to execute them.
What is algo trading?
Algorithmic trading describes the automation of trade executions based on programmatic trade signals generated through a quantitative strategy.
The trades are automatically executed on secondary markets — such as cryptocurrency exchanges — through the use of execution algorithms, which split up the parent order and place child orders on an exchange’s order book.
This practice is based on predefined logic and parameterization relating to timing, quantity, price and volume.
Crypto arbitrage is maturing
Crypto arbitrage is an area where the comparative immaturity of the digital asset market can be viewed as an opportunity. As mentioned above, there are hundreds of crypto exchanges around the globe — a fact that makes crypto liquidity highly fragmented. This state of affairs creates frequent opportunities for cross-exchange crypto arbitrage, as the price of the same assets traded on different exchanges tends to deviate from its mean. The best arbitrage opportunities may only last for a few milliseconds, but some can last minutes.
The recent price growth of BTC and its ever-increasing average daily range make the case for arbitrage trading even more compelling. In fact, when volatility spikes in the market, the price difference of BTC across exchanges can equate to hundreds — or sometimes even thousands — of U.S. dollars.
In order to take advantage of these arbitrage opportunities, you need to be well-capitalized and have a bullet-proof execution platform that allows you to quickly react to price deviations in a matter of microseconds. As you need to scan and respond to the markets 24/7, having fast and reliable access to a wide range of liquidity venues is critical.
Whether you are setting up a proprietary system or choosing a third-party digital asset trading platform, these are the main prerequisites you should keep in mind if you want to engage in arbitrage across crypto markets:
- Reliable connectivity to the largest and most-liquid exchanges — such as OKEx, Kraken, Binance, Gemini, Bitfinex, etc. — over a single connection.
- Ready-made order book data structures and predefined events that allow you to react to order book supply and demand in real time, enabling you to use individual or aggregated order books to find the best bid or ask at any time.
- Tools to monitor the latencies between exchanges in responding to market data and events.
- A wide range of execution algorithms and the ability to create customized alternatives so that you have the flexibility needed to minimize fees and slippage.
- Simulated execution engines to run algorithms on live out-of-sample data to test-drive it with different models of transaction costs.
Since arbitrage falls into the category of high-frequency trading strategies, it is quite a challenge to backtest it properly. Thus, it also makes sense to establish a way to collect and store order book data for future use when backtesting. This is a way to kill two birds with one stone, as you won’t have to buy this data again from third parties. Indeed, this is particularly advisable given the fact that good-quality order-book-level data can be quite expensive in this market.
Confluence trading through a wide range of analysis
Unlike with other asset classes, CRYFIN points out that many crypto traders still only employ technical analysis in their trading strategies — without considering fundamentals or sentiment. The company has adopted an approach — which it calls confluence trading — that aims to marry technical analysis, sentiment analysis and on-chain transaction analysis in a single strategy, and in a fully automated fashion.
In the case of sentiment analysis, a proprietary platform is used to analyze tens of thousands of data sources in real time — ranging from social media to reputable news sites and blogs. The main goal here is to detect and respond to situations where it would be unwise to remain in a long or short position on a given asset.
Another important type of analysis — which is specific to crypto assets — is blockchain transaction analysis. Although growing rapidly, the market capitalization of crypto assets is still small compared to other asset types, like stocks or FX markets. This means that large orders can move the market substantially, and you need to be able to react quickly.
CRYFIN uses a tool called WhaleTrace to detect and respond to larger on-chain transactions to or from exchanges that could potentially indicate upcoming volatility. The firm noted that, since implementing confluence trading, it has observed significant gains in traditional performance statistics — like average trade, net profit, maximum drawdown and the Sharpe ratio.
Hybrid trading through “quantamentals”
As you grow as a trader and manage more assets, you want to eliminate emotion and to systematize as much as possible to avoid human error. Practically, this could mean automating the scanning of markets to minimize your time in front of the screen, or automating trade and risk management.
A degree of automation makes just as much sense for discretionary traders — who trade based on the available information at any given time — as it does for quants. Indeed, to some extent, the differences between discretionary and quantitative trading are beginning to blur as discretionary traders become more systematic about incorporating data into their workflow. In this vein, FT reported last year that a hybrid “quantamental” mindset is one of the biggest trends in asset management, generally.
For digital asset traders, partial automation could involve using pre-built execution algorithms and smart order routing to achieve best-price execution across multiple liquidity venues.
Trade and risk management is another area ripe for automation in this regard. Whenever a discretionary trader at CRYFIN opens a position, the trading platform’s “autopilot” takes control to manage execution, removing the possibility of emotional human decisions or knee-jerk reactions further down the line. Additionally, a system has been implemented to automate position sizing and risk management for discretionary trades — based on predefined rules and current market volatility.
As the crypto market gains new institutional entrants and the market infrastructure matures, traders will need to adopt more sophisticated techniques to beat the market. Research suggests that quantitative strategies outperform the alternatives — but a flexible and robust digital asset trading platform is a key prerequisite for this approach.
This article was written by Jason Blum, head of strategic partnerships and business development at AlgoTrader AG.