By Dr. Anna Becker, CEO and founder of EndoTech
The announcement by Fidelity that it will provide 401K asset management services for Bitcoin is just the latest sign of crypto becoming a mainstream asset. Now, the question for investors is how to approach this immature and extremely volatile market. An emerging option is for retail investors to approach cryptocurrencies the way that institutional investors, especially hedge funds, trade stocks and other assets—- using tools that leverage AI, machine learning and data.
AI-based algorithmic trading for stocks, futures and other traditional assets has emerged as one of the most consistently successful investment methods for institutional investors. Applying those principles to crypto trading could provide similar levels of success for traders – if it’s done right.
But simply porting algorithms for institutional stock investment to crypto won’t work. Having developed AI algorithms for asset classes ranging from equities to commodities to crypto, it’s clear that each asset class requires new computational considerations. While some of the fundamentals that go into stock investment algorithms are transferable to crypto investments – market sentiment, global events, regulatory decisions – there are many issues that apply to stocks that are irrelevant to crypto, including annual reports, production numbers, and a company’s physical asset value, to name just a few. And even with similar fundamentals, the weights given to them in crypto and stock investing will differ; for example, market sentiment is likely to be a bigger factor for crypto than for stocks.
While the basic algorithmic approach for stock and crypto investing will be similar, there are important differences – and nuances when it comes to using data and AI:
Pattern Identification: A good algorithm observes patterns in trading – based on hundreds or even thousands of factors – and analyzes them in order to determine the best move for a trade. One factor that skews that analysis in the stock market is high-frequency trading among institutional investors – with trades made on the basis of slight movements in prices, and the trades themselves often influencing prices. There is also never a repeating pattern for any significant amount of time when it comes to stocks because of the widespread use of automated trading features among institutional investors in these markets; the algorithms detect any patterns right away, use them to their advantage and thus the patterns cease to exist.
In crypto, however, most of the trading moving markets is (still) at the retail level – which means that it’s easier for AI-based algorithms to examine fundamentals, technical analysis patterns, and to suggest successful trades. The patterns are more obvious, clear and repeating; creating the optimal textbook conditions for building an algorithm.
Investor Strategy: Some investors like – and seek out – volatility. Others – such as those allocating a portion of their 401ks to Bitcoin – are likely to hold their investments longer-term, and probably prefer more stability. Currently, volatility levels in today’s bear market are extreme; with Bitcoin recently down more than 60% from its November high. Using AI, a crypto investment algorithm would suggest trades based on an investor’s overall goals – for example, offering those seeking fast, high level returns a certain amount of leverage and short-term systemic trading. Meanwhile, if wealth preservation with the goal of moderated returns is the preference, the algorithm would examine and analyze different factors, and provide trade recommendations that focus on more stable major cryptocurrencies, like Bitcoin or Ethereum, and take a longer-term approach with fewer trades, maybe going in and out of the market only 10 to 20 times a year.
Fundamentals: Factors relevant to crypto include data analysis, pattern identification, asset and portfolio management, and various levels of risk– including trade execution risks, exchange risks, systematic risks, and asset and liquidity risks. Liquidity risk is particularly important in this market; as unlike in the stock market, there is often not enough liquidity to exit at the stop loss, or set lower price limit, when a particular coin completely collapses. All of these need to be weighed and factored into the trading environment. With emerging asset classes, all of these dynamics vary depending on market condition: so factors like volatility, directionality, market depth, and trading volume all impact the complexity and permutations for consideration.
Each of these factors needs to be weighted based on the investor’s preferred strategy. And while most investors are unlikely to do a deep-dive on the algorithms powering their trades, it’s important – both for investors and for the assets – that these algorithms be as transparent as possible, and that investors are clear on what factors and fundamentals a trade is being recommended.
Automated Trading: It’s not enough for an algorithm to figure out a good trade; that trade needs to be executed, preferably immediately. Crypto assets can (and have) seen wild price swings – which is where automated trading comes in. On the other hand, markets can remain quiet, with minor price changes, for days or even months at a time. An investment system needs to know when to buy and sell in order to ensure that the overall goal is met. Price, volume, and threshold are all factors here – as are trading fees, where each transaction can significantly eat into profits (discount brokerage hasn’t reached the crypto market yet).
AI-based algorithms can parse data from thousands of sources that will indicate the optimum time for a trade – and follow through. Fidelity is far from the only big institution delving into crypto; it’s been coming on gradually for several years now, despite misgivings by some other big institutions (and names). Indeed, investing in crypto can be confusing, as the concepts and terms are new – and often esoteric. AI-based algorithmic trading, which takes into account the esoterica and analyzes it–and has proven itself in stock trading– can provide better returns, or at least smarter trading, in crypto as well. AI and data–when used right–can help investors better navigate the volatility, even in a market as new and challenging as crypto.