What's the point of algo trading - and what about crypto?
Trading crypto has its particular appeals - but why trade using an automated algorithm?
An algorithm is “a process or set of rules to be followed in … problem-solving” (according to Google) and as such, is ripe for automation via software. When it comes to the trading of financial assets, coding algorithmic software has been a logical step from the quantitative analysis of market behaviour: once you have modelled a market, algorithms can be written based on the model. For instance, since assets can be modelled as following random walks (like particles in solids), the model can give a prediction of volatility which can suggest a reasonable spread around current mid-market prices to place buys and sells. An algorithm can be coded to place buy/sell orders and cancel them as the price moves.
Of course the further advantage of trading using an algorithm running on a machine - is that it is a machine. It doesn’t get tired an lose focus and it can trade 24/7. For anyone who trades and can code, writing a program that implements a trading algorithm (often called an algo) is therefore at least interesting.
Those who have significant experience of trading, either professionally or on their own retail account, can attest to the difficulty of predicting market change (without special knowledge). This is not surprizing since once a market is predictable, money will pile in and take all profit available through that prediction: the market will change so that it is no longer predictable. Nevertheless, traders can make money - and being a better trader leads to greater profit.
It’s a reasonable thing to buy or sell a financial asset if you know the price will change. How might you know that? A company’s results are worse than predicted. There’s military action in an oil-producing nation - etc: these are the fundamentals behind prices. Algos can be fed indicative information so that they adjust their trading based on it.
However, even without fundamental changes, asset prices fluctuate - sometimes wildly: why? There are many reasons but sometimes it’s just a trend that emerges from the market. Technical analysts look for these trends and try to predict their direction and extent. They can write algos that encode their predictions.
So, armed with a means of exploiting fundamental knowledge or a model of market trends, how does one get going? It can be tricky: to trade stocks you need a broker - they may not provide an API to trade against. If they do, they may charge fees, delay prices, throttle order rate or otherwise stymie the effort to trade efficiently. It’s new to them and they don’t see retail algo traders as an important part of their business.
Cryptocurrencies were invented by programmers who had a libertarian world view. The creators of Bitcoin wanted to exchange value easily without the dependence on an authority like a central bank. As furthur coins were developed and the need for exchanges became clear, the people setting them up were from Silicon Valley, not Wall Street. Cryptocurrency investors also were of the same ilk - technologists.
It was automatic to open these exchanges freely to all traders using well constructed APIs. It was immediately assumed there would be algo trading.
The fundamentals of crypto are strong. Many believe they will replace fiat currencies. This basic belief along with progressively greater publicity and slowly increasing institutional interest has led to a rapid rise in total crypto market valuation. So there is a case to buy crypto.
There is no question however that it is extremely volatile. Coins often change in value typically by multiple percentage points each day.
These facts - general upward price trend and high volatility (and 24/7 trading) lend crypto to algo trading. It’s not unreasonable to see crypto algo as a proving ground for algo. It may lead to an increased interest in algos by retail clients in other asset classes.