Algorithmic trading (also called algo trading, automated trading, or quantitative trading) is an approach to stock trading that relies on automated preprogrammed instructions (algorithms) to carry out a strategy envisioned by the trader. If you manage to successfully identify the upcoming trends by analyzing the existing data (or, rather, manage to write a set of instructions that identifies them), you can, for example, buy a stock when it shows a tendency to rise and sell it before it starts to drop.
Trying one’s hand at algorithmic trading can be a daunting prospect because it requires significant expertise in two unrelated fields: programming and market mechanics. Although there isn’t plenty of information for beginners, there are some manuals and instructions, and some of them are quite good and can lead you through the technical aspects of creating your first strategy. In this article, however, we will talk about what you have to take into account before you even consider entering this field.
Advantages of algorithmic trading
It is often said that success in trading is comprised of 30 percent market analysis, 30 percent emotion control, 30 percent risk management, and 10 percent luck. Thus, a well-written algorithm can take care of 90 percent of the job, leaving only luck unattended.
1. It eliminates the emotional aspect of trading
The most spectacular trading failures happen because traders let their emotions get the better of them. They panic and start to sell at the worst possible moment, or get too greedy and decide for the already good stock price to grow a little more. If you allow an algorithm to make decisions for you and don’t interfere, you don’t have to fear such problems.
2. It is fast
An algorithm reacts to the changes faster than any human ever can, even if he or she spends 24 hours trading.
3. It works independently from you
After you set up your algorithm, it can work without your interference. If you have a demanding day job and cannot afford to dedicate a lot of time to trading, it is your only way to act meaningfully in this field.
Disadvantages of algorithmic trading
1. It requires discipline
What we said about its usefulness for removing the emotional aspect from trading is only true if you can entirely rely on your algorithm and abstain from interfering with it. This can be extremely difficult, especially during the extended drawdown. Quite often, strategies that prove to be highly effective in backtesting are ruined through human interference.
2. It requires continuous research
No strategy remains profitable forever. If you want to maintain a profitable portfolio, you have to dedicate a significant amount of your time to research the field, perfecting your code, and improving your strategies.
3. It requires a solid programming background
Of course, you can use ready-made tools and only have a rudimentary knowledge of programming while outsourcing the majority of development. However, this way, you won’t understand most of your trading infrastructure and won’t be able to make meaningful changes on your own. Therefore, to be successful in this field, you have to have a solid knowledge of a programming language such as C++, Python, R or Java.
4. It is not a get-rich-quick scheme
When done right, algorithmic trading can bring you consistent and stable long-term income. The success will not come overnight — it requires time, steady research, disciplined approach, and patience.
Things to consider
Experimenting with your own money is not the only way of earning money through algorithmic trading. If you have sufficient programming skills and understanding of market mechanics, you can try your hand with hedge funds like Quantopian. They provide platforms where programmers compete against each other for commissions by offering their code and testing it to see whose is the most profitable. It may be a good way of checking if your expertise is enough to try your hand at independent algorithmic trading without risking your own money — or it can become a source of income in its own right.
Artificial intelligence and deep learning
Breakthroughs in the fields of AI and machine learning allow for the creation of algorithms that are able to perfect themselves through the iterative application as a part of deep learning. While it is unlikely that you can benefit from it individually, this too can prove to be a lucrative field of work for a specialist in this area.
Understanding your strategy and avoiding overfitting
Algo trading is based on creating strategies and backtesting them — using historical data to see how the strategy would have performed if it were applied over a past period. One of the common mistakes even more or less experienced traders make is over-fitting — that is, they create a strategy that is over optimized for the existing subset of data and is based on random trends that exist in this subset but are absent in the statistical population. Usually, it is caused by tweaking parameters to show the best possible results over the past period – as a result, you get a meaningless strategy that is based on nothing but the existing knowledge of how things developed, not on real trends.
A good way to avoid it is to ask yourself whether you understand your strategy. Can you explain why it should work without introducing a string of caveats and listing dozens of parameters? Is it grounded in behavioral factors or fund characteristics that cause the patterns you intend to exploit? Or is it just that it would have worked if it were applied in the past?
Algorithmic trading: Where to start
The best way to get into algo trading is to use simple publicly available strategies, backtest them, make your own tweaks, backtest again and see if you are able to improve on them. It will probably not work the first time around but will allow you to get the necessary experience.
Have you already tried your hand at algorithmic trading? What can you suggest to those who are just starting out?
Disclaimer: Nothing in this article should be seen as constituting investment advice, and not all investment strategies are suitable for each individual.
Featured image: Shutterstock