According to Merriam-Webster Dictionary, the word 'algorithm' means "a step-by-step procedure for solving a problem or accomplishing some end," and 'trade' is "the business of buying and selling or bartering commodities." When we put these things together, we get algorithmic trading: the process of executing trades via algorithmic strategies.
If we think about it, this definition tells us that trading of all commodities and stocks is inherently algorithmic even when we execute the trade ourselves or use an intermediary. Based on the information available, we all construct a conscious or unconscious mental model of the financial instruments we own and exchange. While some of us give in to occasional impulsive moves, most of us trade based on these mental models. The term 'algorithmic trading' we will investigate today refers explicitly to using trading strategies based on complex mathematical and statistical models implemented by computer code executed via online trading platforms.
After defining what we mean by algorithmic trading, we must distinguish it from 'high-frequency trading (HFT). HFT is the collection of executing trading strategies of commodities and stocks in very high volumes where execution strategies have latency budgets measured in milliseconds or less[1]. While HFT is available to big players only, algorithmic trading can be used by any small or large players alike.
Since trading financial instruments have become complex, with the introduction of complex models via algorithmic trading, trading firms started employing technically skilled professionals whose only job is to develop and execute trading strategies for these complex instruments. These qualified employers (called quants) usually have advanced degrees in mathematics, statistics, finance, and computer science[2]. Quants develop, implement, test, and execute these trading strategies using algorithmic trading platforms.
Big trading firms employ large numbers of quants to execute HFT and their algorithmic trading. But with the advent of cheap computing tools and open source algorithmic trading libraries, now even small day traders can implement their algorithmic trading strategies. The programming languages such as R and python have open libraries that can easily be installed and used for such purposes.[3]. While the learning curve might be a little steep for a non-technically inclined user, with a bit of effort and elbow-grease, every investor can now develop their algorithmic trading strategy and use it on the field to add to their trading portfolio.
Prof. Dr. Atabey Kaygun
[1]Interested readers might find the movie The Hummingbird Project attractive, partially based on Michael Lewis's book Flash Boys. The film is about two stockbrokers who lay a fiber-optic cable line from Kansas electronic exchange to the New York Stock Exchange to front-run other players.
[2] For detailed information on what quants do, the interested reader might check the following link.
[3] There are a plethora of python and R libraries available. The author finds the following libraries particularly useful: Zipline, FinTa, BackTrader, YFinance, TTR, Quantmod, and PerformanceAnalytics