Do you know that when there were no computers and no internet, how buying and selling of stocks were done? Well, there used to be a ‘ring’ in stock exchanges where buyers and sellers (their brokers, to be precise) meet and carry out trading transactions. Traders or brokers used to shout like crazy, just to make their desired transactions. Stock markets were no better than fish markets. But the rise of computers and the internet changed the game of trading completely. Anyone can now buy or sell stocks sitting at his/her home just with the click of a mouse button.
The Era of Algorithmic Trading
Then came the era of algorithmic and automated trading. We essentially programmed computers to make trades on their own whenever certain criteria are fulfilled. You no longer have to sit glued in front of the screen, analyzing charts and crunching brain with numbers for identifying trading opportunities. Also, we as human beings are subject to a lot of emotions and biases, which certainly hurts trading. But computers don’t have this problem. They rigorously follow what you order them. Algorithmic or automated trading certainly looked like the future of trading. But one thing was still missing.
Also Read: Machine learning algorithms for Trading
The entry of Artificial Intelligence
Under automated trading what we do is that we essentially tell the computer to do X whenever Y happens. For example, we develop a program that buys the stock (X) whenever RSI<20 (Y). Now the program keeps checking for happening of Y (RSI<20) and when it occurs, the program performs the action X (buy the stock). This is automated trading. Everything is done by computers. Humans only have to perform the programming work and specify the conditions X and Y. But maybe humans are so lazy creatures that they didn’t even want to do that work. Or maybe they are suffering from self-doubt and think that machines can perform this work better than them. So here comes the entry of Artificial Intelligence.
Basically, Artificial Intelligence is the pursuit of making computers ‘intelligent’ just like humans and enable them to think, learn and act on their own. By using artificial intelligence in trading, humans want the machines to figure out those sets of X and Y conditions for themselves that can generate maximum returns with minimum risk. There is no need for manual human intervention, expect to monitor and regulate the activities of artificial intelligence. Those days are not far when financial markets would be operated completely by artificial intelligence systems. In fact, AI is already playing a big role in shaping today’s financial markets.
Artificial Intelligence vs. Human Intelligence
AI systems can process such gigantic amounts of information in seconds that it is impossible for humans to compete with them. They are capable of analyzing information from different markets, historical data, financial reports, news announcements, sentiments on social media, economic factors, etc., to quickly decide what is the right trade to make at any point of time. They can make the correct evaluation of any stock by considering all the fundamental, technical, sentimental and other micro and macro factors in real-time. AI systems can scan almost infinite amounts of historical data in a few minutes and identify replicating patterns that are impossible to be observed by humans. There is a limit to what human intelligence can do at any point in time. But the potential of artificial intelligence is limitless.
It is often said that markets are driven by the sentiments of the masses. Human beings often try to estimate the sentiments of the general public, but it is nothing more than pure guesswork. However, the use of artificial intelligence in trading will pave the way for so-called ‘Sentiment analysis.’ By analyzing blogs, social media posts and comments, and news headlines, AI will be able to accurately gauge the sentiment of the masses in real-time. Also using sophisticated technologies such as machine learning and deep learning, AI computers can learn from their past success and failures, just like humans (even better, as computers are not prone to biases and ego problems) and keep evolving and improving their trading skills.
Pitfalls of artificial intelligence in trading
Artificial intelligence is going to dominate the markets. So, AI systems are going to act as money-making machines, right? Anyone with knowledge of markets and AI skills can pump out huge money from the markets, isn’t it? Well, there is no guarantee. It should be noted that there are always possibilities of errors and system failures. Even some small slips can result in big blunders and huge losses. These can occur not only at individual levels but also at a market-wide level. Flash crash of 2010, when Dow fall about 1000 point in a few minutes only to recover in the next few minutes is a good reference to understand how the follies of algorithms can cause blunders in the markets and wipe out trillions of dollars in minutes.
Besides the threat of these inefficiencies of algorithms, there is one more concern. What if they all would be too efficient that they would drive the market to that state of efficiency as mentioned in the Efficient Market Hypothesis (EMH). The efficient market hypothesis states that every stock trades in the market at the fair value and it is not possible to beat the markets in the long run. However, in reality, human beings are not that efficient as to consider every information available before evaluating any stock. Thus stocks do trade higher or lower than their fair value and traders and investors attempt to earn profits by identifying these mispricings. But since, artificial intelligence can factor all the information in evaluating stock prices in seconds, eventually, all stocks would begin trading at their fair value and there will be no legitimate opportunities left for trading or investing. The efficient market hypothesis will come true and no one will be able to beat the markets.
Also Read: Algorithmic trading as a career option
There shouldn’t be any doubt that artificial intelligence is going to replace human intelligence in trading. But no one knows for sure how this is going to turn out. No one knows whether it will improve the profitability of traders or make trading even more difficult and competitive than ever before. Only time can reveal this.