This algorithmic article was written by Yutong Li – Analyst at I Know First, Master’s candidate at Brandeis University.

Highlights

• Despite the controversial opinion on whether Artificial Intelligence could beat the market, AI could still win in many aspects
• I Know First’s proprietary AI-powered algorithm can make accurate predictions and smartly approach the stock market uncertainty
• Based on the 1-year evaluation summary, I Know First has predicted ^IXIC, SPY, and QQQ with 100% accuracy in a 1-year time horizon, and has a performance over 62% for all time horizons

AI Impacts the Whole Financial Industry Today

With the rapid development of technology and computational power, Artificial Intelligence plays a crucial role in the trading equation and allows a more convenient and smarter way for investors to make decisions.  We could put together many examples of how AI is being applied efficiently nowadays across the financial industry. One thing that is being pervasively used is automated customer support. Much like other support that other chatbots could do, an AI financial chatbot could assist customers with simple financial transaction support. Customers could get a fast reply at any time in the day without having to wait for humans to help. On top of this, many companies are also using “Robo-Advisors” services to make long/short position recommendations to customers based on individual requirements and portfolio goals. Another popular application of AI is the predictive forecasting service that provides insights about the stock market trending. The algorithms could gather historical data to generate a predictive model from the past, then make predictions for future trading. Some more examples of how AI has been widely used are banking fraud detection, wealth management, optimization problems, etc. Thus, we could see so many tedious or challenging jobs that are taken over by artificial intelligence efficiently nowadays. And all these technologies and innovations are still seeking to go beyond human capabilities and help us to make decisions in the stock market easier.

Myths About Whether AI Could Beat the Stock Market

This extensive usage of artificial intelligence and the shifts toward machines, however, make many people start to question whether we could trust artificial intelligence completely and whether artificial intelligence could replace humans in the stock trading process ultimately. Even with so many successful applications of algo-trading so far, a controversial idea was put forward in this article, suggesting that AI will never beat the stock market. However, there are some counterfactors to this article that no longer hold over recent years with the expeditious evolution of machine learning and AI-powered algorithms. There is no doubt that our markets are driven by AI to a large extent and we think this proposition needs to be approached critically.

Randomness and Uncertainty Being Handled

The author of the above-mentioned article suggests that some future market moves are not predictable and there are uncertainties and randomness in the complex system that the machine is not able to tell. Admittedly, the market is never constant or stable, rather chaotic. The complexity comes from investors’ differences in goals, economic conditions changes, or “black swan” events. However, it is not true that the stock market is always chaotic, and it also contains predictable patterns. This is exactly the idea held by the I Know First company, and this concern of uncertainty is smartly handled by the company’s AI-based algorithms using Chaos Theory.

Can Artificial Intelligence Make Judgments?

The article also claims that the technical analysis of an asset only considers the historical data and tends to miss a judgment for future fundamentals, thus we still need to make informed decisions ourselves. It is true that the machine should not merely use technical analysis but should see the real market context as a whole. Nevertheless, our judgment for the future also relies on our experience and we humans are also using historical facts and previous knowledge to make decisions. We always have reasons to buy or sell a stock. Same here for artificial intelligence, and it could learn quicker from the past. With improving computing power and more advanced technology, AI might even hold the ability to self-improve and do beyond technical analysis, fulfilling a complete process of reinforcement learning in a trading process.

Black Swan Events in the Era of Artificial Intelligence

Another point the article mentions is that AI is still incapable of predicting a black swan event, and it could be worse because it may cause one. While computer malfunction could sometimes cause problems, as the 2010 flash crash mentioned in this article, this is never the main factor that is responsible for the black swan events. Instead, we should think about the key causes of Black Swan: errors, misconceptions, and uncertainty, etc. The first two factors are determined by a lack of checks and computer reaction ability. These two could be solved by checking trading algorithms and seeing whether it is sustainable. Also, we should bring up the fact that humans tend to make more mistakes in the trading process, and these errors usually vary by case. AI errors, on the other hand, are more predictable because they are programmed and modeled in the system. So once the mistakes are found out from the machine functioning, the same type of mistakes could be eliminated forever. As regards the uncertainty factor we discussed before, a chaotic system has memory, and patterns tend to repeat, so the algorithm is also capable of capturing the changes and uncertainties as well. Therefore, a data-driven algorithmic approach has a lot of potential to deal with these three sources of black swan events. Rather than blaming AI’s mistakes in a black swan event, we need to shift our focus on system maintenance and improvement and anticipate AI’s boosting performance in the future stock market. 

Artificial Intelligence Vs. Human Brains

One another reason for AI outperforming the market could be found out by looking at our brains. Many market meltdowns or flash crashes are actually due to human biases and human misjudgment. According to Prospect Theory, people think in terms of expected utility instead of the actual outcomes. Hence, by relying more on the machines’ objective decisions, people could avoid making irrational decisions and make more profits in the long run, and thereby possesses more power and chances to beat the market.

The I Know First Algorithm

As we discussed above, even though the stock market is a complex system that involves a lot of randomnesses, I Know First is using insights of chaos theory to predict the behavior of the market. The algorithm identifies waves in the stock market and generates forecasts in a heatmap format. Even though the algorithm is complicated, its displaying results are very straightforward to interpret.

The I Know First Market Prediction System models and predicts the flow of money between the markets. It separates the predictable information from any “random noise”. It then creates a model that projects the future trajectory of the given market in the multidimensional space of other markets. The system outputs the predicted trend as a number, positive or negative, along with the wave chart that predicts how the waves will overlap the trend. This helps the trader decide which direction to trade, at what point to enter the trade, and when to exit.

Recent Success Predicting Major US Stock Market Indexes using AI

The above graph displays a performance evaluation summary of I Know First’s long-term horizons forecast for 3 various long-term horizons. Despite volatility and uncertainties happening during the pandemic, the algorithm is still predicting the moves of indexes with great results across different time horizons. We can see that for a 1-year time horizon, the best hit ratio has reached 100% for the Nasdaq index and its ETF QQQ, as well as SPY; and the average hit ratio for this time horizon is 97.2%. Besides, for a 3-month time horizon forecast performance, the hit ratio ranged from 69% to 85%. And the highest one with 85% is achieved by the Nasdaq index. Even during the pandemic, all these long-term forecasts have performed with more than 62% accurately. Thus, we could see the algorithm has successfully picked up the patterns from historical data and has achieved great performances.

Conclusion

Although some articles argued that Artificial Intelligence will never beat the market, many of the concerns could be addressed with the actual results that I Know First AI algorithm achieved. The stock market’s chaotic behavior can be smartly forecasted by algorithms by utilizing chaos theory. Besides, nowadays machines are way more capable of learning historical data and knowledge than humans and actually elaborating on that knowledge in investment. It is also possible that in the future it could fulfill an automated trading process by reinforcement learning.

To subscribe today click here.

Please note-for trading decisions use the most recent forecast.