AI-Driven Algorithmic Trading: Disrupting The Disruptors

I Know First Research Team LogoThis article was written by the I Know First Research Team.

The investment world has recently gone through at times a quiet and secretive, yet dramatic revolution as machines entered an area where man used to reign (and trade) supreme. The first major step in the process came in late 90s as algorithmic trading came in on the back of advanced quant funds and first high-frequency traders. Now, as the world is embracing the artificial intelligence revolution, another tectonic shift is on the way. What is AI-driven algorithmic trading all about, and why should you be excited about it? Keep reading to find out.

Traditional Strategies: An Overview

But before we talk about all the disruption happening around, let us give the disrupted their due and quickly go over some of the more conventional and venerable approaches.


In most essential terms, we can differentiate between short-term and long-term conventional strategies. The former are about trading and speculation, in the sense that you do not really intend to hold on to what you are buying for too long. Long-term strategies, however, like value investing, put the emphasis on the investment part and are thus largely outside the scope of this article. Nevertheless, different speculative strategies imply different time frames, so further differentiation is quite legitimate here.

The next key distinction to make is between swing trading and position trading. The latter suggests a longer-term approach, in the sense that the trader can keep the financial instruments in their possession for up to a few months. The stocks in question would normally be the ones to show the highest potential in terms of their upwards trend. Thus, by holding on to them for a longer period, the trader seeks to make the most out of this trend rather than just cashing it in at once for a quick buck.

Swing traders, on the opposite, do not tend to hold on to their guns for more than a few weeks; in fact, most of the time, the period between them entering and exiting the market is way shorter. Trends are still what they are looking for, not necessarily positive trends: shorting is an option as well, for stocks that are expected to go down.

The details and intricacies differ a lot. Some traders opted for intraday trading, making their profits off the daily changes in stock prices. One of the strategies utilized by such traders is scalping, which seeks to make up for the size of the gains by the number of transactions – a similar principle lies at the foundation of high-frequency trading, by the way, but let us not get ahead of ourselves.        

In picking up the stocks to long and short, traders can utilize both fundamentals and technical analysis. The former is all about trying to find out whether the current price of the company’s stocks reflects its current standing and potential, while the latter mostly focused on historical trading data and the patterns within it, as well as indicators like the moving average.

Now that we have a gist of the conventional strategies, let us take a look at what happened next as the new technology disrupted the financial conventions.

Algo-Traders And Quant Funds: New Players In Town

As obvious as it sounds, the key feature of algorithmic trading is that it relies on a sophisticated computer algorithm to strategize and do the trading. In doing so, it capitalizes on the advances of computer science and other boons of the high-tech era.


High-frequency trading, which dates all the way back to the late 90s, is one of the prime examples of how computers changed the world of investment. Instead of playing on trends, this strategy, most often used by large institutional players, seeks to double down on market movements caused by whales. When a large fund drops off a couple of millions of stocks, its price can slip for a few minutes before bouncing back. These few minutes is where the high-frequency traders make their profit by buying this stock cheap and selling it as soon as the price goes back up.

Granted, the profit margin would be quite small on such transactions, but the idea behind this kind of algorithmic trading is to conduct dozens and dozens of such transactions per second. Another key point is to trade at a high volume. And, finally, by virtue of injecting liquidity into whatever platform they operate on, high-frequency traders are often able to save up on commissions or other dragdown expenses.

Algorithmic trading, of course, requires powerful computers able to pick up the signs of the next temporary fluctuation and stepping in before the market has the time to react to it. The machines have to crunch the numbers to calculate whether the new fluctuation will be followed by a bounce-back or is a long-term trend in the making, which would make it a bad fit for high-frequency trading.

While high-frequency trading was initially as big as it was controversial, heavily increasing the presence of the machines on the stock market, the latest reports suggest that rising competition and other factors have resulted in a shrinking appetite for such operations. But before we go on to the new big disruptors, we need to mention another practice that was given a boon in the times of high-tech: quantitative, or quant trading.

Quant trading is also based on crunching the numbers in order to try to simulate the market dynamics and pick up the best stocks to invest in. The two most important variables such algorithms look at are stock prices and trading volumes. By virtue of operating with large volumes and quite often also adopting the aforementioned rapid-fire approach, some quant funds have quite a lot in common with high-frequency traders.

AI-Driven Algorithmic Trading: The Great Disruption

Enter the age of AI, and with it, new opportunities for smarter trading. The investment world has not been blind to this, and predicting the stock market is a major challenge in contemporary machine learning.


One of the leaders in this sphere is I Know First has, as Israel-based company that has developed a deep learning AI delivering daily forecasts for over 10,500 financial instruments, including stocks, ETFs and currencies. Trained on a historic dataset covering 15 years of trading, it approaches markets from a holistic perspective, looking for signals in the fresh trading data and using those to model the trends and seasonal patterns. In doing so, it looks at the direction which assets across various investment universes are most likely to follow and measures the expected intensity of these price swings.

The forecasts are delivered as a heatmap with two numeric indicators: signal and predictability. Signal shows how a given asset is expected to perform against others on the list. Predictability is an indication of how much of a good job the algorithm has done in its prior predictions for the asset. It ranges from -1 to 1 and is defined as the Pearson correlation rate between earlier forecasts and actual price movements.

The system is thus of value for both institutional and retail investors, who can easily identify top investment opportunities by picking out the stocks with the highest values for both indicators. The predictability indicator also helps in assessing the risks of an investment portfolio.

The algorithm can adapt to periods of market volatility, because its design incorporates elements of genetic coding. To put it simply, it keeps track of its own performance and updates its predictive models as soon as they start to lose their grasp. This also ensures that the prediction accuracy goes up with every prediction delivered.

The AI also draws on chaos theory to account for market volatility. It delivers its forecasts for time horizons ranging from 3 to 365 days, covering short, medium and long-term perspectives, which helps the traders to identify opportunities both for a quick profit from short trading and find the stocks that would generate most value over time.

This approach offers a new tool for traders as they seek to beat the market. The AI predictions are based solely on objective trading data and the advanced mathematics driving the artificial industry as such. No human emotion or bias comes into play at any stage of forecast generation, and the predictability indicator offers a clear measurement of the risk.

Now, the potential issue to look out for is that the predictive AI tends to do better with longer periods of time, which have more room for seasonality patterns. In a string of recent evaluations like this one, it demonstrated an accuracy of around 60% for 3-day predictions, rising up to 80-90% for 3-month forecasts. This means that swing traders would have to make sure to keep their portfolio diversified enough to generate gains on 60% hits.

For position traders, however, this all but guarantees that they will be able to make the most out of their investments. The I Know First AI will help them identify the best stocks to pick up and hold for the next few months as well as a way to seize the moment when the trend turns and it is time to exit the market.