Novin Investment Bank

Algorithmic Trading

If you are active in the capital market, you have likely heard the term “algorithmic trading” frequently these days. In this short article, we aim to provide a definition of algorithmic trading and gain a basic understanding of this concept and its various strategies.
Algorithmic trading, sometimes referred to as automated trading, uses computer software to generate orders with specific instructions. These trades are executed automatically by computers at high speeds, which is not feasible manually by humans.
Algorithmic trading strategies can be categorized in various ways, and familiarizing yourself with some of these classifications can be interesting.
Based on the methods of identifying suitable trading opportunities, these strategies can be categorized as follows:
Trend-Following Strategies
The goal of these strategies is to timely identify changes in trends and take positions in their direction. One of the simplest strategies in this category is the use of moving averages.
Mean Reversion
These strategies are based on the idea that the average price plays a central role, with prices oscillating around it and returning to it after moving away. In this category, statistical arbitrage strategies can be mentioned; these strategies involve taking positions based on changes in the price trends of similar assets when they diverge from each other.
Alternative Data
Today, with advancements in data collection and analysis, the opportunity to use a broader range of data for company analysis has been created. A group of these data, which are not traditionally used for stock analysis, is referred to as alternative data. Data such as text, location, satellite images, and so on fall into this category. An interesting example of the application of these strategies involves analyzing satellite images of parking lots of retail chains to predict their sales before the release of financial reports.
Financial Ratios
In these models, variables such as the price-to-earnings ratio, book value to market value, and other ratios used to determine the relative cheapness of stocks are utilized. For example, one of the simplest of these strategies involves buying stocks with a low price-to-earnings ratio and selling stocks with a high price-to-earnings ratio.
Predictive Models
Another group of strategies focuses on predicting prices using quantitative methods and artificial intelligence. This group of strategies exhibits a high diversity in both the inputs received by the model and the processing methods. Various quantitative and statistical methods such as regression, machine learning models, and deep learning models fall into this category.
Receiving signals is only one part of an algorithmic trading strategy; other crucial parts of a strategy include the necessity of considering transaction costs and methods of executing trades. Many algorithmic trading strategies result in numerous transactions that, if transaction costs are ignored, can even be loss-making. Additionally, since these orders are automatically entered into the market, it is necessary to develop an intelligent model for executing trades that continuously monitors market conditions such as the status of the order book and the liquidity of the symbol.
As mentioned earlier, the common feature of algorithmic trading strategies is the execution of trades without direct human intervention and based on predefined rules.
In future articles, we will become more familiar with the applications of algorithmic trading and its advantages and disadvantages.

Scroll to Top