Progression 

Trading has changed immensely over the years, especially with the introduction of technological innovation. Financial technology (fintech) is one technological innovation that assists with making faster and more accurate financial decisions. In this article, six common trading strategies will be discussed. However, to understand trading concepts it is beneficial to know how trading has developed over the years.  

What Is Discretionary Trading? 

Prior to technological innovation, trading was conducted by brokers. This form of trading is known as discretionary trading. In discretionary trading, a broker makes all trading decisions autonomously for the customer. The broker will make these trading decisions based on experience, knowledge, and a baseline level of intuition. The primary advantage of discretionary trading is the convenience of not needing to personally manage trading decisions and constantly communicate with a broker. The disadvantage with discretionary trading is having to pay fees and potential sub-optimal performance. Discretionary accounts will often require a minimum account balance and 1-2% a year in assets under management (AUM). AUM essentially measures the total dollar value of all financial assets entrusted to a manager or firm by their clients. Poor performance is always a risk with any trading approach. However, discretionary trading can have a higher risk of low performance due to human irrationality and poor decision-making.  

What is Algorithmic Trading? 

As computer technology was introduced to trading, algorithms were created to increase trading efficiency. At its core, algorithms are a series of instructions given to a computer for calculations or other problem-solving applications. An algorithm that follows a conventional technical analysis is called algorithmic trading. Algorithmic trading utilizes the computational power of computers to make quick decisions regarding buying and selling within a financial market. The advantages of algorithmic trading includes: efficiency, elimination of irrationality, and testability (testing against historical data). Using technology can streamline the trading process to seize opportunities via faster information analysis and trading speed. Technology also removes the human intuition factor and replaces intuition with information-based decision-making for better trading results. Algorithms can be tested on historical data to perfect and optimize the system before deployment. A disadvantage of this method of trading is that the computations are based on historical data which may not accurately represent current or future market conditions. Additionally, when black swan events occur, the algorithm is often unable to adapt to the changing market situation without direct human intervention (usually via stopping trading to adjust the algorithm). 

Black-Scholes Model  

One of the first widely used algorithmic models is the Black-Scholes model. It was developed in 1973. The model is used to calculate the theoretical value of option contracts. The Black-Scholes model requires 5 variables:  

• The strike price of the option   

• Volatility  Underlying asset price 

• Risk-free interest rate   

• The time to expiration of the option 

The disadvantages of this model are limited usefulness and assumptions regarding consistency. The Black-Scholds model only prices European options, which makes the model impractical for other markets without significant adaptation. At a technical level, the model assumes volatility is constant, liquidity is infinite, risk-free rate is constant, price changes are random, and price changes are normally distributed (bell curve). These assumptions limit the unmodified model to options trading (which tends to have low market fluctuations). In sum, Black-Sholes showed that algorithmic trading could eliminate many risks associated with human-judgment-based trading. 

What Is Quantitative Trading? 

As technology has advanced further, more advanced mathematical methods have been developed. A subset of algorithmic trading is known as quantitative trading (“quant trading”). This method uses more advanced mathematical models that use a variety of different datasets. Quant trading is used to automate analysis, monitoring, and trading decisions. Similar to algorithms, the advantage of quantitative trading is the elimination of human irrationality. Additionally, if quantitative traders do not alter these complex algorithms, the model will fail when market conditions change.  To correct this disparity, machine learning is implemented into quantitative trading. The hope of integrating machine learning with quantitative modeling is to automate the learning process from historical data, identify correlations, and make more accurate predictions. This eliminates the need for humans to alter models when market changes occur via machine learning automatically altering the model to fit market trends instantaneously. However, there is a risk of hallucination. Hallucination events occur when an AI makes false statements from seemingly nowhere. There is also a risk that an AI-based quantitative model will rely on inaccurate, false, or misleading data in the processes of machine learning to create models that are inaccurate representations of markets. With further development of AI technology, these problems may be resolved.  

Quantitative Trading Strategies 

There are six common strategies for building mathematical models for quant trading, such as: sentiment analysis, mean reversion, statistical arbitrage, algorithmic pattern recognition, trend following, and momentum investing.  

Sentiment Analysis 

Sentiment analysis is a unique strategy of quantitative trading. Sentiment analysis does not use market data. Using natural language processing, the model analyzes mass amounts of text from social media, research reports, or news articles to understand general attitudes toward securities. The process involves categorizing words by assigning numerical values to indicate whether a security is perceived as positive or negative. Sentiment analysis is used as part of larger strategies to better predict the market. The risk associated with sentiment analysis is the subjectivity of assigning value to text and the delay in sentiment being shared.  

Mean Reversion 

The concept of mean reversion is security prices will tend to move towards an average. This means that a security over a long period will return to an average level regardless of fluctuations. Quant trading allows rapid detection of markets with a long-term average. This understanding can indicate when to sell or buy via where the price falls in comparison to the long-term average. This strategy reduces the risks associated with market fluctuations caused by irrational behavior.   

Statistical Arbitrage 

Statistical arbitrage applies the same logic of mean reversion but on a macro level. Mean reversion might be observing a specific security. Statistical arbitrage focuses on analyzing a group of similar assets to determine an average price. The same logic from the mean reversion of buying and selling depending on the average applies to statistical arbitrage. 

Algorithmic Pattern Recognition 

Algorithmic pattern recognition seeks to recognize the patterns of large firm trades. Large institutional firms tend to use algorithms for trading. These trades are made through multiple exchanges/brokers to not affect market price. This quant strategy aims to uncover the otherwise hidden algorithm pattern to make decisions before prices increase due to these large firm trades. 

Momentum Investing  

Momentum investing is a short-term strategy involving looking at trends in equity price fluctuations. The aim is to use volatility as an advantage to find buying opportunities. Once trends are identified, investors will buy rising stocks and sell them near the peak. A common indicator of stock momentum is the relative strength index (RSI), which measures the magnitude of price changes to determine if a security is overbought or oversold. 

Trend Following  

Trend following is similar to momentum investing but is more of a long-term strategy. Additionally, trend following can be applied to a wide range of assets while momentum investing primarily works with equity. Both momentum investing and trend following look at price movements to identify pattern