What is Algorithmic Trading?

Key Takeaways

• Quantitative trading mainly relies on data and models to find investment targets and investment strategies.

• The strengths of quantitative trading mainly include discipline, systematicness, timeliness, and diversification.

• The weaknesses of quantitative trading mainly include sample error and sample bias, strategy resonance, misattribution, and a black box.

Concept Illustration

In recent years, quantitative trading has emerged as one of the investment strategies with increased popularity. So, what is quantitative trading? Simply speaking, quantitative trading is the process of using computer technology with specific mathematical models to find investment ideas and implement investment strategies.

Traditional investment methods mainly include fundamental and technical analysis, while quantitative trading relies on data and models to find investment targets and make investment decisions.

Quantitative trading does not rely on personal feelings to manage assets but uses quantitative models based on appropriate investment ideas and experiences. It uses computers to process massive amounts of information, summarize the market dynamics, and establish investment strategies that can be reused and optimized repeatedly to guide the investment decision-making process.

In terms of application, quantitative trading covers almost the whole investment process, including quantitative stock selection, quantitative timing, stock index futures arbitrage, commodity futures arbitrage, statistical arbitrage, algorithmic trading, asset allocation, and risk management.

Strengths of quantitative trading

Compared with traditional investment methods, quantitative trading has both strengths and weaknesses, and the strengths are mainly as follows.

(1) Discipline

Traditional investing is largely influenced by human emotions such as greed and fear. It is sometimes challenging to ensure the discipline of trading.

Quantitative trading is based on discipline. Strictly implementing the investment instructions given by the quantitative trading model will not change randomly with the change of investor sentiment.

(2) Systematicness

The systematic features of quantitative trading mainly include multi-level quantitative models, multi-angle observation, and massive data processing.

The multi-level model mainly includes a large-class asset allocation model, an industry selection model, and a stock selection model.

Multi-angle observation mainly includes analysis of the macrocycle, market structure, corporate valuation, growth and earnings quality, market sentiment, and other perspectives.

Mass data processing means that quantitative trading can obtain data and information processing capabilities far beyond the human brain through computers, thereby capturing more potential investment opportunities.

(3) Timeliness

Quantitative trading can track market changes, constantly discover new statistical models that lead to excess returns, and look for new trading opportunities. Quantitative trading continuously looks for valuation depressions, and through comprehensive and systematic scanning, it seeks to opportunities brought about by mispricing and misvaluation.

(4) Diversification

The essence of quantitative trading's diversification is to seek by probability. This is manifested in two aspects: on the one hand, quantitative trading discovers rules from historical data, which are primarily strategies with a high probability of profit in the past; on the other hand, quantitative trading profits by selecting a portfolio of stocks, instead of one or a few stocks. Please note that diversification does not guarantee a profit or protect against losses in a declining market.

Weaknesses of quantitative trading

After walking through the strengths, the following lists several weaknesses of quantitative trading.

(1) Sample error and sample bias

Many quantitative trading strategies rely heavily on historical data. Nevertheless, historical data may lack sufficient diversity and long-term accumulation, so sampling could be error-prone due to a small number of samples or deviate due to non-random sampling. The correlation law obtained on this basis may become invalid once it leaves the sample range, thus losing its reference value.

(2) Strategic resonance

Many quantitative strategies are similar to technical analysis strategies. Once a particular strategy has proven effective, its effectiveness diminishes as the number of users increases, known as strategic resonance.

(3) Misattribution

The reason is inferred from the data results in the widely used multi-factor quantification strategy. As long as enough factors are constructed, it is likely to achieve a specific known result.

However, when a quantitative strategy based on this multi-factor combination is used in actual trading, it may fail due to misattribution. Because the cause is reversed from the effect, it is impossible to distinguish accidental and decisive causal factors precisely.

(4) Black box

Various quantitative strategies, including high-frequency trading, hedging, or arbitrage, often have no inherent causal relationship. The effectiveness of these strategies is mainly based on the strong correlation of historical data. The logic of the strategies lies in the fact that if there is a 55% or greater probability of being effective based on historical data, then the odds of winning will accumulate as long as there are enough data duplications.

But with only correlation and no understanding of intrinsic causation, investors cannot predict when history cannot guide the future. It's like a turkey whose owner comes to feed it every day, but the owner comes to kill it on the last day.