Drawing solely from data, such as financial statements, price, volume, and macro reports while employing robust models with artificial intelligence and machine learning, every single argument leads to the conclusion: stock price prediction, while not purely a random walk, should be exact science. This article opts to an alternative viewpoint, asserting stock price prediction is a specialized branch within social science.

Is Stock Price Purely Random?

Many passive investing strategies believe that stock price prediction is purely random and unpredictable. A practical application of this philosophy is to invest in exchange-traded funds (ETFs). For retail investors, this approach has proven effective in reducing huge fees while achieving comparable long-term performance to active investment. However, there’s also evidence that stock prices are not purely random.

First, if stock prices were entirely random, how could the market consistently yield positive returns over the long term? Even naïve investors recognize that the overall economy tends to improve gradually after years. This leads to index growth and resulting profits. This approach with passive investing supports the low-fee approach and proves the stock market index is predictable in the long horizon.

Second, if stock prices were purely random, a company’s market capitalization would be detached from its core business fundamentals. This implies that a startup in Vietnam could be as valuable as Alphabet (Google), which does not make sense.

Third, considering that financial markets serve as the backbone of major global economies, how can stock prices operate purely on randomness? The impact of governance intervention and monetary policy further proves the connection between stock markets and the broader economic systems.

Considering these arguments, it becomes clear that stock prices exhibit some randomness but are not purely random. According to the Efficient Market Theory, stock prices are not random, and no one can make profits in the stock market since all available information is already reflected in the stock price. This theory assumes that all market stakeholders have the same precise knowledge and consistently agree on stock prices for every event. From the global development history, however, achieving global consensus on a random event is an unrealistic expectation.

In short, stock prices are not purely random.

Is Stock Price Prediction an Exact Science?

To answer this question, let’s explore factors that may impact stock prices.

While this list is not exhaustive, it highlights factors impacting stock prices. Despite an investor’s efforts, predicting future outcomes remains challenging due to the unpredictable nature of events. For example, the Covid-19 pandemic is a one-time event and no models can predict how it affects stock prices. Thus, stock price prediction isn’t an exact science. Trying to predict with high accuracy can be futile given the mix of predictable and unpredictable factors.

This leads to an essential insight: there’s no highly accurate model for stock price prediction. To quantify this argument, let’s compare this to natural science. Natural science relies on models. Like in physics, Newton’s universal law of gravitation and classical mechanics models yield convincing outcomes, such as an apple falling to the earth. In production, accuracy often exceeds 99.99%. Scientific models typically prove themselves with over 95% confidence. However, in stock trading, our guess of good model accuracy after all fees and taxes is only around 55%.

Trading stands apart due to a key insight: even completely wrong actions can yield favorable results with a high probability (around 35%). In other scientific fields, wrong decisions only bring disasters.

In short, a wrong action gives a trader 35% chance of success, while a right action increases it to 55%. The challenge lies in discerning whether a successful result comes from a right or wrong action. This makes stock trading extremely hard. In self-driving car design, if the car sees a red light, it must stop. That’s common knowledge, and a right action with known good results. In trading, however, no rewards or penalties can be traced by right or wrong actions.

Fortunately, algorithmic traders leverage the rule of large numbers to profit from robust models. Proving a model’s effectiveness involves applying machine learning and statistical techniques.

Stock Price Prediction – A Discipline Branch Within Social Science

Social science differs from exact science since it deals with social and human aspects. However, there are many interesting links between social science and stock price prediction.

First, there are way too many factors that cannot be fully understood in both disciplines. Imagine solving a decision-making problem with only 2 known factors while the total number of impacting factors is 20.

Second, attributing a single outcome is exceptionally challenging. It’s difficult to trace back a good result to a right decision or a wrong decision. Looking at the past data may give obvious explanations, yet the future remains highly unpredictable.

Third, despite years of research, no concrete methods for accurate predictions exist. In the article “The Forecasting Collaborative. Insights into the accuracy of social scientists’ forecasts of societal change”, experts in social science still cannot foresee societal changes and make predictions reliably. This is similar to the story of stock trading experts.

Lastly, societal changes impact the economy, directly influencing market indices. This is a strong connection between the two disciplines. If one can predict the next societal shift and economic shift, they can make huge profits in trading. In essence, mastering sociology holds immense value for stock price prediction.

In summary, this article suggests that stock price prediction and social science exhibit commonalities. It proposes that grasping sociology could enhance stock price prediction more than learning exact science such as artificial intelligence and machine learning. It’s important to note that predicting societal changes may not directly impact an individual stock, but can significantly influence an entire economic sector or a market index.