What Works on Wall Street Chapter 1: The Folly of Forecasting

"Models beat human forecasters because they reliably and consistently apply the same criteria time after time. In almost every instance, it is the human forecaster's failure to apply his or her rules consistently that results in the model outperforming the human." — James O'Shaughnessy

The Investment Context

James O'Shaughnessy's What Works on Wall Street is a cornerstone of quantitative finance. Rather than relying on theory or anecdotes, O'Shaughnessy backtested decades of stock market data (from 1951 onward) using the Compustat database to definitively answer one question: what investment strategies actually make money, and which ones are just noise?

The first conclusion of his massive data study is devastating to the traditional financial industry: human forecasting is completely useless. When tested against strict, emotionless, quantitative models, human experts consistently lose.

The Wall Street Translation

Wall Street pays analysts millions of dollars to predict the future of companies and the economy. The data proves these predictions are worse than useless; they are actively harmful to returns.

1. The Inconsistency of Humans: Humans are incapable of processing vast amounts of data without bias. We get tired, we get emotional, we fall in love with "story" stocks, and we panic during crises. Even when a human expert has a great set of rules, they will eventually break them because their "gut" tells them this time is different.
2. The Discipline of Models: A quantitative model (an algorithm) never gets tired. It never reads a terrifying news headline and decides to move to cash. It applies exactly the same criteria to every stock in the database, 100% of the time, without exception.
3. The Data Cannot Be Argued With: O'Shaughnessy proves that complex stories about the future of technology or macroeconomic shifts do not drive long-term returns. Cold, hard, historical financial ratios drive returns.

Actionable Trading Rules

1. Stop Listening to Experts: Completely ignore the price targets and buy/sell recommendations of Wall Street analysts. They are humans subject to extreme bias, and their historical track record is mathematically proven to be inferior to simple models.
2. Stop Trusting Your Gut: Never buy a stock because you have a "good feeling" about the company's new product. Your intuition is useless in a complex system like the stock market. Base your decisions purely on the data.
3. Build a Rules-Based System: Write down a strict set of quantitative rules for buying and selling stocks (e.g., "I will only buy stocks with a P/E below 15 and a dividend yield above 3%"). If a stock does not meet the rules, you do not buy it, no matter how much you like the CEO.