Glossary term
Overfitting
When a model learns historical noise instead of a real pattern.
Overfitting happens when a model or strategy is tuned so tightly to past data that it looks impressive in a backtest but fails in live trading. More parameters, more experiments, and more freedom to choose results all increase overfitting risk.
Example: Choosing the best of 500 tested signals after looking at the full history can create a strategy that only worked by accident.