Multiple Hypothesis Testing
Multiple hypothesis testing occurs when a trader tests many different variables or combinations of indicators to find one that appears statistically significant. Without proper adjustments, the probability of finding a false positive result increases significantly as more tests are performed.
In finance, this often leads to the belief that a specific indicator predicts price movement, when in fact, the result is just a product of chance. This is a common pitfall in high-frequency trading and algorithmic development.
To correct for this, researchers use statistical methods like the Bonferroni correction to adjust the threshold for significance. Failing to account for multiple tests leads to the deployment of strategies that have no actual predictive power, resulting in financial loss when the random pattern disappears.