Collider Bias

Collider bias occurs when a researcher conditions on a variable that is affected by both the cause and the effect. This creates a spurious association between the two, even if they are not causally related.

In trading, this often happens when analysts filter data based on an outcome that is itself a result of multiple, independent market factors. For example, focusing only on successful trades might introduce bias because success is a collider influenced by both strategy choice and market luck.

This bias can lead to the false belief that certain trading habits are effective when they are actually irrelevant or even harmful. Understanding collider bias is vital for avoiding the pitfalls of data mining and overfitting.

It requires careful consideration of the causal structure before applying filters or constraints to datasets.