Base rate neglect - ignoring it when evidence arrives - is the most consequential reasoning error in applied probability, and interviews test it directly.
When an event is rare, the base rate dominates the posterior almost entirely. A 99%-accurate test for a 1-in-1,000 condition produces positives that are genuine less than 10% of the time, because the false positives come from a population a thousand times larger.
In trading, the same logic applies to signals: a screen that flags "unusual" activity in a market where genuine events are rare will mostly flag noise, however accurate it is in isolation.