Thinking, Fast and Slow

by Daniel Kahneman

An extreme failure of this kind is unlikely because human learning is normally efficient. If a strong predictive cue exists, human observers will find it, given a decent opportunity to do so.