Judgment has a shape

Human judgment departs from normative logic and probability in systematic, replicable ways — not as random error.

Tversky and Kahneman mapped heuristics that usually work and sometimes fail: representativeness, availability, anchoring. People confuse plausibility with probability. They overweight what comes easily to mind. They cling to the first number they hear.

Prospect theory described how we treat gains and losses differently. Loss aversion, framing, and overconfidence are not insults; they are measured effects. They survive in experts, though expertise can blunt some of them.

The replication crisis trimmed some effect sizes and retired a few famous findings. The core pattern remains: intuition is not a Bayesian oracle. Institutions that want accuracy — courts, clinics, markets — need procedures that do not trust first impressions.