cookie-problem-posterior-0-6

IN premise — summaries/2026/08/24/wiki-Bayesian_inference-chunk-3.md

Created 2026-08-25T02:58:43+00:00

In the canonical cookie-bowl problem, with P(H₁)=0.5, P(E|H₁)=0.75, and P(E|H₂)=0.5, the posterior is P(H₁|E)=0.6.

Summary

This is a hand-verified benchmark for Bayesian updating: starting from two equally likely explanations, observing evidence that slightly favors one should shift its probability from 0.5 to exactly 0.6. It acts as a ground-truth anchor, so if the system's probability engine produces a different number for this standard setup, its core arithmetic is broken.