bayesian-reasoninglisted

Activate when: user says 'Bayesian', 'prior', 'posterior', 'base rate', 'likelihood ratio', or 'update my belief'; someone treats 'the evidence is consistent with X' as proof of X; a high-stakes decision rests on interpreting a test result, security alert, fraud flag, A/B result, or hiring signal; base rates are being ignored in favor of a vivid story. Do NOT activate when: the decision is genuinely deterministic and probabilities do not apply; there is no data or domain knowledge to anchor a prior (Bayes amplifies information, it does not create it from nothing). More: deciqai.com/s/bayesian-reasoning
deciqAI/knowledge-skills · ★ 3 · AI & Automation · score 76

Install: claude install-skill deciqAI/knowledge-skills

# Bayesian Reasoning ## Overview Bayes' theorem: **Posterior odds = Prior odds × Likelihood ratio**. The strength of belief after evidence equals the strength before, multiplied by how diagnostic the evidence is. This skill applies Bayesian discipline where people reason about probabilities informally — and failures follow predictable patterns: ignoring the base rate (prior), confusing P(E|H) with P(H|E) (prosecutor's fallacy), over-updating on vivid confirming evidence, treating correlated evidence as independent. Composes with [`probabilistic-thinking`](../probabilistic-thinking/SKILL.md) (Bayes is the operational engine), [`critical-thinking`](../critical-thinking/SKILL.md) (formalizes considering alternatives), [`logical-fallacies`](../logical-fallacies/SKILL.md) (prosecutor's fallacy and base-rate neglect), and [`first-principles`](../first-principles/SKILL.md) (the prior is bedrock). ## When to Use - High-stakes decision rests on interpreting evidence (medical test, security alert, fraud flag, hiring signal, A/B result) - "Evidence is consistent with X" is being treated as proof of X - Base rates ignored — a rare event treated as probable because evidence "looks like" it - Correlated evidence pieces treated as independent updates - A benchmark score, AI-capability claim, AI-adoption stat, or AI-capex/valuation figure is being treated as proof without asking how often that signal appears when the underlying claim is false - Someone says "Bayesian," "prior," "poster