MENTAL MODEL · M12252

Bayesian Reasoning

DecidemediumDecision Science

Included

Updated 2026-08-15

Loading revision record…

INTRODUCTION

English translation pending.

CORE DEFINITION

Bayesian reasoning is the practice of revising a prior belief in the light of new evidence, where the posterior probability is proportional to the prior times the likelihood of the evidence. Its core commitment is that belief comes in degrees, so the right response to evidence is not to flip from disbelief to belief but to move the probability by the amount the evidence warrants. This guards against both overreacting to weak signals and ignoring strong ones, and it is the formal version of the ordinary intuition that credibility accumulates and erodes rather than switching on and off.

SCAFFOLDING EFFECT

psychology

Reduce cognitive load

- Carry a prior, not a verdict: state how likely you thought the claim was before the new evidence. - Ask what the evidence predicts: judge the evidence by whether it is more likely under your hypothesis or under the alternative. - Move by degrees: shift your belief by the weight of the evidence, and hold the updated value as the next prior.

anchor

Anchor fast decisions

The rule multiplies a prior belief by how well the hypothesis predicts the new evidence, then normalizes, so the resulting posterior reflects both where you started and how diagnostic the evidence is. Because the posterior becomes the next prior, belief updates accumulate over many observations rather than being decided by any single one, which is the mechanism that makes the method patient with noise and firm with real signal. The failure modes are symmetric and opposite: ignoring the base rate over-trusts the new signal, while a too-rigid prior absorbs evidence forever without moving.

MINIMUM ACTION

In progress 0/1

Practice this model in one real situation:

Before receiving the evidence, write down your prior probability for the claim. 2. Ask whether the evidence is more likely if the claim is true or if it is false. 3. Update the probability in the direction and to the degree the likelihood ratio warrants. 4. Carry the new probability forward as your prior for the next round of evidence.

Check to track your progress (stored locally)

Learning progress0%

account_treeGenealogyexpand_more

menu_bookReferencesexpand_more

Source support: Explicit

link
zh.wikipedia.orghttps://zh.wikipedia.org/wiki/%E8%B4%9D%E5%8F%B6%E6%96%AF%E6%8E%A8%E6%96%ADZH · Explicit
verified

RELATED MODELS