Index

Mental Models

Bayesian Updating

A method of revising the probability of a belief by incorporating the strength and relevance of new evidence against prior probability.

Bayesian updating prevents both stubbornness and overcorrection by providing a disciplined method for adjusting confidence as information changes.

Core question

How much should this new piece of evidence shift my current belief?

Example

You believe a market is growing at 15% annually. A new industry report shows 8%. Instead of ignoring or fully adopting the new number, you adjust your estimate based on the report's sample quality and your prior confidence.

How to apply

  1. 1.State your prior belief and its confidence level explicitly.
  2. 2.Evaluate the quality, relevance, and sample size of new evidence.
  3. 3.Shift your belief proportionally — strong evidence warrants large updates, weak evidence small ones.
  4. 4.Record updates over time to build calibration.

Pitfalls

  • ·Anchoring too heavily on priors and underweighting strong new evidence.
  • ·Overcorrecting from a single dramatic data point.
  • ·Failing to distinguish between the quality and quantity of evidence.

Questions

How is Bayesian updating different from just changing your mind?

Bayesian updating is proportional and structured. You adjust by the strength of the evidence rather than swinging between extremes based on the latest headline.

Can teams apply Bayesian thinking in practice?

Yes. Have each team member state confidence levels before and after reviewing new data. This makes belief updates explicit and trackable.

Related