Bayesian thinking
Bayesian thinking is updating your beliefs in proportion to new evidence — starting from a prior probability and revising it as data arrives, rather than holding fixed opinions. Strong evidence should shift you a lot; weak evidence, only a little.
✦ Foundational — cross-referenced 23× across this reference (17 related ideas · 3 comparisons · 2 books) · The State of Thinking 2026 →
How it works
You begin with a prior (your best estimate before new data), weigh how likely the evidence is under competing hypotheses, and update to a posterior. Beliefs become probabilities you revise, not flags you plant.
The skill isn't holding opinions loosely — it's updating them by exactly the right amount.
How to use it
- Hold beliefs as probabilities, not certainties, and update them when evidence arrives.
- Weight new evidence by how surprising it would be if you were wrong.
- Start from base rates (priors) before reacting to a single vivid data point.
Worked example
A test is 99% accurate and you test positive for a rare disease (1 in 10,000). Naively that feels near-certain — but factoring the low prior (Bayes), most positives are false alarms. The base rate dominates.
Where it fails
Garbage priors or motivated weighting corrupt the update — Bayesian reasoning is only as good as the honesty of your priors and your reading of the evidence. It’s a discipline, not a calculator that removes judgement.
- It requires an exhaustive hypothesis space — no amount of correct updating rescues you if the true explanation was never on your list.
- For unprecedented, one-off events there is no principled prior, and the arithmetic dresses an arbitrary starting guess in unearned precision.
- Updating on each item as it arrives ignores where evidence comes from; correlated or selectively-presented sources can push a rigorous Bayesian to a confident wrong answer.
The counter-model: The black swan — Bayesian updating refines probabilities within the hypotheses you hold; black-swan logic warns the decisive event may live outside them entirely — update diligently, but keep exposure to what your priors cannot see.
How to apply it, step by step
- State your current belief on the question as a rough probability, and write it down.
- Before seeing new evidence, decide how much each possible result should move you.
- Collect the evidence, checking whether its source is independent of previous evidence.
- Move your written probability by the pre-committed amount — no more, no less.
- Act when the probability crosses the threshold your decision actually requires.
The deeper point
"Strong opinions, weakly held" misses the point — the skill isn’t holding loosely, it’s updating by the right amount. Most people either won’t move on strong evidence or lurch on weak evidence; calibration is the rare middle ground.
Frequently asked
- What is Bayesian thinking?
- Treating beliefs as probabilities and updating them in proportion to new evidence — starting from a prior and revising toward a posterior as data arrives.
- Why do base rates matter in Bayesian thinking?
- Because a rare condition’s low prior can outweigh strong-looking evidence — most positives on an accurate test for a rare disease are still false alarms.
- How is Bayesian thinking different from being open-minded?
- It’s quantitative and disciplined: you don’t just “stay open,” you shift your probability by how much the evidence actually warrants.
Biases this model helps counter
Related
Keep reading
Just-world hypothesis
Blaming the victim is often the price of keeping faith that the world is fair.
Go deeper
The book behind this idea: The Signal and the Noise by Nate Silver. Hear the whole thing free — start an Audible trial and your first audiobook is on the house.
Read the full summary of The Signal and the Noise →
More canonical picks:
- Thinking, Fast and Slow — Daniel Kahneman
- The Art of Thinking Clearly — Rolf Dobelli
- The Great Mental Models, Volume 1 — Shane Parrish
- Poor Charlie’s Almanack — Charlie Munger
- Super Thinking — Gabriel Weinberg & Lauren McCann
- Seeking Wisdom — Peter Bevelin
As an Amazon Associate, ReadGlobe earns from qualifying purchases and Audible trials — at no extra cost to you.
Cite this page
ReadGlobe. (2026). Bayesian thinking. https://readglobe.com/model/bayesian-thinking/
"Bayesian thinking." ReadGlobe, 29 May 2026, readglobe.com/model/bayesian-thinking/.
Primary source: Wikipedia
Editorial synthesis © ReadGlobe 2026, drawing on the mental-models tradition (Charlie Munger, Farnam Street) and the primary sources for each model. · Last reviewed 2026-05-29.