Bayesian Updating
Arithmetic for changing your mind: start from a prior, weigh new evidence by how much likelier it is under one hypothesis than another, arrive at a posterior. For an investor a thesis is a prior, and each earnings report should move conviction in proportion to its surprise. Refusing to update is still Bayes, just with the evidence weighted at zero.
6 books · 1 people behind it · 4 resources
Start here
- The Signal and the NoiseNate Silver, 2012
- The Theory That Would Not DieSharon Bertsch McGrayne, 2011
- SuperforecastingPhilip E. Tetlock, Dan Gardner, 2015
The people behind it 1
Each person holds a book in this entry or names it in their own tags.
Journalists 1
Resources 4
- Bayes's Theorem: What It Is and Why It MattersShane Parrish · essay · Priors, likelihoods and how much to move.
- Bayes and Base RatesMichael J. Mauboussin · paper · How history should discipline a forecast's starting point.
- Bayes and Base Rates 2.0Michael J. Mauboussin · paper · Updated reference-class data for growth and margins.
- The Value of Probabilistic ThinkingShane Parrish · essay · Bayesian updating, fat tails and asymmetries together.
Books 6
Philip E. Tetlock · 2005
James Franklin · 2001 · How evidence and probability were reasoned about before formal statistics