4 ms·

So Bayesianism is good but only sometimes?

• dagw 8y ago
  Bayesian reasoning is good when applied rigorously, but can be dangerous when applied sloppily and lazily, since people are terrible at subjectively estimating probabilities and tend to be way over optimistic when updating their priors with any new information they receive. Basically people will treat information that should realistically cause them to update their priors from 50-50 to 51-49 as if it was information that should cause them to update to 80-20 (and then they round off too 100-0).