Monday, April 6, 2020

When the facts change ...



"When the facts change, I update my priors"
Dr Bill Hanage of Harvard, as quoted in the press

And so what exactly is he talking about?
Answer: Bayesian statistical methods

The essential matter in Bayes, as different from traditional frequency-of-occurrence statistics is that the starting point is a "guess" leavened by hypothesized or observed conditions.
  • The probability of event A, given the apriori conditions of B existing, has a dependency not only probability of B existing, but also the probability of A itself, and the probability that B can exist in the event of A 
  • Now, the conditioning of A by the existence or likely existence of B is what is called the "priors"
  • If by observation or otherwise it becomes evident that your understanding of B has changed, then an "update to priors" is required, thus bringing about a different understanding of A conditioned on B
That's a lot of words for what is essentially common sense: when the circumstance change, you should change with them, or at least understand the impact if you don't change.

"Update your priors": it's only common sense!


Buy them at any online book retailer!