Comment by Tom Chivers

On the Rootclaim thing: My impression was that they were doing Bayes badly. I have to admit I got a bit bogged down in the incredibly long article about it and all the back-and-forth, but as I understood it they really felt that you could come up with objective probabilities, rather than admitting that Bayes is a subjective process. You can't actually use all the information in the universe and you couldn't compute it if you could! But the fact that superforecasters demonstrably use a highly Bayesian process – base rates/reference classes as priors, inside view as likelihoods – shows that it is a powerful tool. I think it's extremely useful as an informal framework, letting people move away from "is this true/is this false" or "will this happen/won't it happen" to "I think it is X% likely to happen/be true", so they don't have to defend arbitrary bright lines or admit to being flat wrong – they can update and move gracefully from "I think it's likely" to "I think it's less likely" as more info comes in.