Ah, that's what I was missing. That said, regardless of how often these events are occurring, we shouldn't be all that surprised by a 1/5 event.

Thinking further about this, I think framing it in years is highly misleading. If we imagine an event that occurs once every ten-thousand years, with even probability of two outcomes, then if I pick one of the two as my guess and I correctly assign a 50% probability, you can say "you're saying your guess will only be wrong, on average, once every twenty-thousand years!" That's technically true, but I'm clearly not making too bold a claim!

Yeah I sometimes question the sensibility of even trying to assign probability to what are essentially one-time events. I prefer to think of it, the probability is either 0% or 100%, we just don't know it yet!

In a deterministic universe, there's a sense in which that's always the case. But really, probability is always "given the information available". "I know everything about a deterministic universe" is just a special case.

Related to the broader topic: https://en.wikipedia.org/wiki/Scoring_rule#Proper_scoring_ru...