Although note that in that case, although Will did find a coherent way of getting that sort of probability, if anything it underscores that Thomas's essential point there was correct: My probability estimate in that context was at best weird and more likely just poorly thought out, probably because of overcorrecting my overconfidence.