> Unpopular opinion, but FiveThirtyEight constantly got the polls wrong and this isn't surprising. 538 almost never got the polls wrong, they simply reported t

> Unpopular opinion, but FiveThirtyEight constantly got the polls wrong and this isn't surprising. 538 almost never got the polls wrong, they simply reported the results the pollsters got. But you probably meant something else was wrong, and not their reporting of the polls. Like, maybe, the poll-based predictions that the polls supported. But their poll-based predictions of individual results across all of the individual elections they predicted tended to be reasonably accurate in the Silver era, in that results predicted as, say, 70% chances were actually correct about 70% of the time, etc. In the post-Silver Morris period, I haven't seen any after-the-fact analysis, but the universe is a lot smaller, so you'd also expect the actual results on that limited universe to be less accurate in representing the long-term accuracy of the modelling approach. The main problem, though, is that people want predictions to have probability 1 or 0, and don't deal well with things in between.