Is that a 65% chance of a Trump electoral college or popular vote victory? I think last time it predicted that Hillary would win the popular vote. Then she did.
ITT: People not understanding probabilistic forecasts.
Nate Silvers models are not perfect, but pretty close. You can do retro analysis on them to see. If someone with a 5% chance to win actually does win, that doesn’t mean the model was wrong… That just needs to actually happen only 5% of the time.
Getting elections"wrong" as people are talking about here shows they don’t really know how these models work. If you haven’t looked at the retrospective analysis on these results, then you are likely unqualified to declare that they are useless or wrong.
Add comment