National polls don’t tell us much due to the electoral college.
Whether AL is +2 Trump or +20 Trump has no impact on who wins.
That’s a good point. Are there any models out there that “compensate” for that? I know Nate Silver’s one has been popular but I don’t know if that factors in the electoral weighing.
Perhaps the 538 simulation projector would be along those lines?
https://projects.fivethirtyeight.com/2024-election-forecast/
ah yeah, I had forgot about that. He was dead wrong in 2016 ;) I think he had Clinton at like 80% chance of winning, but I believe whatever errors they made back then were mostly corrected since.
Silver isn’t with 538 anymore, I think his model and the 538 model have the same basis though, probably diverging a bit since he left?
In any case, the point that an 80% chance of X is no guarantee of X got pounded home, don’t you think? :)