In Reply to: Your math assumes independent events. posted by mh on September 02, 2026 at 15:44:01
> Just look at the last three presidential elections, where the swing states broke the same way each time.
That only happened in the last 2 of the last 3 Trump elections. Otherwise, the polls are pretty accurate. With Kamala, the polling error was relatively small—about 1% to 2% nationally, which is within the typical margin of error.
As I mentioned, the bigger issue in those elections was that pollsters struggled to accurately model who would actually turn out to vote.
In elections where Trump wasn’t on the ballot—including the midterms before, during, and after his presidency—polling was generally accurate.
I agree that there could be a systemic issue in this election involving the difficulty of identifying the actual electorate. However, that kind of error generally doesn’t persist unless we’re dealing with an unusually different political environment, which I think is a possibility here.
If you assume that you will win every poll that you lead in, that frankly is bad math. If you believe that there is a systemic error in how all of the polls are conducted, then that would make a more logical conclusion that we will win so many races.