[somenameforme]

Oh boy. You don't even know how prediction markets or even odds work, at all, yet are abnormally strongly opinionated on the topic, and just assume you know everything without bothering to even check once. Of course. Prediction markets are binary yes/no. When Polymarket prices Trump as a 2:1 favorite, that does not mean he's expected to win 66% of the vote, but that if you ran the election 3 times, he'd be expected to win twice.

Polymarket's accuracy (as defined by an event corresponding with the majority position) is 90% a month out from an event, and 94% 4 hours out from an event. I suspect a weighted analysis (e.g. a 90% event should happen 9 of 10 times, not 100%) would show even better results. That is dramatically higher than you get from things like pollsters, and it's all freely available to the public.