Nate Silver had a great point about how difficult it is to evaluate who had the most accurate prediction, since people are offering win probabilities and we only observe one realization.

Given the probabilities assigned by Polymarket, the observed outcome was actually very unlikely, so it isn't obvious how much credit they deserve. There might have been other predictions where the colors don't match perfectly, but the actual outcome had a higher probability.

All that said, I do think prediction markets are the best method for doing this sort of thing.