3 ms·
Prediction market prices are more useful than polls, but not probabilities
- zahlman 2y ago>So if poll aggregators, such as 538, show a 50% favorability for Kamala, this converts to a less-than-50% chance of winning due to the aforementioned factors, giving a contract value of perhaps only 40 cents for ‘yes’ instead of 50 cents. 40% is the closer to the true probability or her winning, not 50%. 538 is not simply a poll aggregator; it's a complex model that attempts to estimate a probability by statistical methods (notably including Monte Carlo). It takes the effect of the Electoral College into consideration; and the idea that unknown and a priori unknowable "October surprises* would inherently favour one candidate or the other is simply nonsense. There's a reason Nate Silver gave Trump better odds in 2016 than all the poll aggregators with simpler models. It's not because he thought the polls were all garbage - after all, his model rates and weights the polls based on previously demonstrated predictive power.