Who's calling it

No scores yet: grading starts November 3

Nothing has resolved, so nobody has a score yet. Once each race is called, we'll grade every call here.

How scoring will work. We turn each call into a probability for the side it backs. For a model, that's the number they gave. For a rating, we'll use a fixed scale (Toss-up 50%, Lean 65%, Likely 85%, Safe 97%), which we'll publish before grading. For a plain “X will win,” we'll use 75%. Then we use a Brier score: the squared gap between that probability and what actually happened, where 1 is a win and 0 is a loss. Lower is better. We do the same for Polymarket's price on the day of the call. A forecaster beats the market on a race if their Brier score is lower than the market's was at the time of the call.

Calls are each forecaster's latest public call we found for that race. A “contrarian” flag means the call backs the side Polymarket has as the underdog, or the forecaster's odds differ from the market by 10 points or more. Toss-up calls show how far the market leans from 50-50. Gaps use live Polymarket prices and update each time you load the page. *Implied from the post, for example “about a 1/5 chance” for the other side.