Crowdsourced Election Forecast, As Well As My Own

Thank you to everyone who entered the 2024 election forecasting contest! The winners of the cash prizes will be announced once the election results are in and I can score the forecasts.

In the meantime, I used the forecasts from the contest to calculate a time-weighted average, basically averaging all the contest entries for each question, and giving more weight to the more recent ones, and less weight to the older ones. These predictions are listed in the crowdsourced_forecast column of the table below (shoutout to Datawrapper for the cool visualization).

I’m also publishing my own predictions here, in the mike_forecast column. I don’t have any fancy model or methodology here, they’re just guesses based on vibes. For what it’s worth, I live in Pennsylvania, and I’m also a libertarian and not a huge fan of either major party, so maybe that means I don’t have as much of a bias.

Anyway, once the results come out, I’ll compare these forecasts to the implied probabilities from Manifold (play-money) and Polymarket (real-money) prediction markets, as well as Nate Silver’s predictions for a subset of questions, and report how everyone did in terms of accuracy. Stay tuned!