Nate Silver Suddenly Ditches His Prediction Model

Nate Silver abruptly abandoned his prediction model just before 10:30pm Eastern time on Tuesday, posting on his Substack that his model was not "capturing the story of this election night well." As of 9:05pm, his model had Kamala Harris with a 53% chance of winning, which the Daily Beast calls a "sharp contrast" to the New York Times' voting needle, which, as of 10:30pm, was pointing toward an 81% chance of Donald Trump winning. Silver called the needle "a much better product," and added that his two-person team is simply not large enough to stay on top of results.

"We're spending a lot of time trying to fix code that isn't working right—and it's distracting from our ability to cover the election for you," he wrote. "We think we took on one too many things." On Monday night, historian Allan Lichtman called out Silver's model, posting on X that Silver's "compilation of polls is so unreliable that he now says that who will win the presidency is down to luck." USA Today reports that was in reference to Silver's own post on X Monday, in which he wrote, "It might literally wind up in the range where who's 'ahead' in our final forecast is determined by luck. There's still a little bit of variance introduced by running 'only' 40,000 simulations (we'll run 80,000 tonight but still...)." Both Lichtman and Silver, however, projected Harris as narrowly beating Trump. (More Nate Silver stories.)