Following Donald Trump's decisive win over Kamala Harris, many are left questioning why polls, including Nate Silver's, had consistently shown the vice president leading in the race.
Trump swept to victory on Wednesday, winning 295 Electoral College votes to Harris' 226. Exit polls also suggest Trump will win the popular vote, with 46 of the 48 states declared shifting toward him since 2020.
But on the day before polls opened, Harris was predicted to win by almost every major forecaster. That included Nate Silver, whose projection gave Harris a slight edge in the Electoral College, with a 50-percent chance of victory compared to Trump's 49.6 percent. The model also showed Harris securing 271 Electoral College votes to Trump's 267.
Previous forecasts by Silver were equally close, but showed Trump with a narrow advantage over Harris. On Monday, Silver's forecast favored Trump with a 50.4-percent chance to Harris's 49.2 percent, and the day prior, Trump had a 52.6-percent likelihood of winning to Harris's 47 percent.
Nonetheless, in the final days of the election, Silver noted that the race was a "pure toss-up."
"When I say the odds in this year's presidential race are about as close as you can possibly get to 50/50, I'm not exaggerating," he wrote in his newsletter.
He also said there would be a high chance of a "clean sweep" on election night, with one of the candidates winning every swing state—a prediction that proved to be accurate.
According to Silver's forecast, there was a 20-percent likelihood of Trump winning every swing state, making it the most likely scenario to occur. Meanwhile, there was a 14-percent chance of Harris winning every swing state.
So far, Trump has taken Wisconsin, North Carolina, Pennsylvania, Michigan and Georgia. Nevada and Arizona have not yet been called, but exit polls show that the former president is currently ahead in both states, on 51 percent in Nevada and 52 percent in Arizona, making it very likely he will win both.
But while Silver's swing state predictions may have been accurate, his model gave Harris a more than 70 percent chance of winning the popular vote. Trump's victory in that bracket was perhaps the biggest surprise of this election.
The former president previously failed to win the popular vote in 2020, when President Joe Biden won the election, and famously in 2016, when Hillary Clinton won the popular vote, but failed to win the Electoral College.
On Wednesday, Trump noted his popular vote victory, thanking people for giving him an "unprecedented and powerful mandate." Trump is currently winning the popular vote by 3 points, according to CNN.
Newsweek has contacted Nate Silver for comment via email.
Silver's forecast is not the only one that predicted a win for Harris. On Tuesday, FiveThirtyEight gave the vice president a 50-percent chance of winning, forecasting her taking 270 Electoral College votes to Trump's 268.
Meanwhile, The Economist predicted that Harris would win 276 votes to Trump's 262. Both forecasts also showed Harris ahead in Michigan, Wisconsin and Nevada, while Trump was ahead in Pennsylvania, Georgia, North Carolina and Arizona.
Since 1998, election polls in U.S. presidential, House, Senate, and governor races have typically missed the final vote margin by about 6 points on average, according to FiveThirtyEight's analysis.
For example, in 2020, polls showed Biden with a strong lead over then President Trump. Just before the election, Biden led by 8.4 points in FiveThirtyEight's average. He ultimately won the popular vote by fewer than 4.5 points—just enough for an Electoral College victory.
Similarly, in 2016, FiveThirtyEight gave Clinton a 71-percent chance of winning, but she ultimately lost the Electoral College, despite taking the popular vote.
Scott Keeter, senior fellow at the Pew Research Center, previously told Newsweek that polls in 2020 and 2016 were inaccurate because Trump supporters opt out of surveys due to distrust in institutions like the mainstream media and polling organizations.
English (US) ·