Nate Silver’s election model, released on June 30 after the initial presidential debate, indicates that former President Trump leads President Joe Biden by six points in the probability of winning the popular vote in November.
The last Republican to secure the popular vote was George W. Bush in 2004, with 50.73 percent.
This marks the first time Silver’s election model has forecasted that Trump would surpass Biden in the popular vote. One day before the debate, the forecast showed Trump trailing by two points.
The model gives Trump a 68.4 percent probability of winning the electoral college, the determining process in the election. President Joe Biden’s chances of reelection are 31 percent.
Nate Silver's model gives Trump more than 68% to win the election against Biden. For the first time, Silver also predicts that Trump will win the popular vote by a wide margin. pic.twitter.com/ZXBHEgETvI
— Raylan Givens (@JewishWarrior13) June 30, 2024
Silver, who departed from FiveThirtyEight in 2023, unveiled the model on June 30, shortly after the debate.
Prior to the debate, Silver had assessed Trump’s probability of winning the presidency at 65.7 percent, which was three points lower than his most recent projection.
Silver elaborated on the methodology behind his forecast:
The model is the direct descendant of the f/k/a FiveThirtyEight election forecast2 and the methodology is largely the same, other than removing COVID-19 provisions introduced for 2020. (Other changes are documented here.3) If you’ve received this by email, we strongly recommend that you instead use the web version for interactive charts.
Polling averages are free for all readers, while forecasts, probabilities, and additional detail on model inputs require a paid subscription. Our latest narrative overview of the race — “The presidential election isn’t a toss-up” — can be found here.
The Silver Bulletin polling averages are a little fancy. They adjust for whether polls are conducted among registered or likely voters, the presence or absence of RFK Jr., and house effects. They weight more reliable polls more heavily. And they use national polls to make inferences about state polls and vice versa. It requires a few extra CPU cycles — but the reward is a more stable average that doesn’t get psyched out by outliers.
“Whitmer had phoned O’Malley Dillon with more of an unambiguous SOS: to relay that Michigan, in the wake of the debate, was no longer winnable for Biden” https://t.co/ZZz2uz2GCq
— Jonathan Lemire (@JonLemire) July 1, 2024
