Report Card: 2026 Forecasts

Summary

That's a wrap! The 2026 midterm elections are over. But how did our forecasts hold up? We got a composite score of B+ (an 87.2%, to be exact).

In the races we got wrong, we underestimated Democrats in >95% of them. In fact, of the 500+ races we forecasted, Republicans only won two races that we believed would go to Democrats (AZ-02 and CO-03, in case you were wondering).

The Generic Ballot

This year's polling errors were a major factor in our forecast. Polling data was used to inform our entire forecast, which explains many of the surprises on election night. The underestimation of Democrats in the polls led directly to the underestimation of Democrats in our forecast.

Here's how the generic ballot polling held up against the actual election results:

The Senate

The Senate was our biggest miss of the cycle, with six races forecasted incorrectly. Kansas was only characterized as a battleground on the eve of the election, while South Carolina and Mississippi never were.

Here's how the Senate forecast held up against the actual election results:

The House

A large portion of the House races we forecasted incorrectly were in newly-gerrymandered states, where no data existed under the congressional maps used this cycle. In fact, the state with the single most incorrect House predictions was North Carolina, which turned out to be a surprising Republican dummymander.

Here's how the House forecast held up against the actual election results:

Governors

Aside from Nevada, all three of the gubernatorial races we incorrectly forecasted were in states that also had a Senate race. In all of these, it appears the Senate race boosted the Democratic nominees for governor.

Here's how the gubernatorial forecast held up against the actual election results:

Summary

Just two years ago, in the wake of the 2024 presidential election, we wrote the following:

"Democrats have almost certainly lost the Senate for the remainder of the decade. Their path to 51 seats is nonexistent. Even if every Democratic incumbent wins in 2026 (which would require victories in states like Georgia and Michigan), the party would need to unseat four additional Republican incumbents (which would likely require victories in Maine, North Carolina, Florida, Texas, and/or Ohio). It's over for Senate Democrats."

Our forecast began with Democrats holding 45 Senate seats—they won 56. We had Democrats winning 235 House seats—they won 260. We had them at 19 governorships—they won 30. While our forecast improved from our original predictions, Democrats did things most people never thought possible.

Ultimately, this election cycle fundamentally changed the way we think about electoral politics. We look forward to using this year's results to deepen our understanding of the American electorate and sharpen our future forecasts. 2028 here we come!