Compare · Updated

How our forecast compares with the major forecasters

The raters and modelers below never feed into Who Shows Up; we build the forecast from polls, special elections and past results alone. Here is where we agree with them, where we don't, and how our method would have stacked up against them in past elections.

"Consensus" is the middle rating across the outlets listed below. A disagreement is at least a step and a half apart on the Safe–Likely–Lean–Toss-up scale, in a race someone rates as competitive.

Senate

Governors

Track record against the pros

Read these with a grain of salt in our favor. Our past-year numbers come from a backtest: the same method run after the fact on what was knowable then. Each piece was fit without the year being tested, but we designed the method in 2026 knowing how these elections turned out, while the outlets made their calls live. Our 2024 run also had no race polls (no public archive), which works against us. "Competitive" means at least one side rated the race below Safe. Past ratings are as listed on Wikipedia's election pages on each date, where every outlet's column is dated and cited; FiveThirtyEight's probabilities come from its archived forecast data. Only races both sides rated are compared.