Who won the Ohio Senate race? How the forecasts did
Result as reported Nov 14 · final forecasts frozen Nov 2, 12:46 PM EST · VotePredictor
Sherrod Brown won. VotePredictor's final forecast gave Brown a 72% chance and a margin of Brown +1.8 — the right call. Scored on one scale against every forecast below, it ranked 7 of 14.
How the forecasts did
| Forecaster | Final call | Chance for Brown | Brier | Right? | Margin miss |
|---|---|---|---|---|---|
| FiftyPlusOne | Lean D | 75% | 0.063 | yes | — |
| Fox News | Lean D | 75% | 0.063 | yes | — |
| Sabato's Crystal Ball | Lean D | 75% | 0.063 | yes | — |
| Silver Bulletin | Lean D | 75% | 0.063 | yes | — |
| Split Ticket | Lean D | 75% | 0.063 | yes | — |
| Polymarket | 73% Brown | 73% | 0.076 | yes | — |
| VotePredictor | 72% Brown, Brown +1.8 | 72% | 0.076 | yes | — |
| Inside Elections | Tilt D | 60% | 0.160 | yes | — |
| Cook Political Report | Toss-up | 50% | 0.250 | no call | — |
| Decision Desk HQ | Toss-up | 50% | 0.250 | no call | — |
| The Economist | Toss-up | 50% | 0.250 | no call | — |
| MS NOW | Toss-up | 50% | 0.250 | no call | — |
| NBC News | Toss-up | 50% | 0.250 | no call | — |
| RealClearPolitics | Toss-up | 50% | 0.250 | no call | — |
| Poll average | Brown +2.9 | — | — | no call | — |
Brier score is (forecast probability − outcome)², lower is better. Ratings are converted to a probability the way the rankings board converts them: Solid 99%, Likely 90%, Lean 75%, Tilt 60%, Toss-up 50%. Result source: en.wikipedia.org.
What the forecast is built on
The forecast starts from the 30 public polls in the forecast window, whose plain average is Brown +2.9. Each poll is corrected for its voter screen and its pollster's measured house effect first (the “Adjusted” column below).
- The polls, weighted and corrected, read Brown +2.4.
- The fundamentals (the state's presidential lean, incumbency, the last result and the national environment) read Husted +1.5.
- The published forecast blends the two, mostly on the polls, at Brown +1.8.
How the forecast has moved
| Date | Brown's chance | Predicted margin |
|---|---|---|
| May 7 | 33% | Husted +4.7 |
| Jun 4 | 50% | Tied |
| Jul 2 | 42% | Husted +2.0 |
| Aug 6 | 42% | Husted +2.0 |
| Sep 1 | 55% | Brown +1.1 |
| Oct 1 | 78% | Brown +3.0 |
| Oct 26 | 76% | Brown +2.3 |
| Nov 1 | 74% | Brown +1.9 |
| Nov 2 (final) | 72% | Brown +1.8 |
The polls
34 polls are in the final forecast; 15 older ones have aged out and are shown dimmed. Margins are Democrat minus Republican.
The polls behind it
49 polls · 34 in today's forecast · newest firstDimmed polls are more than 61 days old. They stay on the record, but today's forecast no longer counts them. Each pollster name links to the poll's source; the grade under it is its VotePredictor grade from past elections. Margin is the poll as published. The ± under the sample size is the 95% margin of error on each candidate's share, from the sample size (the gap between them is about twice as uncertain). VotePredictor adjusted (“Our adj.” on small screens) is not a poll: it is our correction of that poll, for its pollster's record in past elections and for registered-voter or all-adult samples. It is the number our forecast reads.
Can you trust this forecast?
Every VotePredictor forecast is scored after the election against the same outcomes as everyone else's. Across 614 past races (Senate, governor, House and president), our final forecasts rank #2 of every forecaster that covered them all (Brier score 0.076, lower is better); FiveThirtyEight is first at 0.072. The full record is on the forecaster rankings and methodology pages.
This article is regenerated every day from the VotePredictor forecast, the poll database, the prediction markets and the rating shops' latest published ratings. Live numbers are on the race page.