# How Accurate Were Prediction Markets in the 2026 Midterms? The Scorecard

*By Benny Ricciardi, Founder, PredictionMarketsPicks — The 7 Oracles at PredictionMarketsPicks*

The scorecard we pre-registered in August: Kalshi's favorite won 31 of 32 called Senate races and 46 of 46 called House districts. Race by race, misses included.

- Source: https://predictionmarketspicks.com/articles/how-accurate-were-prediction-markets-2026-midterms
- Published: 2026-11-04
- Updated: 2026-11-04

---

Three months before the election, we [published the grading rule](https://predictionmarketspicks.com/articles/prediction-markets-2024-accuracy-state-by-state), the race list and the denominator, and said we'd run exactly that whichever way it landed. Here it is, as of Nov. 4.

## The numbers

| Set | Scored | Called so far | Market favorite won | Lost | Pending |
|---|---|---|---|---|---|
| Senate (pre-registered) | 34 | 32 | **31** | 1 | 2 |
| House districts (pre-registered) | 66 | 46 | **46** | 0 | 19, plus 1 no call |
| Governor (not pre-registered) | 15 | 14 | 13 | 1 | 1 |

The price is Kalshi's last hourly close at 11 PM ET on Nov. 2. The side above 50¢ is the favorite. We scored each race win or lose, with no Brier score, the same method we used for 2024.

Race calls come from the Associated Press, or from Fox News where that's the call we could read directly. Each [race page](https://predictionmarketspicks.com/midterms-2026) credits its own call.

## The miss

**South Carolina Senate.** Darline Graham was the favorite at 67¢ and Annie Andrews won, a seat flip. It's the only called Senate race the favorite lost.

The closest called favorite that held was Iowa Senate, where Josh Turek won from 53¢.

## Markets vs. Sabato's Crystal Ball

The fixed comparison was Sabato's ratings, converted on a ladder we published in advance: toss-up 50, lean 60, likely 75, safe 90.

| Ratings used | Races | Market | Sabato |
|---|---|---|---|
| Final ratings, Nov. 2 | 32 called Senate races | **31 of 32** | 30 of 32 |
| The August table we locked | 12 called races | **12 of 12** | 7 of 10, plus 2 toss-ups |

The final ratings missed Iowa and South Carolina, both rated Lean R. The August table had Texas at Lean R, Iowa at Lean R and Kansas at Likely R. Democrats won all three. Sabato's ratings moved during the fall, so the final ratings are the fairer read of where forecasters ended up.

### Where they differed by more than 10 points

The test we called the one that actually matters: races where the market and the final ratings were more than 10 points apart.

| Race | Final rating | Ladder | Market (same party) | Winner |
|---|---|---|---|---|
| North Carolina | Likely D | 75 | 98¢ | Roy Cooper |
| Georgia | Likely D | 75 | 97¢ | Jon Ossoff |
| New Hampshire | Likely D | 75 | 94¢ | Chris Pappas |
| Michigan | Lean D | 60 | 79¢ | Abdul El-Sayed |
| Minnesota | Likely D | 75 | 91¢ | Peggy Flanagan |
| Texas | Lean D | 60 | 74¢ | James Talarico |
| Maine | Lean D | 60 | 73¢ | Troy Jackson |
| Ohio | Lean D | 60 | 72¢ | Sherrod Brown |
| Iowa | Lean R | 60 | 49¢ | Josh Turek |

The market's favorite won **all nine**. Using the August table instead, there were also nine such races, and the market's favorite won those nine too.

## Our three tests

In August we named three ways this could come out badly.

1. **The market's favorite loses more disagreements than it wins.** It did not: 9 wins, 0 losses.
2. **The market's Senate hit rate trails Sabato's.** It does not so far: 31 of 32 against 30 of 32, with two races pending.
3. **The redistricting math beats the chamber contracts on the House result.** Still open. It can't be graded until the House count is final.

## Still pending

**Alaska Senate.** Not called. Kalshi had the Democratic side at 71¢ on the final tape.

**Mississippi Senate.** Not called. Kalshi had Cindy Hyde-Smith at 83¢ and Scott Colom at 16¢.

**House.** 19 priced districts are uncalled, and we'll add them as they're called. CA-40 had no priced market on the final tape, so it's scored as a no call, as the rule requires. The 369 districts without their own contract aren't graded, because no district-level price exists for them.

This page updates as those races are called, whichever way they land.

## Outside the pre-registered set

**Louisiana Senate.** We set this seat aside in September because its market was found after the list was fixed. It's reported here and not counted. Julia Letlow won; Kalshi had her at 87¢.

**Governors.** We didn't pre-register these, so they're reported separately. The favorite won 13 of 14 called races. The miss was Texas: Greg Abbott at 60¢, Gina Hinojosa won. Kansas is uncalled.

Every race is on the [2026 results hub](https://predictionmarketspicks.com/midterms-2026), and the running forecaster comparison is on [markets vs. pundits](https://predictionmarketspicks.com/midterms-2026/markets-vs-pundits).

---

## Disclosure

PredictionMarketsPicks publishes analysis of CFTC-regulated event contracts. Nothing here is financial advice and every position carries risk.

Links to DraftKings, FanDuel in this article are referral links marked "sponsored" — we may be paid if you open an account. It costs you nothing and never changes what the model says.
