Did prediction markets correctly predict the 2026 midterms?

On which party would control each chamber, yes. Kalshi priced a Democratic House at 94 percent and a Democratic Senate at 74 percent on November 2, 2026, and NBC News projected Democrats would control both. They underestimated how many seats Democrats would win, particularly in the Senate.

Who won the House and Senate in the 2026 midterms?

NBC News projected on election night that Democrats would control both chambers, with at least 258 House seats against the 218 needed for a majority. The Alaska and Mississippi Senate races had not been called when NBC last updated its live blog on November 4, 2026, so the final Senate margin was not yet known.

Were Kalshi and Polymarket more accurate than polls and forecast models?

They were more confident in the same direction. Decision Desk HQ gave Democrats a 65 percent chance of the Senate and The Economist's median was 51 Democratic seats, against 74 percent on Kalshi. Both markets and models missed some of the wave, including South Carolina, so neither side can claim it saw the full result.

Why did so many 70 percent favorites win at the same time?

Because the races were correlated. A national swing toward one party lifts its candidates everywhere, so several 70 percent contracts tend to resolve together rather than independently. The same correlation would have made them all miss together in a different year.

How are the uncalled Alaska and Mississippi Senate contracts settled?

Each platform settles by its own written rules. Kalshi can resolve early once at least 4 of its 8 designated media sources call a race with no contrary call, while Polymarket relies on official and credible reporting with a 2 hour challenge window that disputes can extend. Recounts or court challenges can delay payouts, so the contract text is the thing to read.

Did midterm election prediction markets get 2026 right?

Kalshi and Polymarket had Democrats favored for the House and Senate on November 3, 2026, and were right on direction, but low on the size of the wave.

By Reva VermaPublished November 5, 2026Updated November 10, 2026

Did midterm election prediction markets get 2026 right?

Midterm election prediction markets got the direction of the 2026 midterms right: Kalshi and Polymarket had Democrats favored to win both the House and the Senate on election eve, and Democrats won both. What the markets did not get right was the size of the wave, and two Senate seats were still uncalled the morning after, so the scorecard is not finished.

That gap between direction and size is the useful part. A price of 74 percent is a claim about how often it would be right, not a promise.

What happened in the November 3, 2026, midterms?

NBC News projected that Democrats would control both chambers in January. Its election night live blog, last updated at 4:47 a.m. ET on November 4, 2026, projected the party would win 258 House seats, well beyond the 218 needed for a majority. NBC reported Democrats flipped 7 Senate seats: Iowa, Kansas, Ohio, Maine, North Carolina, South Carolina and Texas.

Two Senate races were not called. NBC described Alaska, where Mary Peltola challenged Sen. Dan Sullivan, as too early to call, with a result that could take weeks. In Mississippi, NBC reported that only 28 votes separated Sen. Cindy Hyde-Smith and Democrat Scott Colom, in a state NBC noted has no automatic recounts. Neither had been called when NBC last updated the blog on November 4, 2026, and this post does not guess at them.

What were Kalshi and Polymarket pricing on election eve?

On November 2, 2026, the Boston Globe reported that Kalshi priced a Democratic House at 94 percent and a Democratic Senate at 74 percent, while Polymarket gave a Democratic sweep of both chambers 74 percent and a split Congress 23 percent. It listed 4 Senate races with Democrats in the 70s: Maine at 72 to 73 percent, Michigan at 78, Texas at 74 to 75 and Ohio at 71. CNBC reported the same day that 74 percent was the highest Kalshi had ever priced a Democratic Senate, with the caveat that the platforms are not substitutes for polls.

Interest was unusually high. CNBC reported on October 30, 2026, citing the Anti-Corruption Data Collective, that wagers on 2026 congressional contracts had more than tripled those on the 2024 races.

Did midterm election prediction markets beat the polls and models?

On direction, the markets were more confident than the published models, and it paid off. Newsweek's November 3, 2026, roundup of forecasters showed Decision Desk HQ at 65 percent for a Democratic Senate and 84 percent for the House, and The Economist's model with a median of 51 Democratic Senate seats and Democrats winning the House in 93 percent of simulations, with a median of 235 seats. The Silver Bulletin's final update on November 3, 2026, said a Democratic House looked all but certain and that the polls pointed to a Senate win too.

On size, everyone was low. The Economist's simulated medians had Republicans holding Texas, Kansas, Iowa and South Carolina, and NBC reported Democrats flipping all four. Markets missed some of that too. CNBC reported on November 4, 2026 that before polls closed, markets had Iowa at 50 to 50, South Carolina leaning Republican, and Democrats as narrow favorites for at least 53 Senate seats with a 54th a toss-up. The CBS News live blog on November 4, 2026, put Democrats at 54 or more, with two races outstanding. Markets beat the models on confidence, not on seeing the wave.

How did prices move on election night?

Fast. CNBC reported that Polymarket's odds for Democrat Gina Hinojosa in the Texas governor's race jumped from 39 percent to 73 percent within 5 minutes of the first poll closings, and that Kalshi made Democrat Annie Andrews the favorite in South Carolina about 30 minutes after polls closed there.

Polymarket told CNBC its odds called 10 key races a median of 31 minutes ahead of CNN, Fox News and the Associated Press. Kalshi said it treated a race as called once a candidate passed 99 percent and trading on the other side dried out. A price crossing 99 percent is traders reacting to counted votes, not a decision desk projection. CNBC also disclosed that it has a commercial relationship with Kalshi, including a minority investment.

How do the midterm contracts actually resolve?

The fine print matters most in the races still open. A Covers explainer updated on November 3, 2026 says Kalshi can resolve a midterm contract early once at least 4 of its 8 designated media sources (including the AP, NBC News and Fox News) have declared a winner and none has contradicted them. Polymarket's Senate control market points to state election authorities, Congress and credible reporting, with a 2 hour challenge period under its UMA process that disputes can extend. Covers warned that recounts and legal challenges may delay payouts.

As of November 5, 2026, we found no reported resolution dispute on the House or Senate control contracts. The Alaska and Mississippi race markets are where one would appear, and CNBC reported both priced at 90 percent for the Democrat on November 4, 2026. For more on who decides, see prediction market resolution: who decides.

How should you read a 70 percent after this election?

As a 70 percent. Maine, Texas, Ohio and Michigan were each priced at 71 to 78 percent. NBC reported Democrats flipping Maine, Texas and Ohio, and CNBC reported Abdul El-Sayed winning Michigan. Treated as 4 independent events at those odds, the chance of all 4 favorites winning is about 30 percent. They were not independent: a national wave moves every race together. That is why a board full of 70s can all land at once, or all miss.

Barnard economist Rajiv Sethi told CNBC that a 25 percent outcome, like 2 heads from 2 coin tosses, should surprise nobody. CNBC noted that Kalshi had Lula da Silva at about 20 percent before polls closed in Brazil's election, and he won. One election cannot prove calibration; that takes hundreds of contracts, which is the argument in are prediction markets accurate.

What this election cannot tell you

A good night for the favorites does not show that traders knew something the polls did not. Thin markets can also be pushed around: the October 30, 2026, CNBC report cited the Anti-Corruption Data Collective finding that 88 percent of 18,000 active congressional markets could be moved 10 cents by a single wager under $1,000. Watch the price, and check the volume behind it.

Ravioli is a free-to-play prediction market. You trade with Sense, which is free, and the prizes are real.

Create a free account and browse the board.

Frequently asked questions

Did prediction markets correctly predict the 2026 midterms?: On which party would control each chamber, yes. Kalshi priced a Democratic House at 94 percent and a Democratic Senate at 74 percent on November 2, 2026, and NBC News projected Democrats would control both. They underestimated how many seats Democrats would win, particularly in the Senate.
Who won the House and Senate in the 2026 midterms?: NBC News projected on election night that Democrats would control both chambers, with at least 258 House seats against the 218 needed for a majority. The Alaska and Mississippi Senate races had not been called when NBC last updated its live blog on November 4, 2026, so the final Senate margin was not yet known.
Were Kalshi and Polymarket more accurate than polls and forecast models?: They were more confident in the same direction. Decision Desk HQ gave Democrats a 65 percent chance of the Senate and The Economist's median was 51 Democratic seats, against 74 percent on Kalshi. Both markets and models missed some of the wave, including South Carolina, so neither side can claim it saw the full result.
Why did so many 70 percent favorites win at the same time?: Because the races were correlated. A national swing toward one party lifts its candidates everywhere, so several 70 percent contracts tend to resolve together rather than independently. The same correlation would have made them all miss together in a different year.
How are the uncalled Alaska and Mississippi Senate contracts settled?: Each platform settles by its own written rules. Kalshi can resolve early once at least 4 of its 8 designated media sources call a race with no contrary call, while Polymarket relies on official and credible reporting with a 2 hour challenge window that disputes can extend. Recounts or court challenges can delay payouts, so the contract text is the thing to read.

Start playing free on Ravioli

Trade prediction markets and put your read on the world to the test. No real money, no risk. Just predict, trade, and climb the leaderboard.

Browse markets