2026 midterm markets, one month out: House, Senate, and the close races

Election Day is Tuesday, November 3, 2026. That is 31 days from the afternoon these prices were checked. All 435 House seats are on the ballot, along with 35 Senate seats. The two contracts people actually trade are simpler than the map: which party controls the House, and which party controls the Senate.

This is a reading of those prices. It is not a preference for either party, and it is not a forecast we are asking you to bet. Prices move every day. Anything below is a snapshot from October 3, not a number to memorize.

The House is the expensive favorite

On the afternoon of October 3, the Democratic side of House control was about 92 cents on Kalshi and about 94 cents on Polymarket. A tracker that averaged the two at 1:30 p.m. Eastern put the blend near 93, with the Republican side around 7 cents. Earlier that morning the same markets were a cent or so lower. The direction has been stable. The cents have not been frozen.

A 93-cent price is the market saying the favorite should be wrong about 7 times in 100, if the price is an honest probability. That is a heavy favorite. It is also a number that loses often enough that “the House is over” is the wrong sentence. Control is settled by seats, not by the national popular vote. A contract named “who wins the House” does not pay on who wins the most votes nationwide. Read that sentence twice if you trade these.

The Senate is the market that can still move

Senate control, the same day, was in the mid-60s for Democrats: about 62.5 cents on Kalshi early in the morning, about 64.5 on Polymarket, with PredictIt in the same neighborhood. Another board had Polymarket at 64.5 around noon UTC. Call it the mid-60s, and check the live book before you do anything with it.

Republicans held a 53-47 Senate going into the fall, counting the two independents who caucus with Democrats. A Democratic majority from there is a net gain of four seats. That is why the Senate contract is worth less than the House contract even when the national environment is the same. The map is not the generic ballot.

The alternative people are actually pricing is a Democratic House with a Republican Senate, not a Republican House. One October 3 balance-of-power board had a Democratic sweep of both chambers in the mid-60s and a split (Republican Senate, Democratic House) in the high 20s. A Republican House was a single-digit price. A Democratic Senate paired with a Republican House was nearly a rounding error.

Where the cents are still near a coin flip

Most Senate and governor favorites were not close. One daily scorecard on the morning of October 3 had the favorite at 90 cents or more in 46 of 67 Senate and governor races it tracks. The test of the market is the short list underneath that.

On Polymarket around noon UTC on October 3, the closest Senate boards looked like this. These are yes-prices for a party, not a poll:

• Iowa. Republicans about 56.5 cents, Democrats about 43. A morning snapshot had Republican Ashley Hinson a small favorite on Kalshi, Polymarket and PredictIt, in the mid-50s.
• Maine. Democrats about 58.5 cents. That price had come in over the prior week. A one-point move in a poll moves a market like this. So does a quiet day.
• Ohio. Democrats about 59.5 cents.
• Texas. Democrats about 64.5 cents. Reports through the week had the race between Democrat James Talarico and Republican Ken Paxton, with the traded price in the same mid-60s area. Texas is in the price of Senate control. It is not the whole price.

Governor markets had their own coin flips. Nevada and Georgia were both being quoted with the Republican a slight favorite, low-to-mid 50s, on the morning scorecard. A 54-cent favorite is a market that has not made up its mind. Treat it that way.

If those prices are honest, you should expect some favorites to lose. The same scorecard added up the gaps and got on the order of eight upsets across the 67 races. Zero surprises on November 3 would mean the markets were too timid, not that they were sharp. Races also move together when the national mood moves, so one night can look brilliant or terrible by luck. Seventy races is not a huge sample.

How to read the contract before you read the price

• Control is not the popular vote. Seat contracts and vote-margin contracts are both listed. They settle on different facts.
• Exchanges do not share one rulebook. Kalshi and Polymarket each write their own text for when the House is “decided,” including delays and disputes. A 93 on one site and a 94 on the other can be the same view or a slightly different question.
• A cent is a probability, not a poll lead. Going from 60 to 64 is a real move. It is not “up four points” in the polling sense. The cents note walks through the arithmetic with this weekend’s baseball prices, which are easier to see.
• Recounts take time. Settlement can lag the television call. If you need the contract to pay on election night, the rule you want is the one that says so. Most of them do not.

Frequently asked

Is a 93-cent House price a sure thing? No. It is a price that should be wrong several times out of a hundred. The Senate price, in the mid-60s, should be wrong more often than that.

Which races decide the Senate? Whichever ones are still near 50 in the last week. On October 3 that list included Iowa, Maine, Ohio and Texas, plus a couple of governor races if you are looking past Congress. The list will not be identical on November 1.

What does Polypick do with this? The same thing it does with a game. You screenshot one market. The analyzer reads that market. It does not call the election, it does not prefer a party, and a lean is not a promise. If the screenshot is the House contract, do not expect a paragraph about Iowa.