2026 Governor Race Odds: 36 Boards vs One Ladder

Published 23 October 2026 · Prices frozen 23 October 2026, 21:15 UTC

Every page quoting 2026 governor race odds quotes the state boards, and the state boards are not the contract most people think they are reading. Thirty-six governorships are on the ballot on 3 November. Both Polymarket and Kalshi also sell a single number for the whole map — how many Republican governors there will be when it is over — and on this evening’s prices that number disagrees with the states it is made of.

The gap matters because of which way it cuts. The aggregate ladders are the cheap, thin, easily quoted contracts; the state boards carry seventy times the money. When the two part company it is not a difference of opinion about politics. It is a thin market that has not finished reading a thick one, and this week the state boards gave it a great deal to read.

We priced all seven rungs of Polymarket’s ladder and all eight of Kalshi’s, pulled the Republican side of every one of the 36 contested seats, added the 14 governorships not on any ballot, and multiplied the whole thing out to see whether either aggregate survives its own constituents. All prices are frozen at 23 October 2026, 21:15 UTC; the live boards are linked throughout.

Where the 2026 governor race odds stand tonight

Polymarket’s How many Republican Governors after the 2026 midterm elections? board is a seven-rung ladder over the 50-state total. Republicans go in holding 26 of the 50, so every rung below 26 is a net loss and the modal rung is a net loss of three or four.

Two things jump out before any analysis. The money is in the wrong place: 78.9% of the board’s $708,404 sits on the 28–29 rung, which costs 2.35¢ and would need Republicans to gain two or three governorships on a map where the market prices them losing four. And the book sums to 96.30¢, so the rungs are collectively cheap by nearly four cents before you argue with any of them.

Normalise that 96.30¢ book and the ladder’s implied mean is 23.30 Republican governors, with 57.6% on 23 or fewer and 14.3% on Republicans holding 26 or more — that is, keeping the majority they have now. Hold that mean in mind. The rest of this piece is about the fact that it is wrong by the market’s own lights.

Add the 36 states up and you get a different answer

A count is a sum of the individual races. Polymarket runs a board for 33 of the 36 contested seats — Gamma lists none for Arkansas, Connecticut or Wyoming, so those take Kalshi’s mid. Sum the Republican price across all 36, add the 8 Republican governorships not on any ballot, and you have the same quantity the ladder prices.

14.54 expected wins out of 36, plus the 8 Republican seats not on the ballot, is 22.54 Republican governors. The ladder says 23.30. The state boards are 0.76 of a governorship more pessimistic about Republicans than the contract built to summarise them.

Three quarters of a governorship, and which side to trust

A 0.76 gap invites the obvious objection: the ladder quotes seven to eight points wide on its two central rungs, so maybe the mid is just noise. It is worth testing rather than asserting. Price the whole ladder at every bid and it implies 23.17; price it at every ask and it implies 23.40. The state-board sum of 22.54 sits below that entire range. The gap survives the spread.

The second objection is better: a sum of 36 marginal probabilities is an expected value, and the ladder prices buckets. A distribution with the right mean can still disagree with a bucket. But the direction is the point, and the volume says which side is doing the work.

On Polymarket the state boards are 71 times the ladder. On Kalshi the state markets this article compares are 1,742 times the entire eight-rung count ladder. One caveat on the Polymarket figure: California alone is $42.2m of that $50.2m, and most of it sits on candidates who lost in June — we took that board apart in California governor odds 2026. Strip California out and the other 32 boards still carry $8.0m, more than eleven times the ladder.

Kalshi runs the same market twice and disagrees with itself

Kalshi lists two contracts over the governorship count. KXRGOVCOUNT is a cumulative ladder of “at least N” rungs, filed under Elections with a CFTC product certification. KXGOVWINS is a three-leg board asking which party holds more governorships, which is the same question with the strike fixed at 26. They do not agree.

Four points, inside one exchange, on a binary that either happens or does not. The deeper board is the more Republican one: KXGOVWINS has 269,069 contracts across its three legs and 6,530 traded in the last 24 hours, against 10,575 lifetime on all eight rungs of the count ladder. Its Democratic leg alone — 236,118 contracts — is twenty-two times the whole ladder it contradicts.

Converted into buckets, Kalshi’s ladder is also far more bearish on Republicans than Polymarket’s — the opposite of what its own KXGOVWINS price implies. Both are true at once because the ladder is the board nobody is watching.

Eleven and a half points on the headline question of the cycle. Kalshi’s ladder implies a mean of 22.91 Republican governors, Polymarket’s 23.30, and the state boards 22.54. Kalshi’s thin aggregate happens to sit closer to the state-level consensus than Polymarket’s does — which is luck, not information, on 303 contracts of 24-hour volume.

On the states themselves the two exchanges agree to a point

This is the part that settles it. If the two venues genuinely disagreed about the 2026 governor map, the disagreement would show up in the races. It does not. Seventeen of the 36 seats are quoted on both exchanges with enough depth to compare, and the mean absolute gap between them is 0.96 points.

Sixteen of the 17 gaps run the same way — Kalshi a shade more Republican — and the bias is tiny and consistent, which is what a different fee and spread structure looks like. Note also that Kalshi’s New York market is its deepest governor contract at 3.44m contracts, on a race it prices at 5.35¢ for the Republican. Depth and competitiveness are not the same thing.

So the exchanges agree about the states and disagree about the sum of the states by up to eleven points. That is not a political disagreement. It is two aggregate contracts with almost nothing behind them.

The week the states moved and the ladders did not

The last seven days are the cleanest demonstration available. Six Republican state prices fell by four points or more and only one rose by as much.

Put those through the sum and the state boards went from 23.14 implied Republican governors a week ago to 22.54 tonight — a fall of 0.60. Over the same seven days the Polymarket ladder’s implied mean went from 23.37 to 23.30, a fall of 0.07. The gap between the two widened from 0.23 to 0.76. Over 30 days the ladder moved 0.31 against the states’ 0.62: about half as far, both times.

Two moves are worth naming. Nevada is the only incumbent governor the market now expects to lose: Joe Lombardo was a coin flip at 50.5¢ a month ago. Ohio has gone 38.5¢ to 18.5¢ over the same month, a toss-up becoming a likely Democratic pickup in a state Donald Trump won twice. Neither re-price has reached the ladder in any size.

One governor has already lost, and it counts

Kalshi also sells a ladder on how many governors lose re-election in 2026, and at first glance it looks incoherent. Its modal rung is exactly two defeats at 37¢, with 75.9% on two or more — yet only one incumbent on any general-election ballot is priced below even money on either exchange.

The rulebook resolves it, and it is the kind of clause that makes a price look wrong when it is right. Kalshi counts a defeat if a governor “actively seek[s] re-election” and fails to win, and that “includes losses in primaries”. Dan McKee of Rhode Island lost the Democratic primary to Helena Foulkes on 9 September, the first sitting governor to lose a party primary since 2018. One defeat is already banked. The rung for exactly one is therefore the rung that says nobody else loses, and at 21.5¢ it is priced accordingly.

That clause also decides a case it might have caught. Tim Walz sought a third term in Minnesota and withdrew in January; Kalshi explicitly excludes governors who “withdraw before primaries”, so he does not count, and the seat is now an Amy Klobuchar–Lisa Demuth race the market prices at 4.35¢ for the Republican.

Work the same question from the state boards. Seventeen incumbents are still on a ballot; sum their implied defeat probabilities and you get 1.92, dominated by Lombardo’s 67.5%, with Greg Abbott at 27.4% and Phil Scott at 25.5% next. Add McKee and the state boards imply 2.92 incumbent defeats against the Kalshi ladder’s 2.34. Same direction, same magnitude of error: the aggregate is less dramatic than its own constituents, by roughly 0.6 of a seat, for the second time in this article.

What would actually move these odds before 3 November

Georgia resolving late. Georgia requires a majority, and Polymarket lists a separate board on whether anyone wins the governorship outright in the first round. It is effectively unpriced — 47 bid against a 96.2 ask on $4,631 of lifetime volume, which is not a probability — but the existence of a runoff path matters, because both aggregate rulebooks keep the count contract open until a decisive runoff is called. A 39¢ Georgia that goes to a second round can leave the whole ladder unresolved for weeks.

Alaska, which is not a two-party race. Alaska runs a nonpartisan top-four primary and a ranked-choice general election, and the market has Democrat Jonathan Kreiss-Tomkins at 84.5¢ for an open Republican-held seat in a state Trump carried. A single redistribution of later-round votes decides one of the four Republican losses the ladder is built on.

Correlation, which none of this arithmetic captures. Kalshi sells a board on Democrats sweeping Pennsylvania, Michigan, Wisconsin, Georgia, Arizona and Nevada, quoted 40/44. Multiply Polymarket’s six Democratic prices and you get 28.1% — but those outcomes are not independent, and a national swing moves all six together, so the true joint probability belongs above the product. The 42¢ sweep price is the one number here that argues an aggregate can do something the state boards cannot.

Nothing else, probably. That is the uncomfortable implication of the volume table. The contract that answers “how many governorships will Republicans hold” is, on both exchanges, among the thinnest things either of them lists, and it has spent the past week not reacting to a fifteen-point move in Nevada. The sharper question before 3 November is not what the ladder says. It is why anyone quotes it when the 36 boards underneath it are seventy times the size and say something different.

Related analysis

• How many Republican governors after the 2026 midterms? — the live 7-rung ladder
• 2026 midterms hub — every Senate and governor race we track
• 2026 midterms odds: the Senate flipped in 30 days
• 2026 House odds: markets 90¢, models 76-93%
• California governor odds 2026: Becerra 96¢, 43 dead rungs
• Texas governor election winner — live board

Sources: the Polymarket Gamma and CLOB public APIs (event 101856 plus the 33 state boards named in this page’s source header) and the Kalshi markets and series APIs (KXRGOVCOUNT, KXGOVWINS, KXLOSEREELECTIONGOV, KXDEMSWINGSTATEGOVSWEEP and the GOVPARTY per-state series), all read at 23 October 2026, 21:15 UTC. Resolution criteria are quoted from the descriptions Gamma publishes and from Kalshi’s own rules_primary and rules_secondary fields and its product certifications. The 26–24 partisan split, the 18–18 class on the ballot and the 3 November 2026 election date are from Ballotpedia. The 9 September Rhode Island primary result is as reported by PBS NewsHour, and Alaska’s top-four primary and ranked-choice general are as described here. Our methodology page explains how we read the two venues. Prices here are a frozen snapshot; the live numbers are on the market pages. Nothing here is financial or betting advice.