Kalshi vs Polymarket Election Odds: Same Race, Two Prices
The two biggest prediction markets are pricing the 2026 midterms differently, and one of the gaps is a sign flip on who controls the Senate
GhostLeg, 2026-08-08, 6 min
Numbers in this post are a snapshot from Thursday, August 7 at 12:10 PM ET. Divergence recomputes hourly; the live politics board carries the current state of every price cited below.
Two markets are pricing the same November differently
On November 3, Americans vote in the midterms. On Kalshi and Polymarket, traders have been pricing those races for months, and the two venues do not agree.
As of this snapshot, Kalshi prices Republicans to keep the House at 31.5 cents. Polymarket prices the same outcome at 5.9 cents. That is a 25.7 point gap on one of the most liquid political markets in the world. The other side of the same market shows the mirror image: Democrats to take the House at 69.0 on Kalshi versus 91.5 on Polymarket.
Same chamber. Same election day. Wildly different prices. GhostLeg's politics board tracks this gap, and every gap like it, recomputed hourly across 138 races.
The Senate gap is stranger: the venues disagree on the leader
A big gap on one price is interesting. The Senate control market shows something rarer: the two venues currently name different sides as the favorite.
| Outcome | Kalshi | Polymarket |
|---|---|---|
| Democratic Senate control | 54.0 | 45.5 |
| Republican Senate control | 47.0 | 54.5 |
Kalshi's traders have Democratic control as the favorite. Polymarket's traders have Republican control as the favorite. This is not a magnitude disagreement; it is a sign flip on the headline question of the cycle. GhostLeg reports what each venue prices, not who is right; the point is that "the prediction markets say" is not one number right now.
Most gaps are small, which is why the big ones matter
Zoom out from the headlines and the two venues mostly agree. Across every divergence computation GhostLeg has run since live capture began on August 2, the median absolute gap between Kalshi and Polymarket on the same outcome is 2.0 points. The 90th percentile is 5.0 points. The signed average is +0.33 points, which means neither venue is systematically higher than the other.
Small gaps are normal. Large gaps are the story. A few currently on the board:
- House control: 25.7 points, the largest genuine disagreement we track
- 47 Republican Senate seats (a rung on the seat-ladder market): Kalshi 9.5 versus Polymarket 23.5
- Ohio governor: roughly 5 points apart on both party outcomes
- Government shutdown tracker: Kalshi 24.5 versus Polymarket 20.5
Each of those links to a live page where you can watch the gap move. When a gap closes, one side of it was wrong; the accuracy question is one our graded record will answer race by race as they settle.
The biggest number on the board is not a disagreement
Here is the trap every prediction market reader eventually falls into, demonstrated live. As of this snapshot, our board's largest raw gap involves the CLARITY Act, a crypto regulation bill: a Kalshi contract at 90 cents next to a Polymarket contract at 13.5 cents.
Read the contracts, though, and the "disagreement" dissolves. Kalshi's 90 cent contract asks whether the Senate will vote on the bill before January 1, 2027. Polymarket's 13.5 cent contract asks whether the bill will be signed into law during 2026. A bill can easily get a Senate vote by January without becoming law by December. Both prices can be right at the same time, because they are answers to different questions.
This is the single most important habit in reading prediction markets: the price is an answer to the exact question in the contract, not to the vibe of the headline. The venues also settle differently even when the questions match. Kalshi's election contracts generally resolve on official outcomes, in some cases accelerated by media calls; Polymarket's typically resolve on a consensus of major network calls or certification. On a close, contested race, those rules can diverge in timing and, in rare cases, in outcome. Our race pages carry that caveat because it is real.
Why the venues disagree at all
If both markets are liquid and open to informed traders, why would prices differ for months at a time? A few structural reasons:
Different traders. Kalshi is a CFTC-regulated exchange with US customers onboarding through brokerage-style accounts. Polymarket runs on-chain with a globally distributed, crypto-native user base. Different populations bring different information and different biases to the same question.
Different fee and collateral structures. Cost of capital differs across the venues, which changes how aggressively traders arbitrage small gaps. A 2 point gap may simply not be worth closing after costs; a 25 point gap is another matter, and its persistence is itself information.
Different resolution rules. As above; part of any persistent gap can be a rational price on the difference in settlement terms rather than a disagreement about the election.
Different market structure per race. One venue sometimes lists a race as a multi-candidate market while the other lists party-level contracts. GhostLeg's divergence engine compares like for like, summing same-party candidates where needed, so the gaps we publish are on matched outcomes.
We built the measurement layer for exactly this: the politics board shows both venues' prices side by side on every tracked race, the divergence between them, and how it has moved. The election board runs on the same measurement infrastructure that licenses GhostLeg's sports market data through the Data API, and sits alongside the sports coverage on the intel dashboard. For the general comparison of how the two venues work as products, see Kalshi vs Polymarket: prediction markets compared.
How to read the board
Three practical rules, all derived from the data above:
- Treat sub-3-point gaps as noise. That is the normal disagreement band between two healthy markets pricing the same thing.
- When a gap is large, read both contracts before concluding anything. The CLARITY Act example is what a definitional mismatch looks like; the House control gap is what a genuine disagreement looks like.
- Watch persistence, not snapshots. A gap that survives weeks of trading is telling you something structural. Every race page charts the history.
FAQ
Why do Kalshi and Polymarket show different election odds? Different trader populations, different fee structures, and different resolution rules all contribute. Most of the time the venues land within 2 points of each other; the exceptions are where the interesting information lives.
Which venue is more accurate? That is a graded, empirical question, and answering it honestly requires settled races and a methodology, not a hot take. GhostLeg grades every tracked contract against its own venue's settlement; our public record pages show the results as they accumulate, and a full analysis of election market accuracy is coming in this series.
Do the venues resolve the same way? Not always. Kalshi generally resolves on official outcomes with media-call acceleration in some markets; Polymarket typically resolves on major network consensus or certification. On most races this never matters; on a contested one it can.
Where can I see the current gaps? The GhostLeg politics board tracks divergence hourly across every paged race, with per-race pages for the House and Senate control markets and dozens of state races.
GhostLeg reports what prediction markets price. Prices are not predictions or recommendations. For entertainment purposes only. Past performance does not indicate future results.