Kalshi says its 2026 U.S. midterm markets generated more than $2.5 billion in trading volume, nearly twice the $1.3 billion it reports for the 2024 election cycle. But the figures cover different elections, and the platform’s tally of winning favorites contains unexplained discrepancies.
- Reported volume: More than $2.5 billion for 2026, versus $1.3 billion for 2024.
- Not a like-for-like comparison: The 2024 cycle included a presidential election.
- Favorite tally: Kalshi says favorites won 479 of 487 races counted.
- Unresolved questions: Several category percentages do not match the counts, and the available details are not enough to independently assess the performance analysis.
Kalshi’s volume comparison needs context
Kalshi reports more than $2.5 billion in trading volume across its 2026 U.S. midterm election markets, compared with $1.3 billion during the 2024 election cycle. The rounded figures imply an increase of roughly 92% or more, a difference of more than $1.2 billion.
The comparison is not like-for-like. The 2024 cycle included a presidential election, while the 2026 markets covered congressional, state and local contests. Kalshi has not provided a comparable, market-by-market breakdown here showing how much activity came from equivalent races.
Trading volume measures the value of trades over a period. The same capital can be traded repeatedly, so volume does not show how much money is at risk at one time or how many distinct traders took part. Open interest, by contrast, measures outstanding positions. A pre-election figure cited by Dune put combined midterm open interest on Kalshi and Polymarket at $250 million as of Oct. 18, with $210 million on Kalshi and $40 million on Polymarket. That is a different measure and should not be compared directly with cumulative trading volume.
What does a 479-of-487 favorite tally show?
In its initial tally, Kalshi said market favorites won 479 of 487 races resolved by 6 a.m. ET on Nov. 4, 2026. That works out to about 98.4%, in line with the company’s rounded figure of 98%. The tally included House, Senate and gubernatorial races, along with contests for attorney general, lieutenant governor and secretary of state. Some races, including the Alaska and Mississippi Senate contests, remained unresolved at the cutoff.
The category figures do not add up. The counts Kalshi lists imply favorite win rates of about 98.9% in House races (356 of 360), 97.0% in Senate races (32 of 33) and 97.1% in gubernatorial races (33 of 34). Its reported table gives those categories rates of 95%, 94% and 94%, respectively.
If the counts and percentages cover the same races, cutoff and definition of “favorite, ” they conflict. The combined 479-of-487 calculation is internally consistent, but Kalshi needs to explain the category discrepancy before those rates can be relied on.
Even an accurate favorite win rate tells only part of the story. It counts a candidate priced at 51% as a correct favorite if they win, just as it does a candidate priced at 95%. The tally alone does not show whether the probabilities were well calibrated or how markets compared with polls and other forecasts.
Market odds are projections, not race calls
Kalshi’s election-night recap says its market for Democratic control of the House reached a 99% probability at 8:10 p.m. ET. Its Senate-control market crossed the same threshold at 9 p.m. Live updates also showed individual candidates reaching 99% probabilities at points during the night.
Those were market-implied probabilities, not official race calls or confirmation that contracts had settled. A contract price can indicate the probability traders assign to an outcome, but fees, spreads and liquidity can affect how closely it tracks a straightforward probability. A 99% reading is a strong market signal, not a guarantee.
Kalshi Research separately reported that directional accuracy rose from 93.8% three months before voting to 98.2% at poll closure, while Brier scores fell from 0.040 to 0.016. Directional accuracy measures whether a forecast picked the winner. A Brier score compares forecast probabilities with actual outcomes, and lower scores generally indicate better performance.
The analysis is attributed to Nicole Kagan and Rubens Baiocchi, but its full methodology and underlying data are not available here for assessment. The figures should be treated as results reported by Kalshi Research, not as an independently reproduced evaluation. Details such as the sample, scoring choices and exact forecast snapshots matter when interpreting a change in accuracy.
Key questions about Kalshi’s midterm figures
-
How much volume did Kalshi report for its 2026 midterm markets?
More than $2.5 billion, compared with $1.3 billion for the 2024 election cycle, according to Kalshi.
-
Does that establish a like-for-like doubling of activity?
No. The rounded figures imply growth of roughly 92% or more, but 2024 included a presidential election and the 2026 markets covered different contests.
-
Did favorites win about 98% of the races in Kalshi’s tally?
Kalshi reported 479 wins in 487 races, or about 98.4%. Several category percentages in its table do not match the listed counts.
-
Does the favorite win rate prove prediction markets beat polls?
No. It does not account for the probabilities assigned to each candidate or establish a like-for-like comparison with polls.
-
Were the 99% election-night readings official results?
No. They were market projections, separate from official race calls and contract settlement.
Big volume is not a performance audit
Prediction markets can provide a constantly updated view of what traders expect. That information is useful, especially when incoming results shift the balance of a race. But high volume measures activity, not forecasting skill.
Kalshi’s figures point to substantial reported trading and a high rate of winning favorites. To judge what they say about market accuracy, readers also need comparable election-cycle data, a reconciled race tally and transparent methods for scoring forecasts. Markets can be informative without being infallible, and their claims deserve scrutiny like anyone else’s. Their legal status deserves scrutiny too.
