Prediction market users successfully bet on a blue wave — with a reported 98% success rate.
By the time the first polls closed on Election Day, online prediction traders — and the platforms themselves — faced a major test as bettors heavily priced in a Democratic sweep.
Tuesday’s consequential midterm elections ended the GOP’s control of Congress, already hanging by a thread.
People were interested in these podcasts
Kalshi’s balance-of-power market had a 73.6% chance of a Democratic sweep at 6 p.m. ET on election night, compared to 75.5% on Polymarket, with both fluctuating in the 60s in the month leading up to it.
With Republicans holding narrow majorities in both chambers, Democrats needed a net gain of three seats to take the House and four to take the Senate.
Instead, Democrats flipped at least 35 House seats and seven Senate seats, according to projections as of Wednesday.
Just over 93% of Kalshi traders and 94.5% of Polymarket traders predicted that Democrats would take the lower chamber.
Democrats were favored to win the Senate too, but by a smaller margin: 81.5% of bets on Kalshi and 75.5% on Polymarket backed them.
The midterm elections provided “yet another testament to the accuracy of prediction markets,” according to Americans for Fair Markets, a Kalshi-backed advocacy group.
In 479 of 487 midterm races, Kalshi traders correctly predicted the winners with a 98% success rate, Americans for Fair Markets said, adding that the prediction market’s traders called the winners of the key Senate races nearly two hours before networks did.
Kalshi said that by 5 p.m., 356 out of 360 House seats and 32 out of 33 Senate seats were picked accurately. Markets for control of the House hit a 99% likelihood at 8:18 p.m., the prediction market said, and markets for control of the Senate passed 99% likelihood at 9:02 p.m.
Total trading volume on Kalshi’s 2026 midterm markets surpassed $1.7 billion as of Tuesday afternoon, according to the platform, compared with $1.3 billion for the 2024 presidential election cycle.
By Wednesday, Kalshi said the number reached over $2.5 billion.
The Anti-Corruption Data Collective, a nonprofit watchdog, counted more than 21,000 congressional midterm markets across the platforms, compared with 463 for the 2024 congressional races.
In the Senate battlegrounds, prediction markets saw Democrats as favorites in most of the closely watched races, with Iowa the closest.
Of all the high-profile races — Texas, Michigan, Maine, Kansas, Ohio and Alaska — all leaned toward Democrats winning on both trading platforms.
Texans had not elected a Democratic senator since 1988, and Kansas voters had not since 1932, meaning bettors favored Democrats in two states with long Republican streaks.
In less predictable races where GOP strongholds were not expected to budge, such as Mississippi’s Senate race, bettors stuck with the Republican candidate until the results became unclear.
Another key race, Iowa, was an outlier, where Republican Rep. Ashley Hinson faced Democrat Josh Turek for the open seat.
The race was priced as a toss-up, and Kalshi traders crisscrossed between whether Mr. Turek or Ms. Hinson would come out on top, while Polymarket gave the Democrat a narrow win.
Americans for Fair Markets said these results should not come as a surprise, dubbing prediction markets “truth machines.”
“Tuesday’s results underscore the importance of prediction markets as a critical economic tool that empowers entrepreneurs, investors, and pundits to make informed and economically sound decisions,” the group said in a statement. “As prediction markets continue to grow, so too will the size and scope of the decisions based on them.”
Contact the author
Please read our comment policy before commenting.