The outcome of Wisconsin's Democratic gubernatorial primary delivered a setback not only to the progressive wing of the party but also to Polymarket and Kalshi, two prediction market platforms that had cultivated a near-infallible reputation after correctly forecasting the results of every state in the 2024 presidential election.

Event contracts on both Polymarket and Kalshi tracking the Wisconsin governor's primary had aligned closely with traditional polling, which anticipated a comfortable victory for progressive candidate Francesca Hong. Each platform assigned Hong approximately a 95% probability of winning. When the votes were tallied, however, her opponent David Crowley prevailed by a margin of less than half a percentage point.

Following the upset, Polymarket deleted a post on X in which it had placed Hong's chances of winning the primary at 96%, despite her lack of endorsements from prominent Democratic socialists such as Bernie Sanders and Alexandria Ocasio-Cortez.

Christopher Hale noted the removal on X: https://x.com/ChristopherHale/status/2087532780280303721

They deleted this tweet. pic.twitter.com/BoavMCpSn1 — Christopher Hale (@ChristopherHale) August 12, 2026

In response to questions about the market's accuracy, Kalshi's leadership pushed back against the criticism. CEO Tarek Mansour said in a social media post that the 95% probability assigned to Hong still implied roughly a one-in-20 chance for her opponent—a low-probability scenario that the market had explicitly accounted for. Kalshi founder Luana Lopes Lara struck a similar tone, writing: "5% is not 0%."

The miscalculation arrives at a time when prediction markets are increasingly viewed as a complement to traditional polling for forecasting elections and other events. Their credibility climbed sharply during the final stretch of the 2024 presidential campaign and its aftermath, when Polymarket and Kalshi were widely credited with giving Donald Trump an edge that many conventional polls had portrayed as a tossup. Their reach has since extended into mainstream media: CNN named Kalshi its official prediction-markets partner, and Dow Jones announced plans to integrate Polymarket data across The Wall Street Journal, Barron's, and MarketWatch. Kalshi operates as a federally regulated exchange under the Commodity Futures Trading Commission, while Polymarket, which runs on blockchain infrastructure, agreed to a CFTC settlement in 2022 and has restricted access for U.S. users in certain markets—differences in oversight that have shaped how each platform is received by institutions and regulators alike.

Crowley's victory, however, was not the first instance of prediction markets mispricing a political contest, although earlier misses involved probabilities that allowed greater room for an upset. In June, Kalshi and Polymarket assigned reality-television personality Spencer Pratt roughly a 75% chance of advancing from Los Angeles's nonpartisan mayoral primary to the general election; he ultimately finished third. The prior month, Kalshi gave Rep. Thomas Massie a comparable probability of winning the Republican primary in Kentucky's 4th Congressional District. Massie lost to Trump-backed challenger Ed Gallrein.

For some observers, this week's result served as a reminder that prediction markets are not foolproof replacements for polling. Statistician and political forecaster Nate Silver reinforced that view in a recent social media post.

"I think prediction markets are cool. But people should stop treating them as magic, and I don't think they're a good substitute for polls, or belong in models," he said.

This story was originally featured on Fortune.com.