Polymarket poll trades draw internal investigation

Prediction markets sold themselves on the midterms as fast, accurate and open. Now a set of Polymarket poll trades is raising the oldest question in finance: did someone know first? Researchers at Vanderbilt University flagged bets placed shortly before two New York Times/Siena polls, the Times reported on November 6, and Polymarket says it is investigating.

What happened

In the early hours of October 31, one account bet $191,277 that Nebraska Republican Senator Pete Ricketts would win re-election. Roughly an hour after that, at 5:05 a.m. ET, a Times/Siena survey put Ricketts six points ahead of independent Dan Osborn. Polymarket odds on Ricketts moved from about 67% before the bets to roughly 73% after them and 79% once the poll was public, and the trader made more than $77,000.

The same account put $622,608 on the Democratic nominee in Michigan's Senate race in the same period. A Times/Siena poll on November 1 showed Democrat Abdul El-Sayed leading, and that position returned more than $180,000. Crypto Briefing put total profits above $257,000 and said more than $2 million had arrived in an anonymous wallet around 5:30 p.m. Eastern on October 30 before the betting began. The user also moved large sums to Kalshi, where similar bets followed. Polymarket said it had spotted the trades and opened its own review.

Brett Goldstein of the Vanderbilt Institute of National Security said the activity "requires further investigation."

Why it matters

The researchers did not allege wrongdoing. They noted the trader might have predicted poll timing without knowing results, acted on personal views or used offsetting positions. But scheduled media polls are seen in advance by a small group of pollsters and editors, which makes Polymarket poll trades of this kind a test of how well venues police nonpublic information. On-chain records cut both ways here: the wallet's deposits and the timing of each bet can be traced precisely, but the person controlling it cannot, Crypto Briefing noted. A move from 67% to 79% on one scheduled release also shows how much a single data point can shift a political contract.

The case adds to pressure that was already building in Washington, from House Oversight letters on insider trading to the April charges against a U.S. soldier accused of profiting from Maduro-related contracts. It also complicates the industry's victory lap after Kalshi's midterm accuracy claims. Kalshi declined to comment to the Times.

What's next

Polymarket has not announced conclusions or penalties. Watch for any referral to authorities, account restrictions or new rules on trading around scheduled data and poll releases. Kalshi, which received transfers from the same user, has not said whether it is reviewing the activity, and pollsters may face questions about who sees results before publication.

The Nebraska market at the center of the Polymarket poll trades is archived on Polymarket.