Polymarket investigates suspicious bets placed hours before NYT/Siena polls

Someone appears to have been very good at guessing what a poll would say before anyone else could read it.

Polymarket says it has identified a series of well-timed trades placed shortly before two New York Times/Siena College polls were made public. The company has opened an internal investigation. The trades were flagged by researchers at Vanderbilt University, whose findings were reported by the New York Times.

The pattern includes a bet of approximately $191,277 on Nebraska Republican Senator Pete Ricketts winning re-election. It was placed on October 31, 2026. A poll released soon after showed Ricketts ahead by six points, and the trader pocketed more than $77,000.

A $2 million deposit and two lucky calls

The sequence of events is what caught the researchers’ attention.

On October 30, around 5:30 p.m. ET, more than $2 million landed in an anonymous crypto wallet. The account then started placing sizable bets on Polymarket. According to the Vanderbilt research, roughly $191,000 of that went toward Ricketts holding his Senate seat.

Before the poll dropped, the market gave Ricketts approximately a 67% chance of winning. After the six-point lead became public, those odds climbed to about 79%. That jump is what turned the position into a profit of more than $77,000.

The Nebraska trade was not a one-off. The same trader also put more than $622,000 behind the Democratic nominee in the Michigan Senate race. After a subsequent NYT/Siena poll was released, that position produced profits exceeding $180,000.

Add it up and the account cleared more than $257,000 across the two races.

What the researchers and platforms are saying

Brett Goldstein, the Vanderbilt researcher involved in the work, was careful not to accuse anyone outright. But he made clear the timing looked off.

“Requires further investigation.”

Goldstein said the activity was hard to dismiss as simple coincidence and deserved a closer look.

Polymarket confirmed it had detected the trades and begun its own review. The company did not announce any conclusions or penalties.

The research also covered Kalshi, the other major prediction market named in the findings. Kalshi declined to comment.

Why polls are such an obvious target

Major media polls are scheduled events. A handful of people, including pollsters, editors and production staff, see results before publication.

The crypto element adds another wrinkle. The funding came through an anonymous wallet, which means the trades are visible on-chain, but the identity behind them is not. Researchers could trace the $2 million deposit and the timing of the bets precisely. They could not say who controlled the money.