Big Bets Before Poll Drops: Researchers Flag Suspicious Prediction Market Activity

A Vanderbilt University research team has documented a pattern of large wagers placed on prediction markets just before major political polls become public, raising the prospect that some traders are acting on nonpublic information.

On October 31, a single Polymarket user staked nearly $200,000 on the re-election of Nebraska Senator Pete Ricketts. Roughly an hour later, a New York Times/Siena College poll was released showing Ricketts ahead. The bet paid out more than $77,000. The same account also placed substantial wagers on Kalshi, a rival platform.

“The sequence of events is difficult to dismiss as coincidence,” said Brett Goldstein of the Vanderbilt Institute of National Security. His team has been tracking prediction market activity for months, looking for anomalies that suggest insider access. Polymarket confirmed the trades and said it is conducting an internal review. Kalshi declined to comment.

The episode mirrors a criminal case unsealed in April. U.S. Army Master Sergeant Gannon Ken Van Dyke was charged with pocketing over $400,000 on Polymarket using classified information. Prosecutors allege Van Dyke helped plan “Operation Absolute Resolve,” the January mission that led to the capture of Venezuelan President Nicolás Maduro in Caracas. Before the operation became public, Van Dyke bought “yes” shares on contracts betting Maduro would be ousted by January 31 and that President Trump would invoke the War Powers Act against Venezuela by the same date.

Both cases expose a regulatory gray zone. The Commodity Futures Trading Commission has asserted jurisdiction over some prediction markets, particularly those tied to economically significant events. But whether traditional insider-trading statutes apply to bets on political outcomes remains unsettled. Polymarket and Kalshi operate under different legal frameworks, and neither is registered as a national securities exchange.

Goldstein’s team has identified several other instances where large positions shifted shortly before public data releases. The researchers have shared their findings with federal regulators. As prediction markets grow in volume and visibility, the question of who gets to see the numbers first is moving from academic curiosity to enforcement priority.