Study suggests that a small group of skilled traders primarily drive Polymarket’s price accuracy

A new academic paper questions the assumption of crowd wisdom in prediction markets, revealing that a minority of skilled traders predominantly influence price discovery on Polymarket, raising conc...

A new academic paper has cast doubt on the popular idea that prediction markets are accurate because of broad crowd wisdom, arguing instead that most of Polymarket’s price discovery comes from a tiny cluster of skilled traders. The study, published on SSRN on April 20 and revised on April 25, examined 98,906 events, 210,322 markets and $13.76 billion in trading volume across 1.72 million accounts, according to researchers from London Business School and Yale University.

Using a sign-randomisation test, the authors said only 3.14% of accounts qualified as skilled winners. Those traders were not only consistently profitable but also showed persistence: 44% kept their skilled classification in out-of-sample testing, far above the 10% rate the authors cite for skilled mutual funds in a parallel comparison. The paper argues that the platform’s accuracy is driven by a small, identifiable minority rather than the broader pool of participants.

The study also found that skilled traders were the group most likely to react first to scheduled news, including Federal Open Market Committee announcements and corporate earnings releases, adjusting their order imbalance in the direction of the surprise. By contrast, other groups showed no consistent response. The authors said that while some accounts appeared to trade on non-public information, those cases were too concentrated in a handful of events to explain the platform’s wider pricing behaviour.

That point comes against a more fraught regulatory backdrop. On April 23, the Commodity Futures Trading Commission filed a complaint in federal court in New York against Gannon Ken Van Dyke, an active-duty US Army service member, alleging he used classified information to trade a Polymarket contract on whether Nicolás Maduro would be out of power by January 31, 2026. The agency said the trades generated more than $404,000 in profits. The academic paper separately identified 1,950 accounts that met its timing and conviction criteria for suspected insider trading, including three accounts that, it said, earned more than $630,000 around a secret US military operation in early January 2026. Even so, the researchers concluded that such activity was too concentrated to account for price formation across the platform as a whole.

Source Reference Map Inspired by headline at: [1]

Sources by paragraph: - Paragraph 1: [2] - Paragraph 2: [2] - Paragraph 3: [2] - Paragraph 4: [3], [2]

Source: Noah Wire Services