Study: Polymarket Accuracy Driven by 3% of Traders, Not 'Wisdom of the Crowd'

Researchers from London Business School and Yale University have analyzed Polymarket transactions from 2023 to 2025 and reached a striking conclusion: the platform's prediction accuracy is driven by a small cohort of skilled traders — an "informed minority" — rather than the collective wisdom of the crowd.

Only 3% of Users Shape the Market

The scope of the study is substantial: scientists examined data from 1.7 million accounts and trades totaling $13.76 billion. Their findings reveal that so-called "skilled winners" represent just 3.14% of all users. This group, combined with market makers, captures more than 30% of total profits on the platform.

High overall earnings, however, do not necessarily indicate genuine trading skill. Further analysis showed that only 12% of the most profitable accounts actually employ a deliberate strategy. Roughly 60% of seemingly "lucky users" began incurring losses when market conditions shifted. The majority of participants — 67% — remain in the red due to inexperience or simple bad luck.

Why This Matters

These findings directly challenge the narrative promoted by the leadership of both Polymarket and Kalshi. Executives at both companies have repeatedly argued that prediction markets outperform expert analysis thanks to the collective intelligence of their broad user bases. The new research undermines that claim: pricing mechanisms are effectively shaped by a narrow band of professionals, not the wider audience.

For regulators and the platforms themselves, this distinction is critical. If prediction accuracy hinges on a few thousand skilled traders rather than millions of participants, the "wisdom of the crowd" argument loses much of its persuasive power in debates over the legitimacy of prediction markets.

Signs of Insider Trading

Beyond skill analysis, the researchers identified suspicious activity consistent with potential insider trading. They flagged 1,950 accounts that placed trades immediately before major events and ceased activity right after those events concluded.

One notable example involved bets on the resignation of former Venezuelan President Nicolás Maduro. The trades were executed shortly before the United States officially announced a military operation dubbed "Absolute Resolve," aimed at capturing the political leader.

Insider betting on Polymarket has already drawn law enforcement attention. In April, the U.S. Department of Justice charged active-duty service member Gannon Ken Van Dyke with allegedly using classified information to place bets on the platform.

What Comes Next

The study's conclusions could intensify regulatory scrutiny of prediction markets. Evidence of insider trading combined with the dominance of a small professional trader class raises questions about whether the competitive environment is fair for average participants. Polymarket and Kalshi now face the challenge of responding to well-founded academic criticism that strikes at the very foundation of their business model.