A new academic study challenges the popular narrative that prediction markets derive accuracy from the wisdom of crowds. Instead, researchers from London Business School and Yale University found that just 3.14% of Polymarket accounts—those classified as skilled traders—are responsible for the vast majority of price discovery.
Key Findings: A Tiny Fraction of Accounts Drive Accuracy
The working paper, titled “The Accuracy of Prediction Markets: Wisdom of Crowds or the Informed Few?” was posted on SSRN on April 20, 2026 and revised on April 25, 2026. It analyzed the complete trading history of Polymarket, the world's largest prediction market by volume, covering 98,906 events, 210,322 markets, and total trading volume of $13.76 billion across 1.72 million accounts.
Using a statistical method called a sign-randomization test to distinguish genuine skill from luck, the researchers classified only 3.14% of accounts as skilled winners. These traders traded an average of 79 markets and consistently built positions aligned with eventual outcomes. The remaining 96% of accounts either broke even by chance or incurred losses.
The order flow of skilled traders was found to significantly predict both subsequent price movements and final market outcomes: a one-percentage-point increase in their net buying share raised the probability of correctly predicting the final outcome by 8 basis points. In contrast, traders who were merely lucky showed no predictive power in either test.
Skill Persistence Far Exceeds Traditional Funds
The study tested skill persistence by randomly splitting events into training and test sets. Among those classified as skilled in the training period, 44% retained that classification in the test set. For the unskilled loss group, 51% remained in the same category. By comparison, in a parallel test using mutual fund data, only 10% of skilled mutual funds maintained their classification.
Skilled traders also reacted fastest to scheduled news events such as Federal Open Market Committee (FOMC) announcements and corporate earnings releases. Only the skilled group adjusted their order imbalances in a narrow window around each announcement to account for the surprise component of the news. Other groups showed no consistent reaction.
Insider Trading Cases and CFTC Lawsuit
The research identified 1,950 accounts that met criteria for timing and conviction suggestive of trading on non-public information. These accounts earned an average of about $15,000 each and caused significant price moves when they traded. One documented case involved three accounts that took positions in contracts related to Venezuelan President Nicolás Maduro hours before a secret U.S. military operation began on January 3, 2026, collectively profiting by over $630,000.
On April 23, 2026, the U.S. Commodity Futures Trading Commission (CFTC) filed a complaint alleging that an active-duty U.S. military service member used one of those accounts to engage in insider trading. Despite these price impacts, the researchers concluded that insider activity is too concentrated on isolated events to explain broad price discovery across the platform.
Market Growth Concentrates Gains in Few Hands
Polymarket's monthly trading volume surged from $3.3 million in December 2023 to $1.98 billion in December 2025—a roughly 600-fold increase in two years. Active accounts expanded from about 1,600 to over 519,000 over the same period. Yet skill concentration remained narrow: market makers and skilled buyers together represented less than 3.5% of accounts but captured over 30% of total profits. Unlucky or unskilled losers made up 67% of all accounts and absorbed the entirety of losses.
The authors conclude that the accuracy of prediction markets reflects the behavior of a small but identifiable group of informed traders, whose participation is the mechanism of price formation. Whether these traders will continue to participate as the platform grows and fees increase remains an open question left for future research.