A recent study has examined the dynamics of prediction markets, revealing that a small group of informed traders significantly influences price movements, challenging the notion that market accuracy relies on broad participation. Conducted by researchers from London Business School and Yale, the findings are based on an analysis of Polymarket trades from 2023 to 2025.
The research scrutinized 1.72 million accounts and $13.76 billion in trading volume, concluding that merely 3% of traders are responsible for the majority of price discovery. These informed traders consistently make accurate predictions, while the remaining 97% generally incur losses, providing liquidity but lacking the skill to profit.
To differentiate between skill and luck, the researchers employed a method where they simulated each trader’s bets 10,000 times, altering only the direction of the trades. This approach established a benchmark to assess whether traders’ profits were due to genuine skill or mere chance. The results indicated that only 12% of the top winners surpassed this benchmark, with many initially successful traders reverting to losses upon further evaluation.
The presence of skilled traders enhances market accuracy, particularly in the lead-up to event resolutions. These participants are quick to adjust their positions based on new information, such as Federal Reserve announcements or corporate earnings reports, while the broader trader base tends to react less consistently.
However, the study raises concerns regarding the implications of trading on non-public information. Both Polymarket and Kalshi have stated that such practices violate their regulations. The paper highlights a specific incident involving the U.S. operation to remove Nicolás Maduro from power in Venezuela, where newly created accounts made substantial bets on the outcome shortly before the event, resulting in significant profits. Although there is no evidence of wrongdoing, the incident underscores the risks associated with insider information.
While insider trading can lead to more pronounced price movements, it is rare and typically confined to specific events. The study emphasizes that the ongoing accuracy of prediction markets largely hinges on the consistent performance of informed traders rather than isolated bets.
A study indicates that a small group of informed traders is primarily responsible for price movements in prediction markets, challenging the belief in crowd-based accuracy. The research highlights the risks associated with non-public information and the implications for market integrity.