Stanford University and Singapore Management University researchers have identified significant vulnerabilities in Polymarket's five-minute Bitcoin prediction contracts. According to the academic study, the short-duration contracts create incentives for sophisticated traders to manipulate spot Bitcoin prices just before contract settlement. The researchers found that settlement-price manipulation has shifted approximately $1.28 million from regular market participants to traders exploiting the settlement process during the period analyzed. The working paper, co-authored with Singapore Management University, examined approximately two months of five-minute Bitcoin bets on Polymarket and found repeated bursts of one-sided trading on the Binance exchange that temporarily moved Bitcoin's price in the final seconds before bets closed, benefiting traders positioned in the same direction. The study establishes statistically significant trading patterns around Polymarket's five-minute Bitcoin contracts that are consistent with short-lived attempts to move Bitcoin prices immediately before settlement, though it does not establish intent or prove that the traders influencing Binance prices were the same participants profiting on Polymarket.
The research reveals a fundamental structural problem in financial prediction markets that differs from traditional prediction markets. Financial prediction markets fundamentally change the incentive structure because sophisticated participants can simultaneously trade both the derivative-like prediction contract and the underlying asset, creating opportunities for economically rational manipulation. Settlement becomes part of the trading strategy rather than merely the conclusion of it, with the apparent effectiveness of brief price dislocations in five-minute markets illustrating how point-in-time settlement creates opportunities for manipulation, particularly during periods of thin liquidity such as overnight sessions and weekends. The researchers describe the pattern as a "transitory push to manipulate the spot price" and argue that bets tied to financial assets are vulnerable to manipulation because participants can trade the very asset that determines whether they win or lose. This contrasts with traditional prediction markets used for elections and sporting events, where traders cannot easily influence the outcome.
The Stanford researchers propose several solutions to reduce manipulation risks in prediction markets. According to the study, extending contract duration from five minutes to 15 minutes largely removed the abnormal trading behavior observed during the analysis period. The researchers also recommend alternative settlement methods, including time-weighted average prices, as potential ways to make future contracts more resilient. Polymarket has acknowledged these concerns and is looking to transition certain markets in the next year to settlement methods that use prices over a longer period rather than a single point in time, with a company spokesperson stating the change would "further ensure market integrity." The company currently uses multiple independent pricing oracles to aggregate data and ensure accuracy. Similarly, Cboe's emphasis on broad index-based contracts highlights how diversified reference assets are structurally more resistant to targeted price influence than single-asset benchmarks. The findings come as exchanges expand prediction markets tied to financial assets, with Cboe beginning to roll out products tied to stock indexes and Nasdaq seeking approval for similar contracts, potentially extending the questions raised by the paper beyond crypto and Polymarket.
Despite the manipulation concerns, prediction markets continue attracting record trading activity. According to DefiLlama data, Kalshi processed approximately $9.4 billion in trading volume during June, while Polymarket International recorded roughly $4.3 billion over the same period. Much of this activity came from markets tied to the expanded 2026 FIFA World Cup, with combined trading volume on Polymarket and Kalshi reaching more than $5.4 billion. However, the industry faces increased regulatory scrutiny, with several states challenging operations of companies including Kalshi and Polymarket, while the Commodity Futures Trading Commission maintains exclusive jurisdiction over federally regulated event contracts. The findings arrive at a consequential moment as exchanges including Cboe and Nasdaq expand prediction-market offerings linked to financial assets, with attention likely to move beyond crypto toward broader questions of benchmark construction and regulatory expectations.
