A new academic study has found that Polymarket’s five-minute Bitcoin prediction contracts have created incentives for sophisticated traders to manipulate spot prices and profit at the expense of ordinary participants. Summary Stanford researchers link Polymarket’s five-minute Bitcoin markets to settlement-price manipulation. The study estimates about $1.28 million shifted from retail traders to sophisticated participants. Researchers say longer settlement windows and improved pricing methods could reduce manipulation risk. According to researchers from Stanford University and Singapore Management University, the structure of Polymarket’s short-duration Bitcoin markets encourages traders to influence the cryptocurrency’s spot price shortly before contracts settle. Their paper concluded that the issue stems from the way settlement prices are calculated rather than from prediction markets themselves. The researchers examined contracts that ask users to predict whether Bitcoin will finish above or below a fixed price within five minutes. Because settlements rely on Chainlink price feeds based on Bitcoin’s market price at the end of each trading window, traders who hold large positions may have an incentive to push the spot price in a favorable direction just before settlement. Settlement design creates opportunities for manipulation After comparing market activity before and after Polymarket introduced these contracts in July 2024, the researchers identified a clear pattern in Bitcoin trading. According to the study, spot-market order flow increased sharply near settlement, and prices frequently reversed soon afterward, behavior the researchers said is consistent with settlement-price manipulation. The paper estimated that the trading pattern shifted roughly $1.28 million from regular market participants to traders who exploited the settlement process during the period analyzed. Rather than describing prediction markets as fundamentally flawed, the researchers argued that contract design plays the central role in reducing manipulation risks. Among the changes discussed in the study, extending contract duration from five minutes to 15 minutes largely removed the abnormal trading behavior. The researchers also pointed to alternative settlement methods, including time-weighted average prices, as possible ways to make future contracts more resistant to manipulation. Their findings extend beyond cryptocurrency markets. According to the paper, traditional exchanges such as Nasdaq and Cboe have proposed event contracts linked to asset prices, making settlement methodology an increasingly important issue as similar products move into regulated financial markets. Prediction markets continue expanding despite regulatory pressure Even as researchers highlighted weaknesses in contract design, prediction markets have continued to attract record trading activity. According to DefiLlama data, Kalshi processed about $9.4 billion in trading volume during June, while Polymarket International recorded roughly $4.3 billion over the same period. Much of that activity came from markets tied to the expanded 2026 FIFA World Cup. Data from Polymarket and Kalshi showed their World Cup winner contracts had generated more than $5.4 billion in combined trading volume at the time of writing, including about $4.25 billion on Polymarket and roughly $1.2 billion on Kalshi. At the same time, the industry’s rapid growth has drawn increased regulatory attention in the United States. Several states have challenged the operations of companies including Kalshi and Polymarket this year, while the Commodity Futures Trading Commission has maintained that federally regulated event contracts fall under its exclusive jurisdiction rather than state gambling laws. With those legal disputes now moving through the federal court system, legal observers have said conflicting appellate rulings could eventually require the US Supreme Court to determine whether oversight of prediction markets belongs primarily to the states or to the CFTC.
