A study reveals evidence of manipulation in Bitcoin contracts on Polymarket, showing significant profits for a small number of wallets.
A recent study conducted by researchers from Stanford University and Singapore Management University has uncovered evidence of manipulation in five-minute Bitcoin contracts on Polymarket. The findings are detailed in their paper, "Settlement Manipulation in Prediction Markets."
Through an analysis of over 60 million on-chain records, order book snapshots, and high-frequency data, the researchers identified 821 wallets that collectively earned $8.22 million during periods of unusual market activity, while retail investors and other unclassified participants incurred losses of $7.61 million.
Mechanism of the Alleged Manipulation
The five-minute contracts, initiated on February 12, 2026, pay out $1 if the Bitcoin price at the end of the period meets or exceeds the initial value. The outcome is determined by the BTC/USD price feed from Chainlink, with new contracts being created every five minutes.
The study outlines a potential manipulation strategy where traders buy contracts betting on price movements and execute large spot trades just before the settlement period ends. This activity influences the price recorded by the oracle, allowing the traders to profit before prices revert to their previous levels.
After the introduction of these contracts, directed order flow in the final ten seconds of the trading window surged by about 50%, with price reversals occurring shortly after. In instances where the outcome was nearly even, the intensity of trading spikes was noted to be approximately 3.9 times greater than usual.
On average, the volume of trades during these manipulated periods reached $1.7 million, compared to just $68,000 in non-manipulated cycles. The bulk of these episodes occurred during off-peak hours, suggesting that lower liquidity facilitated price manipulation.
Financial Implications for Participants
The research categorized 1,613 cycles as potentially manipulated, representing the top 10% of observations. The identified wallets, which participated in at least five of these cycles, were deemed likely manipulators, earning significantly more in manipulated cycles than in others.
While these wallets accounted for just 0.34% of the total addresses trading the contracts, they captured substantial profits compared to the losses sustained by retail participants, who represented about 70.5% of the total losses.
It's important to note that the study does not definitively link specific wallets to individual traders, nor does it prove intent behind the transactions.
Impact on Odds and Market Dynamics
Interestingly, the research highlights that even when Polymarket assessed the probability of winning outcomes at 90–100%, directed trading against the favored side led to surprising victories for the opposing side in 34.2% of cases. In contrast, such outcomes without manipulation were rare, occurring only 1% of the time.
Despite considering alternative explanations like hedging by market makers, the researchers found the scale and timing of directed trades inconsistent with typical hedging practices.
Future Changes for Polymarket
In response to these findings, a Polymarket representative stated that the platform utilizes multiple independent price oracles and plans to adjust its contracts to consider prices over extended periods rather than a single point in time. The researchers noted that extending the contract duration could diminish the potential for price manipulation.
To enhance market integrity, they suggested implementing measures such as averaging prices over time, randomizing settlement moments, capping position sizes, and imposing penalties for last-minute trades.
