Analysts from Stanford University and the Singapore University of Management presented a study that revealed systematic manipulation in the five-minute Bitcoin contract market on the Polymarket platform. After analyzing over 60 million on-chain records, order book data, and high-frequency spot flows, experts concluded that a group of 821 wallets (just 0.34% of 243,000 active addresses) earned $8.22 million from anomalous cycles, while retail investors lost $7.61 million.
Mechanism of Alleged Manipulation
The five-minute contract, launched on February 12, 2026, pays $1 if the Bitcoin price at the time of settlement is not lower than the initial value, and $0 otherwise. The Chainlink price feed is used to determine the outcome. Researchers reconstructed the scheme: a trader buys a contract for an increase or decrease, and then, seconds before the window closes, executes a large spot trade, temporarily moving the market in the desired direction. After settlement, quotes partially recover.
A key indicator is a sharp increase in directional order flow in the last 10 seconds of the window, which rose by approximately 50% after the contract's launch. In cycles with near-equal probability of outcomes, the activity surge was 3.9 times higher than normal, and the average spot trade volume at such moments reached $1.7 million compared to $68,000 in regular cycles. Moreover, 56% of such episodes occurred during nighttime hours (UTC) and 44% on weekends, when liquidity is lower, making manipulation easier.
Retail Trader Losses and Platform Response
Particularly telling were cycles where Polymarket estimated the probability of one side winning at 90–100%. In 34.2% of manipulated cycles, directional movement against the favorite led to its defeat, whereas without anomalies, this occurred in only 1% of cases. In more balanced cycles (50–60% probability), manipulation changed the outcome in 65.4% of cases compared to 41.4%.
Representatives of Polymarket have already announced plans to transition some markets to a mechanism that considers prices over a longer period rather than a single point in time. In fifteen-minute contracts, according to the study, the activity surge was significantly weaker, and the price reversal was not pronounced. Other possible solutions include settlement based on the average price, random selection of the fixing moment, position limits, and additional costs for transactions before closing.
My comment: This study is a wake-up call for the entire prediction market industry. The problem lies not in the tool itself, but in the vulnerability of oracle mechanisms, especially with short time windows. Polymarket is responding correctly, but investors should remember: "the wisdom of the crowd" only works under conditions of equal access to information and the absence of technical opportunities for manipulation. As long as such loopholes exist, retail traders will play by rules written not in their favor.