A new study asserted that Polymarket’s five-minute Bitcoin contract became a wealth transfer machine. It moved money from retail bettors to a small group of manipulators, and made the spot price of Bitcoin worse in the process.
The paper, “Settlement Manipulation in Prediction Markets” by David Dai, Ruizhe Jia, and Shihao Yu of Stanford and Singapore Management University, studied a product that did not exist before February 12, 2026.
That day Polymarket launched a binary contract that paid $1 if Bitcoin closed a five-minute window above where it opened, and $0 otherwise. A new contract is opened every five minutes around the clock.
In a few months, Polymarket’s five- and fifteen-minute crypto-up/down markets traded over $4 billion and tripled the platform’s daily volume. The mistake in the polymarket was when the contract settled against the Chainlink oracle that estimates the value of Bitcoin in a large exchange of places.
A trader holding the contract could buy or sell real Bitcoin in the closing seconds, drag that reference price to the strike, and win the bet.
The combination of the exchange oracle seemed like a safeguard, because moving it seemed to require moving many places at once. The authors indicated that it was not defensive. Binance, the largest crypto exchange, sits about two and a half points from oracle and came close to one-for-one with it. It ended up on the same side of the strike as the solution about 85% of the time. The push that drove the price of Binance a few basis points past the strike carried the result.
The pattern was in the Binance data. After the five-minute contract went live, the net order in the last ten seconds before each close jumped nearly 50% above the pre-launch level. The spike was sharpest when the push was most important: in 6% cycles the market judged close-even, the jump was almost 3.9 times that of the rest.
The retreat was lost. Real information resides in value; manipulative push does not. Within ten seconds the price is back, about a quarter in roughly equal cycles. The pressure is concentrated in the small hours, when the dollar flow moves the value the most: 56% arrived overnight and 44% on weekends.
Who won, who paid with these Polymarket badges
In nearly even cycles, a push to the favored side turned into a winner 65% of the time, versus 41% in the normal trade. Even when one side held a 90-to-100% chance before closing, pushing against it reversed the outcome 34% of the time, versus 1% in cycles without pushing. Betting is a market that is considered to be almost a loss one out of three times.
Because Polymarket sits on a public blockchain, the authors keep track of each bag. Only 821 traders matched the fake profile, about one in three hundred out of 243,000 who traded the contract. They took in $8.2 million in push cycles and broke even on some. Of the losses, 93% came down to stores.
The authors ruled out hedging as an innocent explanation. The binary contract had less exposure to hedging when one side closed, however those were reversed cycles. And the trade came in one burst in the last fifty seconds, not as a position built on a window.
The solution
The adjustment was the contract horizon. Fraud was not in the fifteen-minute contract, because the long window took normal trading before the close and made the concentrated push a weak force. The stakes were high in crypto last time: Nasdaq and Cboe each filed with the SEC to list binary asset price contracts on indices, which have similar risk to capital markets.
