Ninety-seven per cent

Polymarket is a prediction market. Users bet real money on yes-or-no questions about the future, and the price of a ‘yes’ share, between zero and one dollar, moves with the crowd’s estimate of the odds. Through the 2026 Iran war it ran contracts on the things the war turned on: whether the United States would strike, whether Iran’s Supreme Leader would survive, whether a ceasefire would be reached and whether it would hold. Some accounts bet on those questions as though they already knew the answer.

This article is about the accounts that were right too often, and too early. It is not a finding that any of them traded on inside information; none has been charged, and the platform is anonymous by design. It is a record of the bets, their timing, the win rate a nonpartisan group measured across all of them, the referrals the company has since made to the Justice Department, and the letters two members of Congress sent in response. What the same war looked like in the oil futures market, where a separate probe is now open, is the subject of this case’s first article.

01The twelve-minute account

The pattern was there from the war’s first days. When the United States and Israel struck Iran on 28 February 2026, about US$529 million had already traded on Polymarket contracts tied to the timing of the strikes, and an analytics firm identified six newly created accounts that together profited about US$1 million betting the strike would happen by that date ¹. In January, before the war, more than US$400,000 was wagered on Polymarket on Venezuela’s Nicolás Maduro being removed, hours before United States forces seized him ⁸.

The clearest case came at the other end of the fighting, on the ceasefire. On the evening of 7 April 2026, Washington time (the morning of 8 April in Australia), President Trump announced a two-week ceasefire with Iran ⁹. In the hours and minutes before he did, a cluster of accounts bet that it was coming.

One wallet placed roughly US$72,000 in bets at an average price of about nine cents on the dollar and cashed out for a profit of about US$200,000. Another, a third wallet, had been created just twelve minutes before Trump’s post; it staked US$31,908 at around thirty-three cents and is estimated to have earned about US$48,500 ³.

Representative Ritchie Torres put the arithmetic bluntly.

Polymarket did not immediately reply to a request for comment ². None of the accounts has been named, and none of these figures is a finding; they are the timing and the size of bets that anonymous accounts placed and were paid on.

Update, 7 October 2026. The ceasefire announcement is now marked as the evening of 7 April 2026 in Washington, which was the morning of 8 April in Australia, so that this date matches the Australian market dates in this case’s fourth article. Reference ⁹ has been added for the timing of the announcement. No fact has changed.

02Ninety-seven per cent

For months the ceasefire bets were a set of anecdotes: a wallet here, a cluster there, each striking on its own but none of them a measurement of the whole. In August a nonpartisan research group, the Anti-Corruption Data Collective, supplied the measurement.

The group analysed a large body of publicly available Polymarket data and found 152 accounts that had profited about US$8 million on the war markets, with what the reporting called a staggering 97 per cent win rate ⁴.

The company was already ahead of the report. A senior Polymarket official told CNN that the platform had referred dozens of accounts showing signs of possible military insider trading to the Justice Department for investigation, and that those referrals predated the group’s findings ⁴.

That is the turn this article records. The prediction market that took the bets is now the source of the referrals against them, and the figure that drew the referrals is not any single wallet’s winnings but the win rate of the field as a whole.

03The letters

Two members of Congress had already written, in April, over the ceasefire bets.

Representative Ritchie Torres wrote to the chairman of the Commodity Futures Trading Commission, Michael Selig, ‘calling for an immediate investigation into suspicious trading activity on Polymarket in the hours before President Trump announced a US-Iran ceasefire’, in the words of the report of his letter. In a companion request over the oil futures market he wrote that the pattern raised serious concerns that certain market participants may have had access to material nonpublic information regarding a market-moving geopolitical event ⁵².

Senator Richard Blumenthal wrote to Polymarket’s founder and chief executive, Shayne Coplan, on 9 April, framing the platform itself as the risk.

The concern in that sentence is not only that someone with advance knowledge might cash in. It is that a public, anonymous market in exactly-timed bets is a place where a foreign service could read the same signals, or plant them. Neither letter is a finding, and neither compels the platform to do anything; they are requests, on the public record, from members of the minority.

04What it is, and what it is not

Everything above rests on data anyone can pull. Polymarket’s ledger is public: the wallets, the bets, the timestamps and the payouts are visible to anyone who cares to read them, which is how an analytics firm, a research group and reporters were able to count what they counted. That openness is why the win rate can be measured at all, and it is also why nothing here is secret enough to have been misappropriated on the platform itself. The question the referrals and the letters raise is what the accounts knew before they bet, and that lives off the ledger, where no chart can reach.

So the record stops where the evidence stops. Anonymous accounts placed well-timed, winning bets on the war’s turns; a nonpartisan group put the field’s win rate at 97 per cent; the company referred dozens of accounts to the Justice Department; two members of Congress asked regulators and the company to look. No account holder has been named, charged or found to have traded on inside information.

The one financial penalty imposed anywhere over trading tied to the president’s own words did not fall on a Polymarket account. It fell in a rival prediction market, Kalshi, on Gabriel Perez, a former White House teleprompter operator who bet on the wording of Trump’s speeches using texts he saw before delivery; the CFTC ordered him in August to repay his profits and pay a penalty, US$172,000 in all, with a three-year ban ⁶. He is the only person who has paid. The winning ceasefire accounts, and the field the Anti-Corruption Data Collective measured, remain anonymous.

In the futures market, where the same war was traded in oil rather than in yes-or-no shares, the numbers are larger and the inquiry is further along: the Justice Department and the CFTC are examining at least four oil trades that together made more than US$2.6 billion betting prices would fall right before they did ⁷. That is the subject of this case’s first article. This one stays with the prediction market, and with the ninety-seven per cent.

The ledger is public, so the win rate can be counted. What the accounts knew before they bet is not on it.

If it’s a rort, we cover it.