A Green Beret allegedly made $400,000 in insider bets during a raid in Venezuela, in what appears to be an isolated breach. A new report suggests this may be the visible edge of something broader.
The Anti-Corruption Data Collective (ACDC), a non-profit research organization, analyzed every settled Polymarket contract from January 2021 to mid-March 2026 – more than 435,000 markets with a cumulative trading volume of $54.4 billion – and found that the winning rates of low-probability bets on military and defense outcomes are difficult to explain by skill or luck.
Across political markets, such “low-risk” bets typically succeed about 14% of the time. In military-related contracts, the success rate exceeds 50% in some cases.
“Markets related to specific government policies (such as military, defense, and foreign affairs) are more difficult to predict using public information alone,” the authors write, making them “more susceptible to information asymmetries,” including insider trading or expertise.
In these markets, the gap between informed traders and uninformed traders is probably the widest, creating a condition where a small group of people can consistently outperform not just by reacting faster, but by knowing more.
Polymarket, for its part, touted its market surveillance team and its work with the Justice Department on the Venezuela case. Like Kalshi, the platform prohibits the trading of confidential knowledge.
Profit concentration
The ACDC report’s findings add to research pointing in the same direction. A working paper from London Business School and Yale University found that approximately 3% of traders found most of their prices on Polymarket.
Separate analysis by blockchain analytics firm Solidus Labs shows profits are more concentrated, with less than 1% of wallets receiving about half of the proceeds. ACDC’s contribution is to suggest where some of the advantages might come from.
The report uses the June 2025 U.S. attack on Iran as a case study. Polymarket lists several date-specific contracts as to whether a strike will occur. Markets related to June 19th and June 20th expired with no events and no winning bets.
The strike took place at 18:40 EST on June 21st. In the hours leading up to the strike, 19 bets were placed on the contract, totaling $164,292, with the final result being “yes.” Eight wallets shared approximately $1.8 million in profits, with one wallet taking away nearly $500,000.
The Pentagon designed the operation to be unreadable from the outside and used decoy bombers and long-range stealth aircraft to avoid detection. Still, a handful of traders made big, well-timed bets on the outcome.
The pattern goes beyond a single event. The report found that in Polymarket’s Military and Defense category, there were more winning bets than losing bets in five of the six two-hour windows before market resolution, contrary to what the market price would suggest.
“Long-shot” bets can also outperform for other reasons, including mispricing or changes in public expectations. But the consistency of these patterns, particularly in markets related to military decision-making, suggests that some players may be operating with information advantages that others do not have.
ACDC is a non-profit research organization funded by the Constitutional Government Fund, and in contrast to Solidus Labs, it has no monitoring products for sale. Solidus Labs’ recent Polymarket analysis also serves as a marketing example for the platform it licenses to Kalshi.
ACDC’s recommendations include identity verification of bettors, conditional payouts for questionable bets, limits on markets where outcomes are determined by small groups, and limits on the level of granularity in contracts.
The report’s conclusions go further, calling for “an evidence-based debate about whether the public should bet on these outcomes.”
