In a new report, the Anti-Corruption Data Collective (ACDC) identified 556 Polymarket accounts, which the organization refers to as "Orcas."
Reportedly, these accounts place small, targeted bets on low-probability outcomes and win more than 75% of the time.
What did ACDC discover about Polymarket?
The Anti-Corruption Data Collective (ACDC) published a report titled "Insider Trading Risk Classification," analyzing 78,496 high-risk bets placed from 12,355 wallets on Polymarket. ACDC defines a high-risk bet as one exceeding $2,500 placed on an outcome priced at 35 cents or less, meaning the perceived probability of the outcome is 35% or lower.
From this sample, researchers identified four groups: orcas, whales, bots, and "small fish." They recorded 556 wallets belonging to orcas, which placed bets on only a few markets and topics but won with an exceptionally high success rate of over 75%.
Large players, controlling 3,278 wallets, trade large volumes across many markets. Bots, with 760 wallets, appear to be partly automated, while the remaining 7,761 accounts are small players.
In its previous April report, ACDC found that across the entire Polymarket platform, only 14% of high-odds bets win, but this rate increases to 25% on political markets and to 52% on military and defense-related markets.
Michelle Kessler-Kret and David Szakonyi, authors of the report, estimated that on military markets alone, winning outsider bets have generated $9.3 million.
Why is insider trading on Polymarket a national security concern?
ACDC found that when betting on major military events, orcas place their bets first, followed by whales and bots, mimicking how experienced traders observe the public blockchain and copy insider actions in real time.
Such "copycat" trading is legal, but it can broadcast information about insider activity to the world.
For example, when an orca placed a bet on U.S. military action in Iran hours before the strikes in June 2025, a bot and a whale placed similar bets worth $200,000 and $100,000, respectively.
The same blockchain transparency that enables copycat trading also allows foreign military and intelligence services to act based on market signals.
Tracing a specific Orca account to a specific person is difficult, as the monetary trail is anonymous, and obtaining any information would require requesting data from exchanges holding funds for thousands of clients.
ACDC concluded that the only solution is a complete ban on the most risky market categories.
There have been several other incidents of insider trading on prediction platforms. In April, the Department of Justice charged a U.S. Army soldier named Gannon Ken Van Dike, accusing him of betting approximately $33,034 using secret information about a raid that captured Nicolás Maduro, and subsequently earning about $409,881.
In May, the U.S. Attorney's Office for the Southern District of New York charged Google engineer Michelle Spagnolo with using internal company data to earn $1.2 million on Polymarket.
Bubblemaps also recently claimed on CBS's "60 Minutes" program that it traced linked accounts that collectively won over $2.4 million in more than 80 bets on the Iran war with a 98% win probability.
"These numbers cannot be explained by luck alone," said Bubblemaps CEO Nicolas Vaiman.
end-content
