A new report titled “Classifying Insider Trading Risk” from the Anti-Corruption Data Collective (ACDC) has identified 556 Polymarket accounts, categorized as “Orcas,” that demonstrate a success rate of over 75% while placing small, targeted bets on events with low odds.
The ACDC study examined 78,496 longshot bets placed by 12,355 wallets on Polymarket. Researchers defined a longshot as a wager exceeding $2,500 at a price of 35 cents or less, representing an outcome with a 35% chance or worse. The research divided the analyzed wallets into four groups: Orcas, Whales, Bots, and Small Fish.
Categories of Polymarket Bettors
The report detailed the characteristics of the four bettor categories:
- Orcas: 556 wallets that bet in only a handful of markets and topics with a success rate exceeding 75%.
- Whales: 3,278 wallets that trade at high volume across many markets.
- Bots: 760 accounts that appear at least partly automated.
- Small Fish: 7,761 remaining accounts.
An earlier April report by ACDC found that across the entirety of Polymarket, only 14% of longshot bets win. However, that win rate climbs to 25% in political markets and reaches 52% in military and defense markets. Authors Michelle Kendler-Kretsch and David Szakonyi documented $9.3 million in winning longshot bets from military markets alone.
National Security Concerns
According to the report, Orcas frequently place wagers first on major military events, while Whales and Bots follow behind in a manner similar to traders observing public blockchains to copy insider-like moves. While this copycat trading is legal, ACDC noted that it can broadcast insider activity globally.
For example, the report cited an instance where an Orca placed a bet on U.S. military action in Iran hours prior to June 2025 strikes, after which a Bot and a Whale placed copycat bets of $200,000 and $100,000, respectively. ACDC warned that blockchain transparency allows foreign militaries and intelligence services to act on these market signals.
Tracing pseudonymous Orca accounts back to specific individuals remains complicated, requiring subpoenas to exchanges. Consequently, ACDC concluded that banning the highest-risk market categories outright is the only solution.
Other Incidents on Prediction Platforms
The report's findings follow other recent legal cases involving prediction markets:
- In April, the U.S. Justice Department charged U.S. Army soldier Gannon Ken Van Dyke with staking roughly $33,034 on Venezuela contracts using classified details of the raid that captured Nicolás Maduro, yielding about $409,881.
- In May, prosecutors in the Southern District of New York charged Google engineer Michele Spagnuolo with using internal company data to win $1.2 million on Polymarket.
- Additionally, Bubblemaps stated on CBS’s “60 Minutes” that it traced nine linked accounts winning more than $2.4 million across over 80 Iran war bets with a 98% win rate. Bubblemaps CEO Nicolas Vaiman stated that luck alone cannot explain those numbers.
