Only 3% of traders drive prediction markets' accuracy, not the crowd, study finds

Researchers show market accuracy comes from a tiny group of informed traders, not broad participation.

What to know:

• A new working paper analyzing every Polymarket trade from 2023 to 2025 finds that about 3% of traders account for most price discovery, undermining the idea that prediction markets rely on the wisdom of the crowd.
• The researchers distinguish skill from luck by simulating each trader’s bets 10,000 times with coin-flip directions, concluding that only a small share of top winners consistently outperform chance and that many apparent winners regress when tested on separate events.
• Only 12% of the biggest winners by raw profit cleared that benchmark, and roughly 60% of "lucky winners" turned into losers when tested against a separate sample of events.

The Green Beret arrested for betting on a classified U.S. raid looked like a one-off scandal for prediction markets. A new study suggests he may be a more troubling data point: an extreme example of the small group of informed traders who, as the soldier is accused of doing, actually move prices on Polymarket, while the crowd loses money around them.

The study, part of a working paper released this week by Roberto Gómez-Cram, Yunhan Guo, Theis Ingerslev Jensen and Howard Kung of London Business School and Yale, directly tests the industry's core claim that the markets work owing to the massed knowledge of their participants.

Using every Polymarket trade from 2023 to 2025, the authors conclude that it's actually a small group of informed traders that moves prices. The researchers analyzed 1.72 million accounts and $13.76 billion in trading volume, and found that just 3% of traders account for most price discovery, meaning they are the ones moving prices toward the correct outcome.

These traders consistently predict outcomes and move prices in the right direction. The remaining 97% mostly do not. They provide liquidity and generate volume, but in aggregate, they are on the losing side of trades against the informed minority, whose profits come directly from those positions.

The hard part is telling skill apart from luck. With more than a million traders on Polymarket, plenty will rack up big winnings by chance alone.

To filter that out, the authors reran each trader's bets 10,000 times, keeping everything the same except the direction.

Same markets, same moments, same dollar amounts — but a coin flip decided whether to buy or sell. That gave them a benchmark for what each trader's profits would look like with no real edge. If the actual results consistently beat the coin flip, that's skill. If not, it's luck.

The findings show among the biggest winners by raw profit, only 12% beat the benchmark, and many apparent winners didn't stay that way: Roughly 60% of "lucky winners" become losers when their performance is checked against a separate sample of events.

Their activity improves market accuracy. When skilled participants account for a larger share of trading, prices move closer to the correct outcome, especially in the final stretch before resolution. They are also the first to react when new information hits, shifting positions in response to events like Federal Reserve announcements or corporate earnings, while other traders show little consistent reaction.

The same edge that makes skilled traders valuable to price discovery raises a harder question when that information isn't public, or isn't supposed to be.

Both Polymarket and Kalshi have said that trading on non-public information is strictly against their rules.

The paper grounds that risk in a concrete case: The U.S. removal of Nicolás Maduro from power in Venezuela in January. In the days and hours before the operation, three newly created Polymarket accounts piled into a contract asking whether Maduro would be removed. At the time, the market priced the odds at roughly 10%.

The new accounts placed unusually large bets, including orders of tens of thousands of shares, before the price moved. When the raid happened, the accounts collectively made more than $630,000. Two stopped trading entirely soon after, and the third went mostly dormant. There is no evidence of any wrongdoing on these accounts.

Insider trades, when they occur, move prices even more aggressively per dollar, about seven-to-12 times more than typical skilled trades. But they are rare and concentrated in a handful of events, not the day-to-day engine of price discovery. Most of the time, the market's accuracy still depends on repeat traders who consistently outperform rather than on one-off bets.

The findings challenge the idea that prediction markets work because of crowds. They appear to work because of who is informed.

More For You

Litecoin's foundation called the weekend exploit a zero-day. The litecoin-project GitHub repository shows the consensus vulnerability was privately patched between March 19 and 26, more than four weeks before the attack.

What to know:

• Attackers exploited a patched, but not fully deployed, vulnerability in Litecoin’s Mimblewimble Extension Block protocol, triggering a 13-block chain reorganization that rewound about 32 minutes of activity.
• Public GitHub commits show the core consensus bug was privately fixed weeks before the exploit, creating a window in which some mining pools...

MiCA's not enough: Bybit CEO says firms need other licenses to turn a profit in Europe

1 hour ago

Coinbase's John D’Agostino says crypto platform stands alone as industry's full-service prime broker

2 hours ago

Litecoin hit by denial-of-service attack, rewrites 13 blocks to reverse effect

5 hours ago

Crypto is built for AI agents, not humans, says Alchemy's CEO

19 hours ago

Trump defends crypto legislation at private event featuring boxer Mike Tyson, Tether CEO

19 hours ago

BlackRock’s bitcoin ETF just hit a massive milestone that proves crypto is now a mainstream bet

19 hours ago

Bitcoin falls after Trump reportedly canceled Steve Witkoff and Jared Kushner's Iran-talks trip

21 hours ago

Clock is ticking for bitcoin to prevent quantum threat as it could drain 6.9 million BTC including Satoshi’s

Apr 25, 2026

Coinbase’s Jesse Pollak says AI agents are the next big wave for crypto payments

Apr 25, 2026