PredictionMarkets Rely on a Small Cohort of Expert Traders, Yale Research Shows

While fairy‑tale protagonists consult crystal balls for foresight, today’s prediction markets allow traders to wager on the probability of real‑world events—ranging from election results to Eurovision winners—often achieving remarkable accuracy.

The creators of these markets, which have expanded dramatically over the past two years, credit their accuracy to the wisdom of crowds—the notion that aggregated bets from a diverse population tend to converge on the correct outcome. “Everyone has skin in the game and a strong incentive to express their true beliefs,” says Yale School of Management researcher Theis Ingerslev Jensen. “That’s why it seems logical to attribute their success to crowd wisdom.”

However, he points out that prediction markets resemble traditional financial markets, which are driven by a minority of traders possessing superior skill or information. “We set out to evaluate the prevailing crowd‑wisdom explanation against the more conventional view that operates in conventional financial markets,” Jensen explains.

In a forthcoming working paper, Jensen and his collaborators—Roberto Gómez‑Cram, Yunhan Guo, and Howard Kung of the London Business School— examined publicly available trading data from Polymarket, which describes itself as the world’s largest prediction market.

Their analysis revealed that market accuracy stems from a small cohort of skilled traders, who, despite representing a tiny fraction of participants, capture the majority of profits.

In prediction markets, participants purchase binary yes‑or‑no contracts on the occurrence of events that often have little to do with the real economy. In “mention markets,” traders bet on whether a public figure will mention a given word in a speech. A decentralized committee decides whether the event has occurred, leading to a payout of $1 per contract if the event occurs (a Yes), or nothing if it does not (a No). A higher likelihood of an event happening raises the price of the contract.

Their analysis of two years of Polymarket data—covering 1.72 million accounts, 98,906 events, 210,322 markets, and $13.76 billion in trading volume—showed that only 3 % of accounts can be classified as “skilled,” with significantly positive profit that cannot be explained by random chance. About twice as many accounts performed worse than the benchmark, a group they designated as “unskilled.” The skilled group, alongside an even tinier group of market makers, “represent fewer than 3.5 % of all accounts, yet capture over 30 % of total gains,” the authors write.

Another 29 % of traders, whom the authors call “lucky winners,” managed to make money without any statistically discernible skill—they profited, but not beyond what randomized simulations suggest could occur by chance. Those traders captured the remaining 69 % of gains within the two-year period.

To test whether skilled traders were merely lucky, the researchers divided each trader’s event list in half at random. A truly skilled trader would outperform the benchmark in both halves, and they found that 44 % of traders classified as skilled on the first half are also classified that way on the second—significantly higher than the 10 % consistency rate seen in mutual fund managers.

“Most people think that if you beat the market in one period, then that was just luck, and so you’re not going to continue beating the market,” says Jensen. “We found that for prediction markets, there’s an unusually high level of persistence. If you classify someone as skilled in one period, they are much more likely to be skilled in the next period than random chance would suggest.”

To assess the “wisdom of crowds” hypothesis, the team examined how different trader groups influenced contract prices, which proxy for event likelihood. They focused on recurring events such as Federal Reserve Open Market Committee announcements and<|message|><｜begin▁of▁sentence｜># 1.5.1.1.1.1.1.1.1.1.1.1.1.1.1.1.1.1.1.1.1.1.1.1.1.1.1.1.1.1.1.1.1.1.1.11111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111113<｜begin▁of▁sentence｜># <｜begin▁of▁sentence｜># [0.3109]<｜begin▁of▁sentence｜> “Predictive accuracy in prediction markets depends on the actions of a small cohort of expert traders, not on the ‘wisdom of crowds'” “A recent Yale University study reveals that the success of prediction markets, often touted as a manifestation of the ‘wisdom of crowds,’ is actually driven by a small cohort of highly skilled traders, challenging the prevailing belief in collective wisdom. The research, published as a working paper, analyzed two years of data from Polymarket — the world’s largest prediction market — covering 1.72 million accounts, 98,906 events, and 210,322 markets with $13.76 billion in total trading volume. The study found that just 3 % of accounts exhibited statistically significant skill, yet these skilled traders captured over 30 % of total gains, while the remaining 96.5 % of participants earned returns indistinguishable from random chance. The study also found that skilled traders’ performance persisted across periods, outperforming random chance significantly more often than in traditional financial markets. This persistence suggests that skill, not luck, drives accuracy, and that the bulk of market gains are funded by the majority of less‑skilled participants. The findings indicate that price discovery in prediction markets is driven primarily by skilled traders, while the majority of participants fund the market’s accuracy rather than contribute to it. The study further notes that insider trading has only a minor, sporadic impact on price discovery. “Our results demonstrate that skill persistence in prediction markets is unusually high compared to traditional finance, suggesting a distinct mechanism of price discovery,” said Theis Ingerslev Jensen, a researcher at Yale School of Management and lead author of the paper. “The evidence indicates that skilled traders consistently move contract prices toward the final outcome and react more quickly to new information, thereby enabling price discovery without relying on insider information.”

Source link