How prediction markets have turned into corporate barbershops
Marketers have effectively lured users to prediction markets like Polymarket or Kalshi with narratives about the "wisdom of the crowd," which anyone can be a part of. But is this really the case?
ForkLog examined two recent instances on Polymarket where whales achieved their desired outcomes despite objective realities and the actual results of significant bets.
The Wisdom of Insider Crowds
Proponents of prediction markets often point out that such platforms vividly illustrate the validity of the "wisdom of the crowd" hypothesis. This concept was popularized by American journalist James Surowiecki in his 2004 bestseller, although it has existed in various forms for centuries.
In 1906, English scientist and founder of differential psychology (and eugenics) Francis Galton conducted a groundbreaking study for his time. He attended a fair where guests were invited to guess the weight of an ox for six pence.
Galton aimed to collect data on the "votes" of a crowd with diverse knowledge and skills to confirm his intuition about the dangers of universal suffrage. He theorized that just as a stockbroker cannot accurately gauge the weight of an animal by sight, a farmer cannot judge the competence of politicians.
The experiment yielded surprising results. After gathering the "votes" of 787 participants and calculating the median, Galton found that the "average voter" was off by only one pound—minimal considering the ox weighed 1,198 pounds.
The idea that the opinion of an uninformed majority can be more effective in forecasting than that of individual experts is undoubtedly appealing to advocates of democratic society. Moreover, as behavioral economics specialists know well, it flatters the egos of consumers across the political spectrum.
But is the "wisdom of the crowd" applicable to modern prediction markets? This has been questioned by researchers from the London Business School and Yale University. In a joint study, they analyzed transactions on Polymarket from 2023 to 2025 and reached an unflattering conclusion for most: 3.14% of platform participants account for 30% of profits.
"Prediction markets are remarkably accurate, but the source of that accuracy is not the 'crowd.' […] The aggregation of information in prediction markets works differently than traditional notions of 'wisdom of the crowd' suggest. Instead of many participants contributing partial information, prices appear to be driven by a small number of informed and qualified traders whose trades move the market toward final outcomes. Other participants provide liquidity and trading volume but have little impact on price formation," the study's authors conclude.
They highlight not only the speculative behavior of traders but also numerous signs of dishonest trading using insider information.
Such accusations are frequently directed at Polymarket and Kalshi, whose strategic advisor is Donald Trump Jr.
After a scandal involving bets on a military operation in Venezuela, Congress began discussing a ban on prediction markets related to significant public issues. Kalshi has already responded to potential regulation by blocking accounts, including politicians who placed bets on the outcomes of their own campaigns.
Much less attention is drawn to other types of incidents. Seemingly minor in the grand scheme, they increasingly affect many ordinary users.
Beirut's 3D Chess
In early March, the radical Shiite group Hezbollah launched rocket strikes in northern Israel. In response, the IDF resumed ground operations in southern Lebanon while also conducting airstrikes in Beirut and Tyre.
When American diplomacy intervened to resolve the conflict, a prediction market was launched on Polymarket, where participants could bet on the date of an officially announced ceasefire between Israel and Hezbollah.
On April 16, President Donald Trump announced a ceasefire effective at midnight on April 17. Shortly after, the market, which had gathered over $112 million, was prematurely closed at the request of a user who had bet on a positive outcome by April 18.
This decision sparked discontent among participants who predicted that a ceasefire would not be reached by the specified date. Firstly, the ceasefire was largely formal: just hours after it was announced, the Lebanese government accused the IDF of new strikes on its territory. Secondly, Lebanon is not an active party in the current conflict, and Hezbollah refused to participate in negotiations until Israel withdraws its troops and releases detained supporters of the group.
Those unhappy with the market's closure attempted to contest the decision through the UMA oracle platform. However, they were unsuccessful: 86.6% of votes, proportional to the number of tokens held by delegates, favored the fairness of Polymarket's decision, while only 1.48% voted against. The "Yes" option received approximately 14.5 million tokens. As of April 16-17, UMA was trading around $0.45.
This triggered a new wave of comments regarding the closed market. Dissatisfied users accused Polymarket and UMA of collusion, criticized the oracle's operational model that allows "whales" to manipulate decisions, complained about the lack of response from platform administrators, and demanded compensation.
"If Polymarket cannot reverse the final result on the blockchain, it should still consider refunding or compensating losses in cases where a disputed market failed to provide adequate protection against contested interpretations of the rules. This is a matter of market integrity, not just trader dissatisfaction," noted one user.
"A $50 Million Crime"
Simultaneously, a similar incident occurred related to the conflict in the Middle East. Co-founder of a Web3 startup, Yuri Yatsenko wrote in his X post titled: "Right now, a $50 million crime is happening on Polymarket."
The trigger was a prediction market on a ceasefire between the American-Israeli coalition and Iran. According to the terms, a positive outcome would be counted if the parties maintained the ceasefire for 14 days, or 336 hours. Thus, according to Polymarket administrators, this occurred. However, Yatsenko notes that on April 7, the day the market was closed in favor of a positive forecast, rocket fire continued for 20 out of 24 hours.
"Contested resolutions were sent to vote on UMA; 50 new wallets opened large 'Yes' positions before the results were announced; one wallet turned about $72,000 into approximately $200,000 in just a few hours," Yatsenko concluded.
Experience shows that such conflicts rarely end favorably for participants in prediction markets who feel cheated. For instance, in March 2025, Polymarket representatives refused to compensate users affected by the premature closure of a prediction market with an incorrect outcome regarding a mineral deal between Ukraine and the U.S.
Researchers found that three large token holders, accounting for 25% of the votes, managed to achieve an unfair decision on UMA. Polymarket acknowledged the occurrence of a governance attack, labeling the case as "unprecedented."
"This is not part of the future we want to build. We will create systems, monitoring, and much more to prevent this from happening again," representatives of the platform stated at the time.
Back in August 2024, Vitalik Buterin wrote that "categorizing Polymarket as gambling means not understanding what prediction markets are." The co-founder of Ethereum, who is well informed about the workings of such platforms, is absolutely right.
By 2026, prediction markets have indeed transformed from advanced betting shops into yet another infrastructure where ordinary users serve merely as liquidity providers for major players.
ForkLog reached out to Polymarket twice—on April 23 and 28—for their perspective on the situation. We have yet to receive a response.