(April 29): Prediction markets are being touted on social media as a lucrative side hustle for young Americans squeezed by rent and student loan bills. In reality, most traders are losing money and a significant amount in many cases.
Over 100,000 accounts lost at least US$1,000 on Polymarket, one of the largest prediction markets, according to a Bloomberg News analysis of every wallet active since the beginning of 2025. That is almost twice the number that made at least that much.
Among the winners, a majority of the profits were raked in by a tiny slice of what look to be automated bots, based on the Polymarket trade records compiled by the data firm Dune. Everyone else, in aggregate, lost US$131 million.
Polymarket’s blockchain ledger, which offers a public record of every wallet on the exchange, gives an unusual window into the financial fortunes of the millions of people who have begun trading on prediction markets over the last year as the nascent industry has exploded, opening up betting on everything from the elections to the Super Bowl to regime change in Iran.
Almost half of the two million wallets active since early 2025 made or lost less than US$10, indicative of people experimenting with this new form of betting. Even among this group, most ended up in the red.
See also: Digital finance providers seek EU carve-out from new legislation
“If you want to participate and you want to make a living out of this, you better be pretty darn good,” said Charles Martineau, a professor at the University of Toronto’s business school.
Martineau recently co-authored a paper that found that since 2022 around 69% of traders on Polymarket lost money, while the top 1% captured three-quarters of the profits. Martineau said that his students, who have been obsessed with prediction markets over the last year, were taken aback by his findings.
“You’d be surprised how many of them did not anticipate that you would see such concentrated gains — that so few make money,” he said.
See also: HSBC says companies aren’t ready to adopt digital finance
A spokesperson for Polymarket declined to comment on the recent research findings.
There is no indication that Polymarket customers are doing worse than those on Kalshi or other event betting platforms. Polymarket is just the biggest venue to offer a public blockchain record of all accounts. A report from analysts at Citizens earlier this year suggested that users are losing proportionately more on Kalshi than they do on sports betting apps, though Kalshi rejected that analysis.
One of the main selling points of event betting exchanges is that, unlike on sports gambling apps, there is supposedly no sportsbook or house that makes money when customers lose. The key innovation of prediction markets is that they allow buyers and sellers to meet on an exchange to bid on the odds of an event occurring, rather than relying on a house to set odds.
While prediction markets have been described as peer-to-peer, the Polymarket records suggest the role of the sportsbook is now largely being played by the sort of automated, high-frequency traders that have long dominated other financial markets. The most active accounts on the site were a small proportion of wallets but accounted for most of the trading volume.
Joshua Della Vedova, a professor at University of San Diego’s business school, constructed a method for categorising traders that labelled any wallet as a bot if it traded an average of at least 50 times on any day it was active or 1,000 times over the course of the data.
To stay ahead of Singapore and the region’s corporate and economic trends, click here for Latest Section
Using Della Vedova’s definition, Bloomberg’s analysis found that since the beginning of 2025, the typical bot averaged 89 trades on each active trading day — compared to 2.2 for non-bots — with trades spread across a greater diversity of markets.
These high-volume accounts collectively turned a profit of US$131 million, mostly concentrated among 823 users that netted more than US$100,000 each. The less active traders, meanwhile, lost the equivalent amount when all their wins and losses were added up.
Della Vedova found that the bots did not outperform because they were better at predicting outcomes but rather because they got into markets earlier and at better prices.
The accounts that Della Vedova identified as retail traders actually picked the right outcome more frequently, but they ended up losing much more money — tens of millions of dollars in aggregate — because they traded late, at bad prices.
“Retail investors, despite being correct, are losing money,” he said. “The execution edge is an underrated aspect of trading.”
Even among the bots, in both Della Vedova’s research and the Bloomberg analysis, a majority lost money. But the gains from the winners were so large that, as a group, they came out on top.
The Polymarket findings echo a long history of research on the underperformance of retail traders in an array of traditional financial markets. But while retail stock traders generally do worse than the overall market, their results are cushioned by the fact that few stocks end up worthless. Not so in prediction markets, where the consequence of a bad bet is losing 100% of your money.
Despite the unparalleled transparency it provides, the Polymarket data still only offers a partial picture of how customers are faring. The blockchain does not provide information on who owns each wallet, so one trader might control multiple wallets pursuing different strategies, potentially taking both sides of the same contract at various points.
Polymarket has also faced allegations that it hosts a high volume of so-called wash trading, coming from customers who are seeking cryptocurrency rewards, rather than a return on their trading. This sort of activity, though, would be unlikely to influence statistics on profits and losses because it would come from people who are trying to rack up lots of trades without taking long-term positions.
The University of Toronto researchers found that the profits in the system flowed mostly to so-called market makers, who offer to trade at a given price rather than just taking the prices that are already available on the exchange. The market makers did particularly well on sports wagers, with the winnings on certain categories, like weather, distributed more evenly, they found.
Pat Akey, one of the co-authors of the study, said that profitable traders generally have a clearly defined trading strategy and money to leave sitting on the exchange — not the sort of thing you’d expect from an inexperienced bettor looking to cover their monthly bills.
“I don't think prediction markets are bad — they can serve a role,” said Akey, a professor at ESSEC Business School. “I just don't think they can be viewed as a good way to supplement your rent.”
Uploaded by Arion Yeow
