Share 
Traders on Kalshi and Polymarket are putting significant money into longshot bets that lose about 98% of the time, according to a Bloomberg analysis of trading activity. Since early 2025, these wagers have accounted for about half of Kalshi’s activity. On Polymarket, they made up 52% of activity over the past 21 months.
The findings challenge both platforms’ image as neutral information aggregators where traders collectively forecast future events. Much of the activity looks closer to gambling than forecasting.
What counts as a longshot bet
Longshots are bets on unlikely outcomes that offer large potential payouts. These wagers chase returns of 1,000% or more, so a $10 stake would target a profit of $100 or more. The contracts trade at low prices because the market considers the outcome highly unlikely.
Odds like these are more common in sports betting parlays than in forecasting markets, raising questions about how traders use the platforms.
Longshot backers lose 15% to 27%
Longshot bets return poorly, Bloomberg reported. Bets on the yes side, which pay out only if the unlikely event happens, produced average losses of 15% on Kalshi and 27% on Polymarket. A trader who wagered $100 lost an average of $15 on Kalshi and $27 on Polymarket. Trading fees of up to 7% can deepen those losses.
Because the contracts lose roughly 98% of the time, the occasional large payout is outweighed by the many losing stakes.
Polymarket and Bloomberg’s numbers sharply diverge
A Polymarket spokesperson said longshot wagers make up only about 1% of its trading volume. Bloomberg’s analysis found they accounted for 52% over the past 21 months. The two figures differ sharply, and it is unclear whether the company and Bloomberg define longshots the same way.
Kalshi, which Bloomberg found had about half of its volume tied to longshots, declined to comment.
Why bettors can’t see they’re losing
Philip Newall, a former professional poker player who researches gambling behavior at the University of Bristol, said the draw is less about the event itself than the thrill of a big, unlikely payout. The bigger puzzle, he said, is why bettors cannot tell when they are on the losing side.
He noted that poker players routinely overestimate their chances in rare, high-reward situations. Prediction market users appear prone to the same misjudgment, he added, when a contract promises a payout 10 or even 100 times the amount staked.
Brian Cai, who previously worked at Meta Platforms, said he mixed long-shot bets on prediction markets with more moderate ones. The longshots’ poor performance became clear to him only over time.
Scrutiny of prediction markets keeps growing
Bloomberg’s analysis is not the first damaging report on Polymarket. An earlier study found that most accounts on the platform lost money between January 2025 and April 2026.
Both platforms are regulated by the Commodity Futures Trading Commission, but legal challenges have questioned whether their sports contracts count as regulated gambling, like traditional sports betting. Prediction News has noted in its analysis of the Bloomberg report that persistent trader losses could hurt customer retention and draw further regulatory attention.
Meanwhile, Kalshi and Polymarket are both expanding their sports and political offerings.
