Parlay bets are burning gamblers as market takes off on Kalshi
Published in Business News
Retail bettors are piling into long-shot multi-leg bets on prediction-market platforms, opening up a lucrative opportunity for sophisticated traders taking the other side.
Combo wagers, known as parlays in sports gambling, let users bundle multiple outcomes into a single bet. One popular combo for the World Cup final had Spain leading at halftime, Argentina winning the match and either team scoring at least two first-half goals. It traded at an implied probability of 2.7% at kickoff.
Combo bets have become the fastest-growing corner of the prediction-market world since first showing up late last year, representing 36% of the contracts traded on Kalshi this month.
They have also been among the most dangerous for ordinary customers: bettors on Kalshi’s app and website have lost a net $294 million on its combos since the start of the year excluding fees, according to a Bloomberg analysis of the so-called taker trades that mimic the bets made on traditional sportsbooks.
The losses on these sorts of wagers used to be largely captured by sportsbooks, where parlays have been among the industry’s most profitable products. Prediction markets are changing that dynamic by promoting similar trades to their small-time customers on their apps, while allowing Wall Street firms and other algorithmic traders to step in as counterparties.
Leonidas Mastrokostas, a 26-year-old Jersey City resident, learned about the opportunity when he worked at FanDuel, one of the biggest sportsbooks. He has since struck out on his own, and he says he has been making seven figures a month by betting against the risk-taking instincts of ordinary gamblers on Kalshi.
“These lottery tickets are what retail is really looking for,” said Mastrokostas. “They don’t quite understand the pricing, but at the end of the day, given the competition of many makers there, they’ll ultimately lose less.”
Mastrokostas’s old employer, FanDuel, is also embracing the opportunity. Parlays already account for an outsized portion of the profits on its traditional sportsbook. But FanDuel has recently started market-making combos on other exchanges as well.
“It’s great we’re making money today from offering this capability, particularly focused on combos and leveraging the pricing expertise we have,” Rod Coldrake, the chief financial officer of FanDuel’s parent company, Flutter Entertainment PLC, said on a call with analysts in May.
Sportsbooks make far more on parlays than on other sports bets because they are better able to stack the odds in their favor. Bettors lose 19 cents out of every dollar wagered on such bets, compared to 6 cents for straight bets, according to state data from May 2025 to June 2026. Sportsbooks have put them at the center of their advertising, while bettors are drawn to the prospect of a quick, large payoff despite their losses.
Unlike sportsbooks, Kalshi does not take the other side of customer trades. But combos have still been a lucrative new business for the company, in part because of the fees it collects on every trade.
Bloomberg’s analysis focused on so-called market takers on Kalshi, who accept the price displayed in the app, similar to how the average sportsbook customer selects a parlay with preset odds. Traders seeking to take the other side of these wagers on prediction markets — the market makers — do so with automated systems, which most ordinary customers are not able to build.
For now, nearly all combo volumes involve sports games, though Kalshi also makes it possible to include crypto into combos, allowing customers to put a bet on the price of Bitcoin alongside the number of points in a basketball game or the winner of a tennis match.
During the World Cup final between Spain and Argentina, Kalshi users traded on more than 30,000 unique same-game combos, with many attracting thousands of bets. In the end, fewer than 3% of these multi-leg markets resolved to “Yes,” fueling more than $5 million of net losses.
Parlays are often dangerous for ordinary bettors because the quickly compounding nature of low probabilities makes it hard to know if a bet is priced fairly. Trading records for 2026 show that Kalshi combos were priced at a 9% chance of resolving to yes, on average, compared with 43% for non-parlay markets. The true odds on combos are often even lower than what bettors see on the screen, which is what makes them so lucrative for the traders on the other side of the bet.
Bets on so-called single game parlays, involving multiple outcomes in the same game, are particularly hard to properly value because of the difficult math behind correlations. Matt Buchalter, a Canadian actuary who runs the blog Plus EV Analytics, gives the example of a combo requiring world No. 1 tennis player Jannik Sinner to win a match, win the first set and have more than 10 aces. While these may all seem positively linked, if he wins in straight sets, the match may actually be shorter, making it less likely for him to hit that many aces.
Mastrokostas said he believes that the sophisticated players vying to offer prices on combos mean that the ordinary gamblers are getting a better deal than they would on sportsbooks, though it’s less clear if this is true after fees.
“Before these prediction markets existed, everyone was effectively a taker and the makers were the sportsbooks,” he said. “It was their price or no price.”
Kalshi first started offering multi-leg wagers last September, and the company’s app now features a handful of eye-catching combos at the top of the sports homepage, before the user encounters the more traditional, single-outcome bets.
Polymarket only launched its combos recently, so it has much more modest volumes, but it is marketing them heavily with a campaign that awards $50,000 every day to the trade with the highest return.
The popularity of these products could create problems for the exchanges, because customers can easily burn through their money, forcing them to walk away from the platforms. That’s why sportsbooks have traditionally tracked bettors to tailor bonuses and manage their exposure to parlays, according to Jordan Bender, an equity analyst at Citizens.
For the sophisticated traders, or market makers, looking to take the other side of combos, this is also a different challenge than offering prices elsewhere on prediction markets.
After a customer submits a combo, market makers on Kalshi generally get about a second to offer their best price in a mechanism known as request for quotation, or RFQ, with the winning market maker matched with the customer. (It’s akin to the system with the same name used to trade bonds and typically less liquid securities on Wall Street.)
Pricing parlays is so difficult that many RFQ market makers are happy minting profits by pulling the odds from sportsbooks. Lately, though, competition has become fiercer, so market makers have to figure out where they can afford to drop their price to win trades, says Gianni Settino, a 36-year-old software engineer in Los Angeles.
“If everyone’s using the exact same algorithm to come up with a price, you’re never going to reach the user because you’re just at the same level as all the other market makers,” said Settino, who started responding to RFQs as a side hustle. “You have to find spots where you can be more competitive.”
Sometimes, he ends up on the other side of unexpected combos, like one in May with 13 legs spanning six sport leagues, among them the WNBA, women’s tennis and cricket.
As competition intensifies, Mastrokostas believes he has an edge from estimating the fair value of each combo on his own rather than relying on sportsbook odds, since he will know better where he can afford to drop the price. He specializes in the NBA, where he simulates games in order to compute the correlations between various bets.
Some customers believe they can also turn a profit on traditional parlays by noticing when the odds are off. Buchalter, the actuary who specializes in combos, says he has managed to make money this way, but he knows that the average bettor does not.
“Most recreational public bettors have no clue how to price these things,” Buchalter said. “The public loves them because the public is very easily confused about all the hidden correlations and what the value might be.”
(With assistance from Aaron Gordon.)
©2026 Bloomberg L.P. Visit bloomberg.com. Distributed by Tribune Content Agency, LLC.











Comments