Parlays on prediction markets may have lower margins on average than sportsbooks, but users who make these bets lose money more quickly than sportsbook parlay bettors, according to an analysts’ note. However, Kalshi argues the research from Citizens Bank is “misleading and inaccurate.”
Analysts at investment bank Citizens published a note Thursday that looked at data from Juice Reel to examine customer economics on Kalshi.
Juice Reel is a bet-tracking service that allows users to sync their accounts from both sportsbooks and prediction markets in order to record wins and losses. Using data from Juice Reel users, Citizens was able to examine customer-level data — how much an individual loser might win or lose. The note looked at return-on-investment rates (ROI), or how much of a user’s bankroll they use, rather than losses as a percentage of stakes.
While Kalshi publishes a large amount of public data, each trade is anonymized, meaning that based on Kalshi’s own data there is not a way to know how quickly an individual user might be winning or losing their money.
The Citizens note is based only on Juice Reel users, who may not necessarily be representative of all Kalshi parlay takers.
Citizens: PM parlay bettors fare worse
Citizens found that prediction market parlay users on Juice Reel had an ROI of -18%. In other words, between October 2025 and publication of the note, an average parlay bettor with a $100 bankroll would be left with $82. This compared to -12% for sportsbook parlay bettors.
Looking at all users on sportsbook and prediction market bettors, Citizens found prediction market users lose 8% of their bankroll, compared to 6% for sportsbook users.
The finding comes despite the fact that parlays on prediction markets generally come with lower margins than parlays at sportsbooks.
Citizens says that sportsbook parlay hold for the year to date is 18.5%. In comparison, on Kalshi the figure is 13%, according to InGame research.
Therefore, if prediction market users are depleting their wallets more quickly, it would appear to be because those making parlays on prediction markets are betting more or doing so more frequently. The note said that larger parlay stakes appear to be the reason for the discrepancy.
“We believe there is a common misconception that prediction market companies, with lower fees or take rates compared to sportsbooks, must therefore be more customer-friendly or less predatory,” the Citizens note said. “As we discuss below, operator economics and customer economics are not the same, helping explain our analysis that suggests retail customers face worse ROI when using prediction market products.
“We estimate prediction market exchange fees blend to the low-single digits as a percentage of volume, while historically, sportsbook gaming margins are 5% for single-game betting, 18% for parlays, and 9.7% blended in the TTM. That said, a single trade or bet does not illustrate the complete story of a customer’s experience as it fails to capture factors such as trading frequency, time spent on device, session counts, fees, or psychological differences between trading and betting.
“For example, users cycle through their wallets faster as the average trade for a prediction market combo is $66, 92% higher compared to $34 for a parlay.”
Citizens also flagged the specifics of Kalshi’s request-for-quote process on parlays as a possible factor in the differing returns on investment. “Combos go through an RFQ (request for quote) process, allowing takers to identify makers, effectively enabling sharper participants to target more recreational bettors or filter out toxic/sharp flow,” the note said. “Put differently, customers receive dynamic pricing, but market makers can identify the counterparties they are trading with, creating an advantage because they can avoid trading against superior competition.”
Kalshi finds fault with the data
Kalshi did not feel that the data was reflective of how things worked on its exchange.
“Citizens Bank’s report is misleading and inaccurate,” Kalshi spokesperson Elisabeth Diana told InGame. “It’s based on a flawed subset of data, and it ignores a large cohort of some of our most successful traders and market makers. Of course outcomes look worse if you don’t count winners.”
Part of the objection from Kalshi was due to the fact that Citizens only appeared to look at parlay takers, rather than market makers. In theory, anyone can make markets on Kalshi, though in practice for parlays it takes more work and resources than taking trade offers. Parlay takers on Kalshi tend to be more similar to recreational sports bettors, while market makers on parlays are often professional bettors or institutions.
Kalshi has also had disputes with Juice Reel in the past. In February, when Bloomberg reported on a similar note from Citizens — looking at overall ROI rates, rather than specifically for parlay users — Diana initially claimed the data was part of an “extortion” plot by Juice Reel. Bloomberg reported that she later “backed away from the allegation of extortion,” but that Kalshi continues to dispute the findings.
