According to a report by Bloomberg, participants in Kalshi's parlay betting have incurred losses amounting to $294 million since the beginning of the year. However, the actual figure is likely much higher, as explored by Gambling Insider.
This statistic reflects a common way of measuring losses in prediction market parlays, which combines two distinct groups: recreational bettors who back a parlay and those who bet against it. When these two groups are analyzed separately, the losses for those betting against the parlay are estimated to be two to four times greater than the reported figure. In fact, over the past two weeks, these losses have exceeded those experienced at traditional sportsbooks.
The key takeaway from the ongoing discussion about prediction market parlays is that the reported figure encompasses two different types of bettors. Bloomberg's July report indicated that bettors had lost a net of $294 million on Kalshi's multi-leg “combo” markets—essentially a parlay that only pays out if all predictions are correct. This figure is accurate, as it reflects the losses of “takers,” or those who accept displayed prices, similar to how sportsbook customers accept odds on their bets.
Gambling Insider corroborated this calculation using the same public data, confirming that for every dollar wagered, approximately 9 cents were lost. However, the “takers” in the combo market are not a monolithic group; they are divided into two sides: the YES side, which bets that the parlay will succeed, and the NO side, which bets against it. Each side can be taken at the displayed price.
For the purposes of this analysis, a parlay buyer refers to someone betting on the YES side, similar to a customer at a sportsbook who risks a small amount for a potentially large return. Conversely, the NO side allows participants to bet against someone else's parlay, such as paying 95 cents for a contract that pays a dollar if a five-leg combo fails, thus earning a small profit.
In examining both sides over a two-week period from July 14 to July 27, we reviewed every recorded trade across approximately 4.6 million combo markets that settled during this timeframe. The findings reveal stark contrasts between the two groups.
Parlay buyers placed $226 million in bets and incurred losses of $79 million, translating to a loss of 35 cents for every dollar wagered. Even when excluding the World Cup final day—when buyers lost 86 cents for every dollar staked—they still lost an average of 16.6 cents per dollar over the remaining 13 days.
On the other hand, NO-side sellers staked $204 million, nearly matching the buyers, and achieved a profit of $43 million, resulting in gains of 21 cents for every dollar staked. These figures represent the same metric of profit or loss relative to the amount wagered, but one reflects a loss for the bettor while the other indicates a profit.
Interestingly, these two groups do not directly bet against each other; rather, they typically trade against professional market makers who set the prices. The results of these market makers are the missing component in this analysis.
The near equivalence in stakes between the two groups is noteworthy, especially considering that the product is marketed as a lottery-like opportunity for average bettors. Almost half of the total taker dollars are actually bets against these lottery tickets.
On five occasions during the 14-day period, including the day before the World Cup final, the NO side outperformed the buyers in terms of stakes.
Ultimately, when the two groups are combined, the overall loss for takers appears to be 8.4 cents per dollar, a figure that aligns with the year-to-date measurement. However, this blended figure can be misleading, as it mixes the losses of ordinary customers with the winnings of those betting against them, creating an illusion of stability.