Combo contracts fuel prediction‑market volume spikes, analysts say

In this photo illustration, apps for online prediction market sites are shown on an electronic device on Feb. 25, 2026, in Chicago, Illinois.

Scott Olson | Getty Images

The growing popularity of combo contracts on prediction markets is driving a sharp increase in trading volume, even though they represent a small share of overall platform activity. Combos—multiple contracts linked together that pay out only if all underlying outcomes occur—have surged in usage, resembling parlay bets in traditional sports betting.

Last month, combos accounted for more than half of notional volume on Kalshi, propelled by the start of the NFL season. The majority of these combos involve multiple sports contracts.

Polymarket, which expanded its U.S. exchange this year, also made combos a priority after its official May launch. Combos now constitute nearly half of daily volume on Polymarket’s U.S. platform, buoyed by the NFL season.

Large share of volume

Although combos dominate volume metrics, they are not the primary vehicle for most prediction‑market traders. This discrepancy stems from how the Commodity Futures Trading Commission (CFTC) requires exchanges to report volume.

“The multi‑leg contracts are ‘helping [Polymarket and Kalshi] claim a higher number with lower realistic cash output,’” said Chris Park, a researcher and founder of MSR Decode, a research firm analyzing prediction‑market data.

On any prediction‑market trade, platforms count the position as $1 of notional volume regardless of the cash deposited. Each binary contract has an outcome value between $0 and $1, and a trade’s full $1 is counted when someone takes the opposite side.

Combos can have payouts far exceeding $1, so a trader can deposit only pennies while the recorded notional volume can reach thousands of dollars as market makers absorb the opposite side of the high‑payout bet.

A CNBC analysis of Kalshi trades on Sept. 27 illustrated this effect. For single contracts, the average cash outlay was less than 47 cents. For combo contracts, the average cash placed was about 9 cents.

Similarly, while combos made up 58 % of Kalshi’s trading volume in September, they represented less than 13 % of total transactions, according to Dune data. Other than sports, combos dominate other categories on Kalshi, accounting for over 35 % of total notional volume.

The methodology behind combo volume reporting is not always transparent, but the inflated headline numbers can make platforms appear to grow faster than the underlying cash activity suggests.

Apples and oranges

Some observers caution that notional volume is not directly comparable to the “handle” used by traditional sportsbooks. The handle reflects the total amount of money wagered, whereas prediction‑market volume counts both sides of each trade.

“I have spoken to investors who see the notional number and compare it to the handle in sports betting, and then make the argument that these prediction markets have already exceeded the scale of sports betting,” said Ian Moore, an analyst at Bernstein.

He explained, “When a client buys a deep out‑of‑the‑money contract or a combo, prediction markets may show dollar volumes 20 to 100 times greater than sportsbooks because CFTC reporting requirements differ from sportsbook reporting. Same basic bet, same basic idea, but you’re getting a completely different perception of activity.”

Kalshi’s spokesperson Jack Such responded, “Our notional volume does not overstate activity. If people are more interested in other measurements, they can look at other metrics.”

While notional volume is the standard metric for prediction markets, some argue that taker volume—a measure of how much liquidity traders remove from the order book—provides a clearer picture of actual trading activity.

Polymarket agrees: “We really like to look at taker volume in terms of buying ‘yes’ on combos,” said Kyle Gesuelli, the company’s head of revenue and analytics. “That feels like a better measure for us to truly understand the underlying activity.”

Source link