Prediction-Market Traders Are Spread Evenly Across All Fifty States. The Industry Has Spent Two Years Saying They Are Concentrated in the Banned Ones

Editor's note: TrueEdge builds odds tools and earns affiliate commissions from licensed sportsbooks. This piece bears on a contested regulatory question. Weigh it accordingly.

Everyone who wrote about S&P Global's third-quarter consumer survey on Tuesday led with the same number, and it is the right number to lead with: 70 percent of prediction-market users have traded a sports event contract. Ten percent of American internet adults have used a yes/no exchange at all. Football is the business.

Two sentences further into the Kagan write-up there is a finding nobody picked up, and it does more damage to the industry's public case than the sports share does. Traders, S&P reports, "tended to be spread across all states, rather than skewed toward states where sportsbook gambling is prohibited."

Read that against the last two years of testimony. The core public-interest argument for sports event contracts — the one made to Congress, to the CFTC, in press releases and on panels — is that a federally regulated exchange serves Americans whose states have refused to legalise sportsbooks, pulling them out of offshore books and into a supervised venue. Texas. California. Georgia. The whole map of the unserved. It is a good argument, and it is the only argument in the industry's arsenal that is about consumers rather than about statutory construction.

The survey says the user base does not look like that. It looks like the country.

What the geography finding does and does not prove

Be precise about the claim. S&P is describing the distribution of traders relative to population, not an absolute headcount, and a single consumer panel is not a census. "Spread across all states" is a qualitative summary of a cross-tab we have not seen, and we could not see it: S&P's own research page returned a 403 to us three times — to WebFetch, to a browser user-agent and to a bare client — so every figure here comes through Casino.org's rendering, and we have not read the primary note. That is a real limitation and we are not going to bury it in a source line.

What the finding does establish, if it holds, is that the marginal prediction-market trader is not a Texan who had no other option. He is someone in New Jersey or Ohio or Michigan with four licensed sportsbook apps already on his phone who opened a fifth thing. That is a product competing inside the regulated market, not a product filling a hole outside it.

And it explains a number people have found puzzling. Twenty-one percent of respondents either bet with a traditional sportsbook or traded sports derivatives over the past year; 23 percent did one of those or traded a prediction market of any kind. The incremental contribution of the entire prediction-market category to the share of Americans betting on sport is about two percentage points. Most of these users were already betting. They are not new to the activity. They are new to the venue.

The number that cuts the other way, which most coverage got wrong

Two percent of respondents said they traded exclusively non-sports contracts over the past year. Several write-ups have treated that as "only two percent of users are non-sports," which is a misreading worth correcting, because the correct reading helps the exchanges.

Users are 10 percent of respondents. Exclusive non-sports traders are 2 percent of respondents. That is one in five prediction-market users who have never touched a sports contract — a fifth of the base trading Fed decisions, elections, weather and word counts and nothing else. That is not a rounding error and it is not a fig leaf. It is a real constituency, and any argument that event contracts are merely sports betting in a trench coat has to account for it.

It is also, in absolute terms, about one adult in fifty, against roughly one in five who bet on sport. Both of those things are true at once, and a writer who only tells you one of them is selling you something.

The strongest version of the industry's reply

Here it is, and it is not weak.

First, every exchange in American history began with a concentrated wedge. The Chicago Mercantile Exchange spent decades on butter and eggs. "Your users are currently concentrated in one product" describes a young venue, not a fraudulent one.

Second, the geographic finding may be measuring advertising rather than demand. These platforms have spent enormous sums on national television during live sport — Casino.org's write-up attributes the 42 percent of respondents who know the category but have not used it directly to that spend. A product advertised nationally acquires users nationally. That is a media-buying artefact, not evidence about who needs the product.

Third, and most seriously: none of this is what the courts are deciding. The Sixth Circuit ruled for Ohio and Tennessee on September 25 on whether a sports event contract is a "swap" under the Commodity Exchange Act, and Kalshi's petition for rehearing en banc went in on October 9, the last day of the window. The CFTC's October 9 proposal to define event contracts on sports, politics, culture and weather as swaps — paired with an interim final rule excluding state-licensed sportsbook wagers — is an argument about instruments. No judge is going to read a consumer panel and change a construction of the statute.

All three are fair. The third is the one we would press back on, because it concedes the point. If the user data is irrelevant to the legal question, then the industry should stop making the user data its public argument. It cannot run the unserved-states case in Washington and then disclaim consumer evidence when the consumer evidence arrives.

The second has a tell in it too. If national advertising explains the national user base, the advertising is the thing doing the work — and advertising is precisely what no federal rule currently constrains for an exchange, and what the reported FTC review is said to be examining only at the sportsbooks. Explaining your distribution by reference to your unregulated ad spend is not the exoneration it sounds like.

Where the volume data agrees and where it does not

Pew's Kaitlyn Radde, working from The Block's notional taker-volume series, put combined Kalshi and Polymarket monthly volume at $25.7 billion in May and $53 billion in July. Sports in July alone: $31.4 billion on Kalshi, $10.6 billion on Polymarket. Kalshi's second category was crypto at $6.1 billion. Politics — the thing these venues were founded on — ran about $726.8 million on Polymarket and $169.2 million on Kalshi.

The honest caveat about volume is that it counts both sides, counts market makers, and counts the same dollar turning over repeatedly; a professional quoting two-sided prices in a World Cup market generates billions and is not a retail bettor. That is exactly why the survey matters. It counts people. And the two instruments disagree in an instructive way: by dollars, non-sports is a sliver; by heads, it is a fifth. The gap between those is the market maker.

What would change our mind

A state-level cross-tab from S&P showing per-capita trader density materially higher in non-sportsbook states would retire this piece's thesis, and we would print that. So would a demonstration that the 2 percent exclusive non-sports group accounts for a meaningful share of non-sports volume rather than being a long tail of small accounts.

Until then the sentence stands, and it is the industry's own survey saying it: the people on these exchanges are not the people the exchanges said they were built for.