New York bettors staked more than $2.4 billion in September, and the nine sportsbooks taking those bets collected less gross revenue over the half-year than they did a year earlier. Both things are true at once, and the gap between them is the real story behind the new New York sports betting record.
The New York State Gaming Commission numbers show September handle up 5% against September 2025, an increase of nearly $114 million. Calendar year to date through September, volume is up 3.7%, roughly $697 million more wagered than over the same nine months of 2025. The fiscal year looks stronger still: across the six months from April, handle reached almost $12.6 billion, a 6.9% rise.
Why the New York sports betting record didn’t pay off for operators
Gross sportsbook revenue over those six months fell 5.8% to $1.1 billion, because the average win percentage dropped from 10% to 8.8%. Bettors simply did better. Handle is what gets wagered; hold is the slice the books keep, and hold moved against them.
Run the arithmetic and the scale of it is clear. A 1.2-point drop in hold is a 12% relative decline in the share of every dollar retained. Volume grew 6.9%. Twelve percent down on the margin beats seven percent up on the turnover, which is why the top line shrank despite a record month inside that window.
That matters more in New York than almost anywhere else, because New York taxes mobile sports betting gross revenue at 51%, the highest rate in the United States. The state’s cut is levied on what the books keep, not on what players wager. A record handle paired with a softer hold means less tax money, not more. Roughly $1.1 billion in gross revenue at that rate implies state receipts in the region of $560 million for the six months, down with the operator line rather than up with the volume line.
Is a one-point hold drop skill, pricing, or just luck?
Mostly it is results. Six months is long enough to look like a trend and short enough to be dominated by which favorites covered and which parlays landed. A sportsbook’s theoretical margin is set by the prices it posts and the bet mix it attracts; the actual win percentage wanders around that theoretical figure, sometimes for a long time.
The honest answer is that the published data cannot separate the causes. Competitive pricing, heavier promotional credit, a shift in what people bet on, and plain variance all push hold in the same direction, and the monthly reports do not break them out. Anyone telling you with confidence that New York’s margin compression proves books are cutting prices to defend share is going beyond what the numbers permit.
Are prediction markets really siphoning off betting volume?
New York’s handle growth is the single most inconvenient data point for the argument that prediction markets are draining regulated sportsbooks. The largest sports betting state in the country, home to Bally Bet, BetMGM, BetRivers, Caesars Sportsbook, DraftKings, theScore, Fanatics, FanDuel and Resorts World Bet, is putting up record volume at the same time federally regulated sports contracts are growing.
The American Gaming Association frames those platforms as illegal sports betting operating outside state and tribal regulatory frameworks, bypassing licensing, consumer protections and taxation. Its running ticker puts projected lost state gaming tax revenue at close to $1.6 billion since prediction markets began offering what the AGA calls back-door sports betting in late 2025. The association’s position is blunt: prediction markets contribute nothing to the schools, hospitals and infrastructure that gaming taxes fund.
The platforms answer that their users trade against each other rather than against a house, which is the entire basis of the federal regulatory argument. That dispute will be settled in courts and by regulators, not by a handle report.
What the state-by-state numbers do and don’t prove
Three large markets, three different directions, and no tidy conclusion. Note the reporting periods differ, so these are not strictly like-for-like.
| State | Period | Handle | Year-over-year change |
|---|---|---|---|
| New York | January to September 2026 | +$697m vs 2025 | +3.7% |
| Illinois (online, pro and college) | Through August 2026 | $9.3bn (from $8.8bn) | +5.7% |
| New Jersey | Through September 2026 | $8.3bn (from $8.6bn) | -4.1% |
If prediction markets were pulling volume out of regulated sportsbooks in any uniform way, the two largest states would not be growing while the third contracts. New Jersey’s 4.1% decline is real, but New Jersey has been live since 2018 and has mature per-capita participation, a different promotional environment, and its own tax and operator dynamics. Attributing that fall to prediction markets requires an assumption the data does not supply.
The reverse is also true. New York’s growth does not prove prediction markets are harmless. Substitution can happen at the margin inside a growing market: handle could be up 5% when it would otherwise have been up 8%, and nothing in a state gaming report would reveal the difference. Both sides are arguing about a counterfactual nobody can observe.
What this changes for people actually placing bets
Short term, a lower industry hold means bettors in aggregate kept more of what they staked over those six months than they did a year earlier. That is a description of outcomes, not a strategy. The house edge is built into the prices, and over a long enough run the posted margin, not a six-month wobble, is what determines results. No sequence of favorable months changes that.
The more durable consequence is competitive. Operators facing flat or falling revenue on rising volume have limited levers: tighten promotional spend, push higher-margin parlay products harder, or accept thinner margins to hold share. Each of those is visible to customers, usually in the form of smaller bonuses and more aggressive same-game parlay marketing. In a 51% tax state, that pressure is sharper than almost anywhere else.
There is also a question of where the competition is allowed to happen. Thirty-nine states run legal sports betting under state licensing, with age verification, deposit limits, self-exclusion registers and advertising rules attached. Federally regulated event contracts sit under a different framework. Readers choosing between the two are choosing between consumer-protection regimes, and that distinction deserves more attention than it usually gets.
The part nobody can settle yet
Handle tells you how much money moved. It does not tell you where it came from, who it was taken from, or whether it would have been larger under different rules. New York’s record September is strong evidence that the regulated market is not collapsing. It is weak evidence about prediction markets either way, and anyone presenting it as proof is selling a position.
The number worth watching next is not the handle headline. It is whether hold reverts toward 10% as results normalise. If it does, the revenue decline was variance. If it sits near 8.8% through another half-year of growing volume, something structural has changed in how New York’s nine sportsbooks price and promote, and the competition question gets a lot more interesting.
If betting stops being entertainment and starts being a way to chase losses, use the deposit limits, cool-off and self-exclusion tools every licensed New York operator is required to offer, and contact a problem gambling helpline.