Payout Reports: What State Revenue Data Actually Shows
Every state that licenses online casino games publishes the financial results, and those filings are the closest thing American players have to an audited scoreboard.
The short answer
- What a hold percentage is: win divided by handle. The share of everything wagered that stayed with the house.
- What it is not: the house edge. Rhode Island’s 2.4 percent online table hold sits far above the roughly 0.5 percent edge a competent blackjack player faces.
- Why the gap: imperfect play, side bets, a wider denominator and a blend of table games. None of it involves a rigged shoe.
- Best single source: Rhode Island, the only state that publishes handle and an iSlots versus iTables split together.
Which States Publish Numbers, and Where to Find Them
Seven jurisdictions currently host licensed online blackjack, and every one of them reports. They report different things, on different schedules, at different levels of detail.
State revenue filings are badly misread. Someone who wants to know whether the game is honest opens a revenue report, sees a hold figure many times larger than blackjack’s theoretical house edge, and concludes the deck is stacked. Nothing is stacked. The two numbers count different things against different denominators. What follows is a working guide to what the seven regulated states release, how to parse each line item, and where the honest limits sit for anyone comparing legal blackjack sites on evidence rather than marketing.
These filings cost nothing to read, arrive on a predictable calendar, and are submitted by operators under penalty of license action. That last point is the whole reason they matter.
The seven are New Jersey, Pennsylvania, Michigan, West Virginia, Connecticut, Delaware and Rhode Island. The agencies behind them are profiled across our regulator directory, and the market conditions in each are covered on the state pages.
Pennsylvania: the only complete game-type breakout
The Pennsylvania Gaming Control Board is the most useful source in the country for a blackjack player, because it is the only regulator that splits online casino revenue three ways every month. Its revenue reports page carries six report families, one of which is Interactive Gaming Revenue, in both PDF and Excel, filterable by fiscal year back to 2006-07. The board also issues a monthly press release.
The release covering July 2026, dated Aug. 19, 2026, put total iGaming revenue at $249,007,485 from 11 casinos, up 9.09 percent against $228,258,244 in July 2025.
| Pennsylvania line item | July 2026 |
|---|---|
| Total interactive gaming revenue, 11 casinos | $249,007,485 |
| Online slots | $199,243,118 |
| Online table games, blackjack included | $47,426,563 |
| Online poker | $2,337,804 |
| Tax on the month’s iGaming | $115,548,069 |
| Same month a year earlier, July 2025 | $228,258,244 |
Table games, blackjack included, thus accounted for roughly 19 percent of Pennsylvania’s online casino revenue, and slots produced four times as much. Nowhere else can you watch that ratio move month to month.
New Jersey, Michigan and Delaware: totals without the split
The Division of Gaming Enforcement issues a monthly press release and statistical summary plus a set of monthly internet gross revenue reports, posted four business days after the filing date at 2 p.m. Those filings show win by peer-to-peer and other authorized games, the tax calculation, promotional gaming credits wagered, and revenue by brand. Note where the line falls: poker is separated, and blackjack, roulette, baccarat and slots are pooled as “other authorized games.” July 2026 internet gaming win was $276.9 million, up 12.0 percent on $247.3 million a year earlier, with $1.87 billion year to date. None of it tells you anything about table play specifically. Market conditions are covered on our New Jersey page.
The Michigan Gaming Control Board reports monthly on iGaming and internet sports betting jointly, giving gross and adjusted gross receipts by operator across 15 authorized licensees. July 2026 iGaming gross receipts were $302.77 million against $290.65 million adjusted, released Aug. 18, 2026. Again, one iGaming line, no game-type detail.
Delaware, the first state to launch online casino play back in November 2013, posts an “iGames Monthly Proceeds and Distribution” file under the iGaming section of the Delaware Lottery site, organized by fiscal year since July 1, 2023 and archived to 2014. The published index shows no slots-versus-tables division. Three video lottery agents share one BetRivers platform, so the totals are small and a single month can swing on a handful of players.
West Virginia: weekly, which complicates everything
The West Virginia Lottery is the outlier on cadence. Interactive wagering figures go up weekly through the business portal at business.wvlottery.com, with reporting weeks running Sunday through Saturday. The Lottery also files a Monthly Report on Lottery Operations with the state Legislature, and the financial reports page carries audited year-end figures. Interactive wagering delivered $49.8 million in transfers in fiscal 2025.
Weekly reporting weeks do not align to calendar months
There is no clean monthly West Virginia total, so every “West Virginia did X in July” headline is somebody’s approximation from adding weeks together.
Rhode Island and Connecticut: the two that publish handle
Here is the distinction that decides what any of this is worth. Pennsylvania, New Jersey, Michigan and Delaware publish what operators won. They do not publish what players wagered. Without the amount wagered you cannot compute a hold percentage at all, no matter how carefully you read.
Connecticut fixes half of that. The Department of Consumer Protection publishes a monthly dataset, “Selected Online Casino Gaming Data,” on the state open data portal and through its gaming revenue and statistics page. The dataset fields are unusually complete: licensee, fiscal year, month ending, wagers, patron winnings, cancelled wagers, online casino gaming win or loss, promotional coupons or credits wagered, promotional deduction, total gross gaming revenue, and payment to the state. Because wagers are included, you can calculate a blended hold for the whole state. Because game types are not, you cannot isolate blackjack. Only two licensees exist, the Mashantucket Pequot and Mohegan tribal entities, so the DCP file is short and easy to work through.
Rhode Island is the only state that gives you both. The Rhode Island Lottery publishes iGaming revenue information by month and year on its financials page, alongside Annual Comprehensive Financial Reports, and the twice-yearly Revenue Estimating Conference presentations filed with the General Assembly carry fiscal-year detail. Background on the market is on the Rhode Island page.
Rhode Island reports handle separately for iSlots and iTables
That makes it the single place in the United States where an online table-game hold percentage can actually be computed, which is why a very small state carries disproportionate weight and why the numbers further down this page come from there.
The regulators behind these filings
Each agency below publishes the reports described above, and each page explains what that regulator can actually do to a licensee.
The Vocabulary: Handle, Win, Hold and Adjusted Revenue
Four of these five line items are estimates of player behavior. Only one of them is a hard fact, and it is not the one most readers quote.
Handle
Sometimes labeled amount wagered or simply wagers, it is the gross sum of every bet accepted. On an online blackjack table the system books the opening bet, the double, each split hand, insurance and any side bet as separate wagers, so handle counts the same dollar repeatedly as it is won back and rebet. Handle is not money lost. It is money cycled.
Gross gaming revenue
Also called win or gross receipts, it is handle minus what was paid back to players. This is the number reporters quote, and it deserves a blunt translation: the operator’s win is the players’ aggregate loss. When Pennsylvania announces $47,426,563 in online table game revenue for a month, that is $47.4 million that players collectively did not take home.
Hold percentage
Win divided by handle. It is the share of everything wagered that stayed with the house. It is the figure most often mistaken for the house edge, and the rest of this page is largely about why those two are not the same measurement.
Adjusted gross revenue
Subtracts allowable deductions, chiefly promotional credits, from gross revenue before tax applies. Michigan’s July 2026 gap between $302.77 million gross and $290.65 million adjusted, about $12.1 million, is the bonus play the state let operators write off. Connecticut splits this into two fields, promotional credits wagered and promotional deduction, which are rarely the same figure. What qualifies depends on the fine print of the offer, the subject of our guide to bonus terms.
Tax paid
The state’s cut of adjusted revenue. It is the only line in the whole exercise that is a hard fact rather than an estimate, because the money moved.
Hold Percentage Is Not the House Edge
This is the section nearly every site gets wrong, so take it slowly. A published payout or hold figure and a game’s house edge are two different measurements of two different things.
What Rhode Island’s numbers say
Across fiscal 2025, Rhode Island iGaming handle was $1,198,632,835 with $44,993,915 in net revenue. That total splits into $716,257,278 wagered on iSlots and $482,375,557 wagered on iTables. Run the division and online table games held about 2.4 percent, against roughly 4.6 percent for online slots. In the first quarter of fiscal 2026 the pattern repeated and widened slightly.
| Product | Handle, fiscal 2025 | Hold, fiscal 2025 | Hold, first quarter fiscal 2026 |
|---|---|---|---|
| Online table games, iTables | $482,375,557 | About 2.4 percent | Around 2.2 percent |
| Online slots, iSlots | $716,257,278 | Roughly 4.6 percent | Around 5.0 percent |
Two conclusions follow immediately. First, table games are by a wide margin the cheaper product per dollar wagered, holding roughly half of what slots hold. Second, 2.4 percent is nowhere near the 0.5 percent house edge a competent player faces on a decent blackjack game.
Why 2.4 percent sits above 0.5 percent
Four things account for the gap, and none of them involves a rigged shoe.
Most players do not play basic strategy
The sub-1-percent edge quoted for blackjack is the edge against a player making the mathematically correct decision on every hand. Deviate, and the edge climbs fast. Standing on soft 17, refusing to split eights against a nine, hitting a 12 against a dealer four, taking even money on a good hand: ordinary mistakes push a player’s effective edge to 2 percent or beyond without any help from the house. Aggregate hold measures what a whole population of players actually did, not what an optimal player would have done. Our page on house edge and RTP works through the arithmetic.
Side bets carry far heavier edges
Perfect Pairs, 21+3, Buster Blackjack and their cousins commonly run 4 percent to 8 percent, and sometimes worse. Those wagers land in the same handle bucket as your main bet. A table where 10 percent of the money goes on side bets can post a blended hold near 2 percent even if every main-bet decision at the table were flawless.
The denominators differ
House edge is conventionally quoted per initial wager resolved. Reported hold divides win by every wager the system accepted, including doubles, splits and insurance. Those extra wagers do not carry the same expectation as the opening bet, so mixing them into one denominator produces a number that is not the house edge and was never meant to be.
The reported figure is a blend of games
Rhode Island’s online table lineup was six live-dealer tables as of June 30, 2025, covering blackjack, roulette and speed baccarat, against 272 online slots. Roulette carries a structurally higher edge than blackjack, so a 2.4 percent blended table hold almost certainly conceals a blackjack-only figure below it. Anyone quoting the blended number as blackjack’s hold is quoting the wrong number. How those tables are dealt and streamed is covered under live dealer fairness.
One more trap, and it catches experienced readers
In a land-based casino, table hold means win divided by drop, the cash and markers exchanged for chips at the table, not the amount wagered. Brick-and-mortar blackjack hold routinely lands between 15 percent and 25 percent on that basis. It is a measure of how long a buy-in survives, not a house edge, and comparing it to 0.5 percent is meaningless. Online reporting uses actual wagers, which is why online hold numbers look so much smaller and why the two can never be set side by side.
Strip out everything the hold figure cannot support and one thing survives intact: the ordering. Tables hold roughly half of what slots hold, in fiscal 2025 and again in the first quarter of fiscal 2026, and that ranking is the part worth acting on. The absolute number is a behavioral artifact. The gap between the two products is not.
The measurements this page keeps apart
Hold, house edge and return to player get used interchangeably in casino coverage. These three pages take them one at a time.
What These Numbers Cannot Tell You About Fairness
Aggregate hold is a behavioral statistic dressed as a mathematical one, and reading it as proof of anything about the shoe gets the causation backwards.
Hold moves when the player population changes, when the game menu changes, when side-bet uptake changes, and when promotions push traffic between products. It barely responds to the underlying game math at all, because that math is fixed by an approved prototype and verified before launch.
A high table hold in a given month therefore means players played worse, bet more side action, or a few large accounts ran cold. It does not mean the shoe was fixed. The reverse is equally true: a quarter in which several high-stakes players ran hot drags the number down with nothing having changed in the game.
What actually establishes that a game deals fairly is a separate chain of evidence: laboratory certification of the random number generator, approval against a filed prototype, and ongoing audit. That chain is set out on our pages covering whether online blackjack is rigged and RNG certification. Revenue reports sit alongside it as a sanity check on scale, not as a substitute.
For a number that genuinely predicts what a game will cost you, read the rules on the table
Start with whether it pays 3-2 or 6-5 on a blackjack. That one rule swings the edge by roughly 1.4 percentage points, dwarfing anything you will infer from a state filing.
The Seven States Compared
The breakout column is the one that matters. Where table games are not separated from slots, a blackjack player learns nothing specific from the filing.
| State | Publisher | What is published | Cadence | Tables split from slots? | Handle published? |
|---|---|---|---|---|---|
| Pennsylvania | Gaming Control Board | Interactive gaming revenue, PDF and Excel, by operator; monthly press release | Monthly, archives to FY 2006-07 | Yes: iSlots, iTables and iPoker separately | No |
| New Jersey | Division of Gaming Enforcement | Press release, statistical summary, internet gross revenue tax filings by brand | Monthly, posted four business days after filing | No: poker split out, everything else pooled | No |
| Michigan | Gaming Control Board | Gross and adjusted gross receipts by operator, iGaming and sports betting together | Monthly | No | No |
| West Virginia | State Lottery | Interactive wagering figures; monthly report to the Legislature; audited annual reports | Weekly, Sunday to Saturday | No | No |
| Connecticut | Dept. of Consumer Protection | Wagers, patron winnings, win, promotional deduction, gross revenue and payment, by licensee | Monthly | No | Yes |
| Delaware | State Lottery | iGames monthly proceeds and distribution, by fiscal year | Monthly, archives to 2014 | Not shown on the published index | No |
| Rhode Island | State Lottery | iGaming revenue by month and year; annual audited report; revenue conference filings | Monthly, plus fiscal-year summaries | Yes: iSlots and iTables | Yes |
One state of seven, Rhode Island, publishes both the game-type split and the handle. That is the entire national supply of data from which an online blackjack hold percentage can be derived.
The Tax Gap That Shapes Your Lobby
Revenue reports become far more interesting once you know what each product is taxed at, because the tax rate explains operator behavior that otherwise looks arbitrary.
Pennsylvania taxes interactive gaming under 4 Pa.C.S. 13B52, and the rates are wildly asymmetric. Games that simulate slot machines are taxed at 52 percent of daily gross interactive gaming revenue, and with the local share assessment added the effective burden reaches 54 percent. Interactive table games and peer-to-peer games are taxed at 16 percent all in. An operator therefore keeps roughly 84 cents of every dollar won at an online blackjack table and roughly 46 cents of every dollar won on a slot.
Rhode Island’s statute is even starker, and it is worth reading because the allocations are itemized. Online slot revenue divides under R.I. Gen. Laws 42-61.2-5.1(b); online table revenue divides under 42-61.2-5.2(b).
| Rhode Island recipient | Share of online slot revenue | Share of online table revenue |
|---|---|---|
| The state | 61 percent | 15.5 percent |
| iGaming platform vendor | 22.55 percent | 35 percent |
| Game vendor | 15 percent | 48.5 percent |
| Host communities of Lincoln and Tiverton | 1.45 percent | 1 percent |
| Statute | R.I. Gen. Laws 42-61.2-5.1(b) | R.I. Gen. Laws 42-61.2-5.2(b) |
The state takes almost exactly four times as large a share of a slot dollar as of a table dollar.
Now put that beside the hold figures. Slots hold roughly twice as much per dollar wagered and, in Pennsylvania at least, generate revenue at four times the volume. Even after a punishing tax rate, slots remain the product the economics favor promoting. That is why free spins outnumber table-game offers in almost every lobby, why wagering requirements so often exclude or heavily discount blackjack, and why the table-game share of revenue is worth watching quarter over quarter. The reports let you see the resulting game mix directly rather than guessing at it, and our bonus overview explains how the same asymmetry shows up in the offers you are shown.
Where the tax asymmetry reaches you
The rate gap between slots and tables is the reason blackjack is treated the way it is in promotions and terms.
What Offshore Sites Publish
Nothing verifiable. That is not a rhetorical flourish, it is the literal state of affairs.
A state-licensed market
- Handle, win, hold and tax filed with a named agency under penalty of license action.
- Pennsylvania publishes what each of 11 licensed operators won on table games last month, and what each paid in tax, as a downloadable spreadsheet.
- Rhode Island itemizes the statutory allocation of every dollar, slot and table separately.
- Connecticut lists wagers, patron winnings, promotional deduction and payment to the state, licensee by licensee.
- Filings arrive on a predictable calendar and are archived for years.
An offshore brand
- No unlicensed operator serving American players files handle, win, hold or tax with any authority a player can query.
- No offshore licensing body publishes per-operator financial results.
- Some sites post an internal “payout percentage” figure with no stated methodology, no period, no game breakdown and no auditor named.
- Where a testing certificate is displayed, it typically attests that a game’s math was correct when it was submitted, a narrower claim than a monthly revenue filing and a much narrower one than most readers assume.
- What you can actually pull up is a marketing page.
Our comparison of offshore and licensed sites takes that difference through the other areas where it shows up, including complaints, reserves and withdrawal enforcement.
Running Your Own Quarterly Check
The whole exercise takes about 20 minutes a quarter once you know the release rhythm.
Timing
Most releases land in the second half of the following month. Michigan issued its July 2026 report Aug. 18, Pennsylvania issued its July release Aug. 19, and New Jersey posts its internet gross revenue filings four business days after the filing date at 2 p.m. West Virginia updates weekly rather than monthly, so check it whenever you like.
Fiscal years differ
Pennsylvania and Rhode Island run July through June, so a “fiscal 2026 first quarter” figure covers July to September 2025. Do not compare a fiscal quarter to a calendar quarter without adjusting.
Start with Pennsylvania for the mix
Divide online table game revenue by total iGaming revenue. If that share is drifting, something in the promotional environment changed.
Use Rhode Island for hold
Divide iTables revenue by iTables handle and compare it to the prior period. Do the same for iSlots. A stable gap between the two is the normal, healthy picture.
Use Connecticut for the promotional read
Compare promotional credits wagered against the promotional deduction and against total gross revenue. A rising promotional ratio usually precedes a change in what gets advertised.
Do not read one month as a trend
Table games are far more volatile than slots because a single large player can move a small state’s monthly figure by a full percentage point of hold. Two or three quarters is the shortest honest window.
Treat the result as market intelligence rather than as an audit of any individual game. It tells you where the money is going, what the states are collecting, and which products the tax code is pushing operators to promote. It will never tell you whether the last hand you were dealt was fair, and any page that claims otherwise is selling you something.
Where the Evidence Actually Comes From
Revenue filings are a sanity check on scale. The pages below carry the chain of evidence that establishes a game deals fairly, and the routes open to you when it does not.
The rest of the fair play section
Each guide takes one part of the fairness question and follows it to a primary source rather than to a marketing claim.
If a number on your own account looks wrong
Aggregate state data will never resolve an individual dispute. These pages cover what will.
If gambling has stopped being entertainment, help is available
The National Council on Problem Gambling has used 1-800-MY-RESET (1-800-697-3738) since Jan. 29, 2026. The older 1-800-GAMBLER line still connects and remains required in several states’ advertising rules.
Facts on this page were checked against regulator publications and statute text on Aug. 25, 2026. Revenue figures cited are the most recent published as of that date and are superseded each month.
