The Evidence That Regulated Online Blackjack Is Not Rigged

Testing and approval come before launch, not after complaints

In every US state with a regulated online casino, a blackjack game is a piece of licensed equipment. It cannot be offered to the public until an accredited independent testing laboratory has examined the random number generator, the game logic and the payout rules, and until the state regulator has signed off on the submission. The laboratory verifies that the virtual deck contains what it claims to contain, that the shuffle is statistically indistinguishable from random, that the game pays what the rules screen says it pays, and that the software running on the live server matches the version that was certified. That last check matters more than people realize: regulators verify the software build by signature, so swapping in an altered version after approval is detectable.

How labs test a shuffle, what standards they apply and what a certificate actually covers is the subject of a separate page on this site. See RNG certification for the process and its limits. Regulators also require that the game keep a record you can inspect. In New Jersey, N.J.A.C. 13:69O-1.4(p) requires an internet gaming system to let a patron view the outcome and the resulting account balance change for the previous game, including a game that finished after an outage. N.J.A.C. 13:69O-1.2 requires the operator to publish a method for the patron to obtain account and game history, and requires the terms and conditions to spell out what happens if a patron is disconnected mid game. A rigged game is a bad idea in a system that is required to hand the player a receipt for every hand.

The house edge already wins, and it wins enormously

This is the argument that should end the discussion for anyone who looks at the numbers. A well dealt six-deck blackjack game where the dealer stands on soft 17, doubling after splitting is allowed and a natural pays 3:2 gives the house roughly 0.4 percent against a player using correct basic strategy. That sounds trivially small. It is not, because the edge does not apply to your bankroll once. It applies to every wager, and you make hundreds of them an hour.

The public revenue statistics show what that compounding does. The Nevada Gaming Control Board’s monthly gaming revenue reports track statewide win by game. The “Twenty One” line has sat around 15 percent for decades. That is not the house edge. It is the hold: the share of the money players bought in with that the casino keeps. A 0.4 percent per-hand edge turns into roughly 15 percent of buy-in because the average player recycles the same chips through the shoe many times over before leaving. The house is already converting about a sixth of everything walked in the door into revenue, legally, with published rules, at zero legal risk. Manipulating a shuffle to raise that number is the worst risk-adjusted decision available to a licensed operator.

What an operator would be risking

What regulators actually catch

The most useful evidence is not the theory, it is the enforcement record. State gaming boards publish their consent agreements and fines, and if licensed operators were quietly rigging blackjack, that record would show it. It does not. What it shows instead is a steady stream of compliance failures around the edges of the game. The Pennsylvania Gaming Control Board has fined the largest live dealer supplier in the US market three separate times over matters at its Philadelphia studio.

Date	What the regulator found	Penalty
March 22, 2023	An unlicensed employee dealt several games of blackjack in the studio (consent agreement with Evolution US, LLC)	$7,500
August 2024	An unauthorized man got into the live dealer studio, including employee areas	$7,500
Dec. 18, 2024	A power outage caused a server failure that stopped surveillance video from 14 tables being saved for nearly two days, even though the cameras kept running (fine approved)	$25,000

Not one of those cases is an allegation that a card came out wrong. They are staffing, physical security and recordkeeping failures, and in each case the regulator found them, published them and charged for them. That is a far better indicator of whether a market is honest than an absence of complaints would be.

Pennsylvania also produced a useful example of a genuine game math error. In a consent agreement over an incident on Feb. 29, 2012, International Game Technology admitted that four progressive slot machines went into operation at Rivers Casino in Pittsburgh before final board approval, carrying incorrect settings that made them pay at an elevated level for about four hours. IGT was fined $5,000, and the casino paid $11,570.63 to correct the resulting tax underpayment. The error ran in the players’ favor, it was caught, and it was disclosed. Configuration mistakes happen in every regulated market. They are not the same thing as a rigged game, and the system is built to surface them in both directions.

Why It Feels Rigged Anyway

Dismissing the suspicion as ignorance is both rude and wrong. The feeling that something is off comes from a set of real statistical properties of blackjack that almost nobody is briefed on before they start playing. Here they are with the numbers.

The dealer makes 20 or 21 about a quarter of the time

Ask a casual player how often the dealer busts and you will usually hear something between 40 and 50 percent. In a six-deck game where the dealer stands on soft 17, the dealer therefore ends on 20 or 21 roughly one hand in four, and busts a little more than one hand in four.

That gap between the expected 45 percent bust rate and the actual 28 percent is where a lot of the rigging suspicion is born. A player who believes the dealer should be busting nearly half the time will experience a perfectly normal shoe as an unbroken run of dealer miracles. Even the famous bust cards are weaker than their reputation: a dealer showing a 6 busts about 42 percent of the time, which means the dealer beats or ties you on that upcard more often than not.

Normal variance is far larger than the edge

The standard deviation of a single blackjack hand is about 1.14 times your bet, driven by splits, doubles and 3:2 naturals. Over a session of 400 hands, the standard deviation of your result is about 1.14 multiplied by the square root of 400, which is roughly 23 betting units. At $10 a hand on good rules, your expected loss over those 400 hands is about $12. Your one standard deviation swing is about $228 in either direction, and two standard deviations is about $456.

Put plainly: on a fair game with a tiny house edge, finishing a 400-hand session down $450 is an ordinary outcome. It happens on the order of one session in 40 from variance alone, which for someone who plays a few times a week is several times a year. Nothing has to be wrong for that to occur, and no amount of correct play prevents it.

Losing streaks are common, not evidence

Excluding pushes, a basic strategy player loses roughly 54 percent of resolved hands and wins roughly 46 percent. The edge is small because the wins are worth slightly more on average, thanks to 3:2 naturals and profitable doubles and splits. But the raw frequency of losing is what you feel, and it is above half.

Run that forward. In a 400-hand session with roughly 366 resolved hands, you should expect to hit an eight-hand losing run two or three times. A ten-hand losing run has a probability of about 1 in 500 per starting position, which works out to roughly a coin flip that you see one somewhere in a single session. Those streaks are not a malfunction. They are the arithmetic of a game you lose slightly more than half the time, and they will keep arriving no matter how the shuffle is generated.

The gambler’s fallacy, and the deck that owes you nothing

The belief that gets players into real trouble is that a run of bad results makes a good result more likely. In a game where the virtual shoe is reshuffled between hands, which is the standard for computerized blackjack, that is exactly false. Every hand is dealt from a fresh, complete shoe. The software has no memory of your last twelve hands, and neither does your bankroll’s future. Over 100 hands, a fair game will produce dealer bust rates anywhere from roughly 20 to 36 percent without anything being wrong.

Selective memory and the bad beat archive

Human memory encodes surprising and painful events far better than routine ones. You will remember the hand where you doubled an 11 into a 3 and the dealer turned over a 21. You will not remember the eleven hands that session where the dealer broke on a 15 and you collected without making a decision. Over a few thousand hands, the archive you carry around is a heavily edited highlight reel of your worst outcomes, and it feels like data. It is not a sample, it is a filter. Confirmation bias finishes the job: once you suspect a game is crooked, every dealer 20 is filed as evidence and every dealer bust is filed as noise. That is precisely why the only reliable check is a documented one, covered further down this page.

Your own mistakes cost more than most people admit

The 0.4 percent figure describes a player making the correct decision on every hand for the specific rule set in play. Very few people do that, especially at 400 hands an hour with a timer running. Common errors are expensive in a consistent direction: standing on 16 against a dealer 7 or higher, never splitting 8s against a big upcard, doubling too rarely, taking insurance, and hitting soft totals incorrectly. Typical recreational play gives up somewhere between 1.5 and 2.5 percentage points on top of the rule set. That is why casino management models the realized edge against ordinary table game players at roughly 2 percent rather than 0.4, and it is another reason nobody needs to touch the shuffle.

A 99.5 percent game still empties your account

The most useful number to internalize is not the house edge, it is the expected loss per hour, because that is what actually leaves your account. Expected loss per hour equals your bet size multiplied by hands per hour multiplied by the house edge. Speed is a full multiplier in that equation, and online blackjack is very fast.

Format	Typical hands per hour	Why
Casino floor, full table of six or seven players	60 to 80	Shuffling, chip handling, other players deciding
Casino floor, heads up with the dealer	150 to 200	No other players, but physical dealing still takes time
Live dealer online	40 to 70	Fixed betting timers, streamed dealing, multiple seats
Computerized online blackjack, one hand at a time	300 to 500	Instant deal, no shuffle, no other players, no chips
Computerized online blackjack, multi-hand	600 or more	Several boxes resolved per deal at full speed

Take a player betting $25 a hand on a good rule set with a 0.4 percent edge. On a full casino table at 70 hands an hour, the expected loss is about $7 an hour. The same player, the same bet, the same rules, playing a computerized game at 400 hands an hour, has an expected loss of about $40 an hour. Nothing about the game changed except the clock. Move that player to a bad rule set at 2.4 percent, and the figure becomes roughly $240 an hour.

Four Specific Beliefs, Checked Against the Numbers

The Real Culprit: The Rule Set You Did Not Read

Blackjack does not have a fixed return the way a slot machine does. A slot is built to a target payback percentage, and a laboratory certifies that number. Blackjack’s return is produced by the rule set plus the quality of your decisions, and the rule set is chosen by whoever is offering the game. Two blackjack tables sitting side by side in the same lobby, with identical artwork and identical shuffling, can differ by more than two full percentage points in what they cost you. Nobody has to cheat to build the expensive one. They just have to change the payout on a natural and hope you do not check.

6:5 blackjack is the single most expensive change in the game

When a natural pays 3:2, a $10 bet returns $15. When it pays 6:5, the same $10 bet returns $12. You are naturally dealt about one blackjack every 21 hands, so this is not an obscure edge case; it fires roughly 19 times in a 400-hand session, and each time it costs you 30 percent of what the win should have been.

Converted into house edge, moving a natural from 3:2 to 6:5 adds about 1.39 percentage points. On a base game holding 0.40 percent, that is a jump to about 1.79 percent, which is more than four times the cost of the same game with a proper payout. There is no strategy adjustment that recovers it, because the money is being removed from a hand you already won.

Some online and studio variants go further and pay a natural at even money. That adds about 2.27 percentage points. Where you see it, it is nearly always paired with a headline feature such as free doubles, free splits or a multiplier, and the even money payout is how the feature is being financed.

Dealer hits soft 17

If the dealer must draw on soft 17 rather than standing, the house edge rises by about 0.22 percentage points. This is the change players are least likely to notice, because it only alters the outcome of a specific subset of hands, and it also changes correct basic strategy in several spots. It is worth more than half the entire house edge of a good game, and it is now the default in a large share of online and live studio blackjack.

Doubling, splitting and surrender

Everything in this group works the same way: it removes options that are only worth using when you already have the advantage, so restricting them costs you money without ever appearing on screen as a loss. Removing double after split, restricting doubling to 10 and 11, forbidding a second split, forbidding resplit aces and omitting late surrender together add roughly half a percentage point, and the exact figures are in the table below. The one most players never notice is a dealer who does not check for a natural before you act, so that a doubled or split bet is lost in full to a dealer blackjack. That is common in European-styled online games and costs about 0.11 percentage points.

Deck count

Fewer decks favor the player, but the effect is much smaller than folklore suggests, and it is the rule most often used as a marketing hook to disguise worse rules elsewhere. Holding everything else constant, a single-deck game is worth about 0.46 percentage points better than six decks and eight decks about 0.02 worse. A single-deck game paying 6:5 is a substantially worse bet than a six-deck game paying 3:2, and single-deck 6:5 is one of the most common traps in the business.

What each rule costs, in one place

Rule change from the reference game	Effect on house edge	In your favor or against
Natural pays 6:5 instead of 3:2	1.39 percentage points	Against you
Natural pays even money instead of 3:2	2.27 percentage points	Against you
Dealer hits soft 17	0.22 percentage points	Against you
No doubling after a split	0.14 percentage points	Against you
Doubling limited to 10 and 11 only	0.18 percentage points	Against you
Doubling limited to 9, 10 and 11	0.09 percentage points	Against you
No resplitting, one split maximum	0.10 percentage points	Against you
Dealer does not check for a natural before you act	0.11 percentage points	Against you
Eight decks instead of six	0.02 percentage points	Against you
Two decks instead of six	0.17 percentage points	In your favor
One deck instead of six	0.46 percentage points	In your favor
Late surrender offered	0.08 percentage points	In your favor
Resplitting aces allowed	0.08 percentage points	In your favor
Hitting split aces allowed	0.19 percentage points	In your favor
Six-card Charlie pays automatically	0.16 percentage points	In your favor

These figures are standard results of combinatorial analysis of blackjack and have been stable in the literature since Peter Griffin’s work in “The Theory of Blackjack” and Don Schlesinger’s “Blackjack Attack.” They assume correct basic strategy for the specific rule set in play. A player using approximate strategy gives up more, typically another 0.5 to 2 percentage points depending on how rough the play is, which is a separate cost worth naming honestly.

A realistic bad game against a realistic good one

Neither of the rule sets below is invented. Both are live in the market right now, sometimes inside the same casino lobby.

Rule	Expensive game	Good game
Decks	Eight	Six
Natural pays	6:5	3:2
Dealer on soft 17	Hits	Stands
Double after split	Not allowed	Allowed
Doubling	Hard 10 and 11 only	Any two cards
Resplitting	One split maximum	Split to four hands
Resplit aces	Not allowed	Allowed
Surrender	Not offered	Late surrender
Approximate house edge	2.45 percent	0.24 percent

The gap is about 2.2 percentage points, which means the expensive game costs roughly ten times as much to play. At $25 a hand and 400 hands an hour, that is about $10,000 wagered per hour in both cases: an expected loss near $24 an hour on the good game and near $245 an hour on the expensive one.

Side bets are where the money really goes

The base game is one of the cheapest bets in a casino. The side bets bolted onto it are among the most expensive, and online blackjack interfaces are designed so that placing them is one tap away from placing your main bet. Exact figures depend on the deck count and the specific paytable, which vary between suppliers, so treat these as ranges rather than fixed values.

Side bet	Typical house edge	What it pays on
Insurance	About 5.9 percent on six decks	Dealer having a ten-value hole card under an ace
21+3	About 3 to 8 percent	Poker hand from your two cards plus the dealer upcard
Perfect Pairs	About 4 to 11 percent	Your first two cards forming a pair
Any Pair, common in live studio games	About 4 to 11 percent	Same as above, with studio-specific paytables
Buster Blackjack	About 6 to 8 percent	The dealer busting, scaled by cards drawn
Royal Match	About 4 to 7 percent	Suited first two cards, more for king and queen
Blazing 7s and similar seven-based bets	Around 11 percent	Sevens in your opening hand
Lucky Ladies	About 17 to 24 percent	A two-card 20, with premiums for queens
Bet Behind	Same as the hand you back	Another player’s hand, played by that player

Now run the arithmetic that the interface never shows you. A player betting $10 on the main hand of a good rule set at 400 hands an hour is expected to lose about $16 an hour. Add a $5 side bet with a 6 percent edge on every hand and you are wagering an extra $2,000 an hour at more than fourteen times the cost per dollar, for an additional expected loss of about $120 an hour. The small side bet is costing seven and a half times what the main bet costs. Players who feel their bankroll evaporating faster than the blackjack math should allow are usually right, and this is usually why.

Multiplier and feature variants

The live studio format has produced a family of blackjack variants built around a mandatory extra charge. In multiplier games of the Lightning type, every hand carries a fee equal to your main bet, and that fee funds the random multipliers applied to winning hands. In free double and free split formats, the casino covers the extra wager but the dealer pushing on 22 is what pays for it, which converts a large number of hands you would have won into ties. Both are legitimate, disclosed, certified games. Both cost meaningfully more per hand than plain blackjack on good rules. If you want to see how the specific mechanics work, this site covers Lightning Blackjack and Free Bet Blackjack in detail.

The bonus trap that makes it worse

One more structural cost belongs here, because players routinely misread it as the casino moving the goalposts. Blackjack is either excluded outright from bonus wagering requirements or weighted at around 10 percent of turnover at most regulated and offshore casinos, precisely because its house edge is so low. That means $100 wagered at blackjack may count as $10 toward a rollover, or as nothing at all. Players who accept a bonus, grind blackjack toward the requirement and then find their withdrawal blocked usually ran into the weighting table, not a rigged game.

Shoe Versus Reshuffle After Every Hand

A large share of the suspicion comes from a real design decision that nobody advertises: most computerized blackjack does not use a persistent shoe. The software builds a full virtual shoe of six or eight decks, deals your hand from it, discards the whole thing and builds a fresh one for the next round. Nothing carries over. This is standard across the major suppliers, it is disclosed in the game help files, and it is not rigging. It has three consequences worth stating precisely.

First, card counting is dead on arrival. Counting tracks the depletion of high cards from a shoe dealt down over many rounds; if the shoe is rebuilt every hand there is no depletion to track. That is a design choice, not something an operator hides. Counting itself is not a crime in the United States, and this site covers where the law actually stands on card counting.

Second, and this surprises people, it runs slightly in the player’s favor. Physical shoe games with a cut card carry a small structural cost to the player known as the cut-card effect, first described by Peter Griffin: because a fixed cut card ends the shoe after a number of cards rather than a number of rounds, more rounds get dealt out of card-rich, low-count stretches than out of favorable ones, shaving roughly 0.06 percentage points from the player in a typical six-deck game. Reshuffling after every round removes that entirely.

Third, and this cancels the second many times over, a game that never shuffles, never handles chips and never waits for another player deals three to six times as many hands per hour. Whatever fraction of a percentage point you gain per hand is dwarfed by the increase in hands.

Live Dealer Blackjack Is a Different Fairness Question

With live dealer blackjack, the shuffle is not a random number generator at all. Real cards are dealt from a real shoe by a real person in a studio, streamed to you. That removes the RNG question and replaces it with the same set of questions a physical casino faces: who handled the cards, who watched, and what is on tape.

Cards, shuffles and machine reading

Every table is also managed by a game control unit that encodes the video and links the physical table to the game server, so the shoe, the table and the server are certified as one instrumented system rather than as a video feed with a database bolted on. Angel Group’s scanning shoe technology and comparable systems built by the studio operators are standard equipment.

It is often claimed that live dealer blackjack runs on continuous shuffling machines. Mostly it does not. The common setup is a scanning shoe dealt down to a cut card and then reloaded, which is why live tables display a shuffle break and computerized tables usually do not. Some live tables and variants do use continuous shufflers, and unlimited-seat formats have their own dealing arrangements, so this is another item to confirm in the rules panel rather than assume. The practical consequence is narrow: a live shoe dealt to a cut card is theoretically countable, which is precisely why studios use deep-penetration limits, table maximums and betting timers that make it impractical, and why operators reserve the right to close accounts they believe are counting.

What happens if you disconnect mid hand

This is the single most common source of live dealer complaints, and the answer is not the same as in computerized blackjack. In a computerized game, the round is yours alone. If your connection drops mid hand, the game state is held server-side and the hand is normally waiting for you when you log back in, or it is auto-completed and the result posted to your account. Under New Jersey’s rules the system must let you view that outcome and the resulting balance change even when the game finished after an outage.

What Actual Misconduct Has Looked Like

Real cheating in online gambling exists. It is worth knowing what it looked like, because the shape of it is nothing like the shape of the suspicion most players carry.

The one case that really was cheating

The clearest documented case in the history of the industry is the superuser scandal at UltimateBet and Absolute Poker, two sites licensed by the Kahnawake Gaming Commission. Between 2003 and 2008, accounts with a modified client were able to see opponents’ hole cards. The commission fined Absolute Poker $500,000 in January 2008. It later fined UltimateBet $1.5 million and ordered roughly $22 million repaid to affected players. Findings released in September 2009 identified 117 user names operated through 23 accounts, tied the great majority of the devices and IP addresses used by those accounts to former World Series of Poker champion Russ Hamilton, and named a wider group of individuals involved.

The far more common problems, which are not rigging

For every rigged shuffle allegation that has ever been substantiated, there are thousands of substantiated complaints in these four categories. They are the real risk, and they are worth more of your attention than the RNG is.

If you have a live dispute, the mechanism that works is a documented one. In regulated states you can escalate past the operator to the regulator, and the regulator has the operator’s own logs. The process, the deadlines and the state-by-state contacts are covered on this site’s page about filing a complaint.

The enforcement pattern across states

State regulators publish what they find. Read enough of it and the pattern is consistent: self-exclusion failures, geolocation failures, know-your-customer failures, responsible gaming failures, licensing and surveillance failures. Michigan’s board, for instance, has spent the past two years running a sustained campaign against unlicensed offshore sites rather than against its own licensees, announcing on April 7, 2026 that it had issued cease-and-desist orders to 45 offshore operators in four months. If licensed blackjack were being rigged, the same agencies producing that volume of enforcement paperwork would have found it.

Two Pieces of Standard Advice That Do Not Work

Checking the published blackjack RTP

This is the most repeated bad advice on the subject. Blackjack does not have a certified return-to-player figure the way a slot does. A slot is built to a target payback and the laboratory verifies that number against the game’s math model. Blackjack’s return is produced by the rule set combined with how well you play, so the same certified game returns about 99.6 percent to a flawless basic strategy player and something closer to 97 percent to a typical one, on identical software. What the laboratory certifies is the generator, the shuffle, and that the game follows and pays the rules it declares. If a site advertises a headline blackjack RTP without stating the rule set it assumes, the number is marketing, not measurement. Read the rules panel instead; that is where the return actually lives.

Trusting “provably fair” as a synonym for honest

Some crypto-facing offshore sites offer provably fair verification. The genuine version works like this: before the hand, the site publishes a hashed server seed; you supply or accept a client seed; after the hand the server seed is revealed, and you can recompute the hash and the shuffle yourself to confirm the result was fixed before you acted. Implemented properly, that is a real and checkable guarantee, and in one narrow respect it is stronger than a certificate.

How to Check for Yourself, in About Ten Minutes

You do not have to take anyone’s word for any of this, including this page’s. Here is the practical sequence, in the order that gets you the most information for the least effort.

1. Open the rules panel before you place a bet

Every certified blackjack game has a help, information or rules screen, usually behind an “i” or a menu icon in the corner of the table. Open it first, every time, on every new table. You are looking for seven things:

1. What a natural pays. If it does not say 3:2, close the table.
2. Whether the dealer hits or stands on soft 17.
3. How many decks are in play, and whether the shoe reshuffles after every hand.
4. Whether doubling is allowed on any two cards, and whether you can double after splitting.
5. How many times you may split, and whether aces can be resplit or hit.
6. Whether surrender exists.
7. Whether there is any non-standard rule such as the dealer pushing on 22, a per-hand fee, or a cap on the maximum win.

2. Read the side bet paytables, then decide once

Each side bet has its own paytable in the same panel. The paytable is what determines the house edge, and suppliers ship several versions of the same named bet. Rather than deciding hand by hand while the timer runs, decide before you sit down whether you are playing side bets at all, and if the answer is yes, size them as entertainment spending rather than as part of your blackjack bankroll.

3. Verify the license, not the logo

A certification badge in a website footer is an image file. The check that means something is the regulator’s own list. Every state with regulated online casinos publishes the operators and platforms it has licensed, and you can search the name of the company that actually holds the license, which is often not the consumer brand on the door. This site walks through how to do that on the page about licensing, and covers what a lab certificate from a firm such as Gaming Laboratories International, eCOGRA, iTech Labs or BMM Testlabs does and does not guarantee.

Two badges deserve specific skepticism from a US player. A UK Gambling Commission or Malta Gaming Authority logo tells you nothing about the site you are on if that site is serving you in the United States, because neither authority licenses US-facing play and neither will take your complaint about it. And a logo is not a certificate: the document that matters names the specific game, the specific software version and the date, and a legitimate operator will produce it on request.

4. Pull your hand history

Regulated operators are required to give you access to your account and game history, and the New Jersey rules cited earlier are typical of the standard. In most apps it is under account, then history, then gaming or transactions. What you want is the per-round detail: the round identifier, the cards, your actions, the result and the balance before and after. Export it or screenshot it.

Two things this gives you. First, an actual record to argue from if something goes wrong, which is worth far more than a memory of what you think happened. Second, the ability to check the game against itself. If the rules panel says a natural pays 3:2 and your history shows a $10 natural returning $12, you have caught something specific and provable, and that is exactly the kind of complaint a regulator will act on.

5. Use demo mode, but know what it proves

Free-play versions are useful for two things: confirming that the interface behaves as described, and learning correct strategy for the specific rule set without paying tuition. This site’s free blackjack section covers practice play. Be clear about the limits, though. A demo game running on a different rule set from the real money table teaches you the wrong strategy, so check the rules panel in both. And a demo session cannot validate a real money random number generator, because it is not necessarily the same instance and, more importantly, no session you can personally play is large enough to detect a small bias.

6. Understand what your own sample can and cannot show

Players sometimes record a few hundred hands and compare the dealer bust rate to the expected 28 percent. It is a reasonable instinct, and it is worth doing for your own understanding, but the statistics are unforgiving. Detecting a shift of one percentage point in a rate near 28 percent with any confidence requires a sample in the tens of thousands of hands, and a manipulation subtle enough to be worth an operator’s risk would be far smaller than that. What a short sample can reveal is a gross error: payouts that do not match the stated table, a rule applied differently from the one described, a bet resolving in the wrong direction. Those are the things worth watching for, and they are the things that actually get found.

Red flags that are worth acting on

Offshore Versus Regulated, Stated Honestly

As of August 2026, seven states have live, regulated online casinos and therefore legal real money online blackjack: New Jersey, Delaware, Pennsylvania, West Virginia, Michigan, Connecticut and Rhode Island. Maine has passed an online casino law, LD 1164, structured around tribal exclusivity, but it has not launched. Nevada, despite the assumption that Las Vegas implies legal online play, licenses online poker only and has no regulated online casino or online blackjack. The minimum age is 21 everywhere it is legal. Everyone else in the country who plays online blackjack for money is playing at a site licensed somewhere else, most often Curacao or Panama, and regulated by no US authority.

On the narrow question this page is about, the honest answer has three parts.

For the legal picture rather than the fairness picture, see this site’s pages on whether offshore sites are legal and on how regulated and offshore sites compare.

The Bottom Line

Online blackjack at a licensed US operator is not rigged, and the evidence that it is not is stronger than the evidence for almost any other claim in this industry: mandatory pre-launch certification, auditable systems, mandated hand histories, an enforcement record full of small compliance failures and empty of rigged shuffles, and a business model where cheating would risk a license generating hundreds of millions of dollars a year in order to improve on an edge that already returns roughly 15 percent of every dollar players buy in with.

But the player who suspects something is wrong is not being paranoid, and should not be told they are. Something usually is wrong. It is just written on the rules screen instead of hidden in the code. A game paying 6:5, hitting soft 17, refusing double after split and offering a one-tap side bet at a 6 percent edge, dealt at 400 hands an hour, will strip a bankroll roughly ten times faster than the same-looking game two tiles over. That is not cheating. It is worse in a way, because it is entirely legal, fully disclosed, laboratory certified, and nobody has any obligation to stop you from walking into it.

Read the rules panel. Skip the side bets. Slow down. Those three habits are worth more than any assurance about a random number generator, and unlike the rigging question, they are entirely within your control.

This page is general information about how these games and their oversight work. It is not legal, financial or tax advice, and licensing rules change; verify anything that matters to you with the relevant state regulator.

Is Online Blackjack Rigged FAQ

Play Within Limits

Understanding the math does not make blackjack a safe way to make money, and the faster a game deals, the faster a plan can come apart. Set a loss limit and a session clock before you open a table, use the deposit and time limit tools every regulated operator is required to provide, and treat the total you expect to lose per hour as the price of the entertainment rather than as a target to win back. Chasing losses is the real risk at any blackjack table. Help is free and confidential at 1-800-GAMBLER. More on tools, self-exclusion and where to get support is on this site’s responsible gambling page.