Published on February 25, 2026
Blackjack carries the lowest house edge of anything on a casino floor — under the right conditions, around half a percent. Those conditions are printed on the table itself, and most players never read them. This is a guide to which rules change the number, how much basic strategy is worth, and which widely repeated tactics do nothing whatsoever.
Two things are worth saying before anything else. Blackjack is still a negative-expectation game: correct play minimises the house edge, it does not reverse it. And the difference between a good blackjack table and a bad one is larger than the difference between a good player and a bad one.
If a table pays 6:5 on a natural blackjack instead of the traditional 3:2, it adds roughly 1.4% to the house edge — more than every decision you make in a session combined. A $10 blackjack pays $15 at 3:2 and $12 at 6:5. It is the single most expensive rule in the game, it is increasingly common, and it is written on the felt in plain language. Checking for it takes two seconds and is worth more than any strategy on this page.
The rules matter more than your play
Blackjack has been solved by exhaustive computation, so the cost of each rule variation is known rather than estimated. These are the standard published approximations, expressed as the change to the house edge:
| Rule | Effect on house edge | What it means for you |
|---|---|---|
| Blackjack pays 6:5 rather than 3:2 | about +1.4% | The worst rule in the game. Walk away from the table. |
| Eight decks rather than one | about +0.5% | Fewer decks favour the player; most online tables use six or eight. |
| Dealer hits soft 17 | about +0.2% | Printed on the felt as "dealer hits soft 17" or "stands on all 17s". |
| Doubling after a split not allowed | about +0.14% | Small individually, but these restrictions cluster together. |
| Late surrender offered | about −0.08% | In your favour, and rarely used by players who have it. |
| Re-splitting aces allowed | about −0.08% | In your favour. Uncommon. |
Stack the bad rules together and a game advertised as blackjack can carry a house edge above 2% — worse than many slots. Stack the good ones and the same game runs near 0.5%. You choose which table to sit at before you make a single decision, which makes it the highest-leverage choice available.
Basic strategy is arithmetic, not a system
Basic strategy is the mathematically optimal decision for every possible hand against every dealer upcard, computed from the rules in force at that specific table. It is not a betting system and it makes no prediction about what comes next. It simply answers, for each situation, which of hit, stand, double, split or surrender loses the least over time.
Playing it correctly is worth roughly 1.5% to 2% against playing on instinct — comparable in size to the 6:5 rule, and entirely under your control. A few of its instructions feel wrong at the table, which is exactly why they are worth knowing:
- Always split aces and eights, including eights against a dealer's ten. Sixteen is the worst hand in blackjack; two hands starting at eight are less bad.
- Never split tens. Twenty is already a winning hand. Breaking it up to chase two more is how a strong position becomes two mediocre ones.
- Stand on hard 16 against a dealer 2 through 6. The dealer is likely to bust; your job is to be present when it happens.
- Hit soft 17. An ace-six cannot bust on one card, and standing on seventeen against a strong dealer card wins far less often than it feels like it should.
- Decline insurance. It is a side bet on the dealer holding a ten, and it is priced against you. This holds even when you are holding a blackjack yourself.
Correct play depends on the table rules, so a chart for a single-deck game where the dealer stands on soft 17 is not the right chart for an eight-deck game where the dealer hits it. Match the chart to the table.
Side bets undo the whole point
Perfect Pairs, 21+3, Lucky Ladies and the rest sit beside the main game offering payouts that look like the reason you came. They typically carry a house edge between 3% and 6% — several times the edge on the hand itself.
The arithmetic is unkind. Playing perfect basic strategy on a good table and then placing a side bet every hand produces a worse expected result than playing badly with no side bet at all. If you have gone to the trouble of finding a 3:2 table, do not hand the difference back one chip at a time.
Card counting: what it is and where it does not work
Card counting is real, and the way it is usually described is not. It does not involve memorising cards. It tracks a single running number that rises as low cards leave the shoe and falls as high ones do, because a deck rich in tens and aces slightly favours the player. The counter then bets more when the count is high.
Even executed perfectly it yields an edge of roughly half a percent to one percent — enough to matter across thousands of hands and a large bankroll, and worth nothing across an evening. It is a job, not a trick.
It does not work in online blackjack at all. An RNG blackjack game reshuffles a virtual shoe before every hand, so no information carries from one hand to the next and there is nothing to count. Live-dealer tables mostly use continuous shuffling machines, which reintroduce played cards immediately and achieve the same thing. In a physical casino counting does work, and casinos are permitted to refuse service to anyone they suspect of it.
Edge sorting, and why the Ivey case settles it
Edge sorting exploits tiny asymmetries in the pattern printed on the backs of some cards, which lets a player who can get the deck oriented a particular way identify certain cards before they are turned. It is often presented as a clever legal edge.
It has been tested in court twice, and the outcome was the same both times. Phil Ivey won approximately £7.7 million at Crockfords in London and about $9.6 million at the Borgata in Atlantic City using the technique. Crockfords refused to pay; the UK Supreme Court ruled against Ivey in 2017, holding the technique to be cheating. A US federal court reached a comparable conclusion in the Borgata case and Ivey was ordered to return the money.
So the honest summary is this: a world-class poker professional executed the technique successfully against two casinos and was left with nothing but legal costs. Whatever edge sorting is, it is not a strategy an ordinary player can use to win money and keep it.
Three tactics that do nothing
Dealer tells. A blackjack dealer follows a fixed, published rule: draw to sixteen, stand on seventeen. There is no judgement to read, because there is no judgement being made. In a hole-card game the dealer has not looked at the down card either, so there is no information behind the body language even in principle.
Betting progressions. Doubling after a loss, raising after a win, and every variant in between only rearrange the order in which your money moves. The house edge applies to each bet independently of its size, so no sequence of stakes changes the total. Progressions produce many small wins and occasional catastrophic losses, which is why they feel effective right up until they are not.
Hot and cold tables. A table that has been "running well" has no memory of it. Streaks are what randomness looks like from the inside; they are visible only in hindsight and they carry no information about the next hand.
All three share the same defect: they are patterns imposed on independent events. The psychology behind slot machine wins examines why the human brain manufactures those patterns so insistently, and why the tendency survives knowing about it.
Bankroll at the table
Blackjack's low edge is only worth having if you are still at the table when it expresses itself, and hands arrive far faster than most players estimate — an online table can deal several hundred an hour. A session bankroll of at least forty times your base bet keeps a normal losing run from ending the session, and doubling and splitting mean your actual exposure per hand is sometimes twice the base bet, so plan around that rather than around the minimum.
The longer piece on bankroll management covers how to structure this properly.
Question-Answer Section: Blackjack Odds Explained
Does basic strategy let me beat the house? No. It minimises the house edge to roughly half a percent on a well-ruled table, which is the lowest available anywhere in the casino, but the edge remains against you. Basic strategy is how you lose the least, not how you win.
Why is 6:5 blackjack so bad? Because a natural blackjack pays $12 on a $10 bet instead of $15, and naturals occur often enough that the difference adds around 1.4% to the house edge — more than every playing decision in a session combined. It is the one thing to check before sitting down.
Can I count cards in online blackjack? No. RNG blackjack reshuffles the virtual shoe before every hand, so no information survives from one hand to the next, and live-dealer tables mostly use continuous shuffling machines that do the same. Counting only functions where cards are dealt from a shoe that is not reshuffled.
Should I take insurance when I have blackjack? No. Insurance is a separate bet on the dealer holding a ten, and it is priced in the casino's favour regardless of what you are holding. Taking “even money” on your own blackjack is the same bet under a friendlier name.
Are blackjack side bets worth playing? Not if the result matters to you. Side bets typically carry a house edge of 3% to 6% against roughly 0.5% on the main hand, so betting one every round costs more than the advantage you gained by finding a good table and playing correctly.
Discipline at the table is one half of the story; what happens to the money afterwards is the other. Our look at the myths and legends of famous casino heists traces what happened to people who decided the odds were not worth respecting at all.
And if you would rather your hands were dealt on infrastructure that is not quietly burning through electricity, our piece on eco-friendly gambling and sustainable practices in online casinos looks at what operators are actually changing.