Blackjack is the one casino game where your decisions change the outcome. That makes it worth learning properly, and it makes the table rules worth reading before you sit down.

The rules #

You are dealt two cards and so is the dealer, one of theirs face up. Number cards count face value, picture cards count ten, and an ace counts one or eleven, whichever helps. The aim is a total closer to 21 than the dealer’s without exceeding it.

Your options each hand:

• Hit — take another card.
• Stand — keep your total.
• Double down — double your stake for exactly one more card.
• Split — with two cards of the same rank, separate them into two hands with a second matching stake.
• Surrender — where offered, forfeit half your stake and end the hand.

Go over 21 and you lose immediately, whatever the dealer then does. That rule is the entire house edge: you act first, so you can bust before the dealer plays at all.

A “blackjack” is an ace with a ten-value card on the first two cards. It should pay 3:2 — £15 on a £10 stake.

The rules that move the edge #

Blackjack is not one game. The same table name can carry very different maths.

If you remember one thing: check the blackjack payout before your first hand. A 6:5 table wipes out most of what basic strategy earns you. It is usually printed on the table felt or in the game’s rules panel, and a site that leads with 6:5 tables is telling you something about itself.

Basic strategy #

Basic strategy is the mathematically optimal decision for every combination of your hand and the dealer’s up card. It is not a system or a prediction — it is simply the play with the best expected value in each situation, worked out exhaustively. Deviating from it costs money, on average, every time.

The full chart covers every combination, but a handful of rules carry most of the value:

• Always split aces and eights. Two aces played as one hand is a poor 12; split, each starts on eleven. Sixteen is the worst total in the game, and two eights escape it.
• Never split tens. Twenty is a winning hand. Breaking it up to chase two is giving away a strong position.
• Stand on 17 or higher. Always.
• Stand on 12 to 16 when the dealer shows 2 to 6. The dealer is most likely to bust with a low card up; let them.
• Hit 12 to 16 when the dealer shows 7 or higher. The dealer is likely to make a strong total, so a weak one will not survive.
• Double on 11 against almost any dealer card.
• Never take insurance. More on this below.

Standing on 16 against a dealer’s 10 feels terrible and is still correct — both plays lose more often than they win, and standing loses slightly less. Basic strategy minimises losses as much as it maximises wins, which is why following it through a losing session is exactly when it matters.

Insurance is a bad bet with a reassuring name #

When the dealer shows an ace you may be offered insurance — a side bet, up to half your stake, paying 2:1 if the dealer has blackjack. It is presented as protecting your hand. It is a separate wager on whether the dealer’s hole card is a ten, and it pays worse than the odds of that being true. It carries a house edge of roughly 7% in a standard game.

Decline it. Every time. The same applies to “even money” on your own blackjack, which is the identical bet wearing a different label.

Side bets #

Perfect Pairs, 21+3 and similar side bets are offered alongside the main game and typically carry house edges of several per cent — many times the main game’s. They exist because blackjack’s low edge is unprofitable per hand, and side bets restore the operator’s margin. Playing the main game at 0.5% while feeding a 6% side bet each hand means you are no longer playing a low-edge game.

Card counting, and online #

Card counting works in principle because a deck rich in tens and aces favours the player, and it is not cheating. It is also irrelevant to almost every online player: software blackjack reshuffles a virtual shoe every hand, so there is no persistent count to track. Live dealer games use real cards but generally shuffle early enough to remove any advantage. Treat any product sold as an online counting system accordingly.

What blackjack costs you #

Even played perfectly, blackjack is a losing game over time — around 0.5% under good rules, which is roughly £5 per £1,000 staked. It is the cheapest entertainment on the casino floor, and it is still negative. Playing without basic strategy typically pushes the edge to 2% or more, so most of what you can control is simply not giving away the difference.

Set a loss limit before you play, keep stakes flat, and treat the 0.5% as the price of the session rather than an obstacle to be beaten. If it stops being entertainment, our safer gambling page lists free support available 24 hours a day.

Why the dealer’s up card decides your play #

Basic strategy looks arbitrary until you see the dealer’s bust rate. The dealer has no choices — they must draw to a fixed rule — so their up card determines how likely they are to go over 21. Approximate bust rates:

This is the whole logic. Against a 5 or 6, the dealer busts roughly two times in five, so standing on a poor total wins more often than drawing to it. Against a 10, they bust about one time in five, so standing on 15 loses about four times in five — hitting is bad, and standing is worse.

It also explains doubling. You double on 11 because you cannot bust and a ten-value card is the most common draw. You double on 10 and 9 against weak dealer cards for the same reason, and not against strong ones.

A hand, played through #

You stake £10 and are dealt 8 and 8. The dealer shows a 6.

Sixteen is the worst total in blackjack — too low to win often, too high to draw to safely. But you do not have to play it as sixteen. Basic strategy says split eights, always, and against a dealer 6 that is emphatic: you put out a second £10 and play two hands, each starting on 8 against the weakest card the dealer can show.

Say the first hand draws a 3 for 11. Double: another £10, one card. It comes a 9, for 20. The second hand draws a queen for 18. You now have £40 staked across two strong hands against a dealer who busts about 42% of the time.

The dealer turns over a 10 for 16, must draw, and takes a 9. Bust. Both hands win, returning £80 on £40 staked.

That sequence — split, double, win both — is not luck applied to a bad hand. It is what the strategy was for. It also loses sometimes, which is the part that makes people abandon it.

The complete basic strategy chart #

This is the whole of basic strategy for the standard online game: multiple decks, dealer standing on soft 17, doubling permitted after a split. Where a rule differs at your table, the notes underneath say what changes.

Read it as: find your hand down the left, the dealer’s up card across the top, play what the cell says. There is no judgement involved and no situation where a hunch beats it.

Hard totals (no ace, or an ace counting as one)

H hit · S stand · D double if allowed, otherwise hit

Soft totals (an ace counting as eleven)

Soft 18 is the hand people misplay most. Holding A,7 against a 9, 10 or ace is a losing 18, and hitting it is correct even though it feels like breaking a made hand — you cannot bust, and 18 is not good enough against those cards.

Pairs

P split · S stand · H hit · D double

Two entries deserve explanation. 5,5 is never split — it is a hard 10, one of the best doubling hands in the game, and splitting it turns a strong position into two weak ones starting on five. 9,9 stands against a 7, which looks inconsistent until you notice the dealer’s most likely total with a 7 up is 17, and your 18 already beats it; splitting risks a good hand to chase a marginal improvement.

Surrender, where offered

Late surrender forfeits half your stake and ends the hand. It applies to very few situations, and outside them it is a mistake:

• Surrender 16 against a dealer 9, 10 or ace.
• Surrender 15 against a dealer 10.
• Do not surrender anything else. Never surrender a pair of eights — split them instead.

Surrender exists because some hands lose more than half the time by any line of play. Giving up half a stake on a hand that loses roughly three-quarters of the time is the cheaper outcome.

What changes if the rules differ

• Dealer hits soft 17. Slightly worse for you. Double A,7 against a 2, and surrender 17 against an ace where allowed.
• No doubling after split. Do not split 4,4 at all, and split 2,2 and 3,3 only against 4 through 7.
• Single or double deck. Marginally better for you; a handful of borderline doubles become correct that are not in a multi-deck shoe.
• 6:5 blackjack. Nothing in the chart changes, and none of it matters much — the payout has already added roughly 1.4% and swamped the value basic strategy recovers. Find another table.

Why doubling and splitting carry so much of the value #

A player who hits and stands correctly but never doubles or splits gives away a substantial part of what basic strategy is worth. These are the only moments in blackjack where you can put more money on the table in a favourable position, and the game is close to break-even precisely because they exist.

Doubling

Doubling is a bet that one more card will win the hand more often than not. You double on 11 because you cannot bust and roughly a third of the deck is a ten-value card. You double on 10 for the same reason, except against a 10 or ace, where the dealer is too likely to already hold a strong hand.

The soft doubles are different in character. Doubling A,6 against a 5 will often produce a mediocre total — but you cannot bust, and the dealer is showing the card they bust with most often. You are not doubling because your hand is strong; you are doubling because theirs is weak. Understanding that distinction removes most of the discomfort.

Splitting

Splitting turns one hand into two, each with its own stake. It is correct when two separate hands are worth more than the combined one, and that is usually because the combined total is bad rather than because the halves are good.

Eights are the clearest case: sixteen loses more than any other total, while two hands starting on eight are ordinary. Aces are the other: A,A played as one hand is a soft 12 that will almost never make 21, while two hands each starting on eleven are the strongest opening in the game. Most tables deal only one card to each split ace and do not allow re-splitting, and it is still correct.

Blackjack variants, and what each one costs #

Online lobbies carry a dozen names that all look like blackjack. Most are the standard game with a rule changed and a compensating rule bolted on, and the compensation is rarely equal.

Blackjack Switch

You play two hands and may swap the second card between them — a genuine advantage, and a large one. It is paid for elsewhere: blackjack usually pays even money rather than 3:2, and a dealer 22 pushes against most player hands instead of busting. Those two changes are worth more than the switch, which is why the game exists.

Free Bet Blackjack

Doubles and splits on certain hands are free — the operator stakes them. In exchange, a dealer 22 pushes rather than busting. Since the dealer busts with 22 reasonably often, that push takes back more than the free doubles give.

Spanish 21

All the ten-spot cards are removed from the deck, which is significantly bad for the player, and a long list of bonus payouts is added to compensate. Whether it comes out ahead depends entirely on the specific paytable, and it demands its own strategy chart — the standard one does not apply.

Multi-hand and speed tables

Playing three or five hands at once does not change the edge on any individual hand. It multiplies how much money is exposed to that edge per round, and it accelerates the session. The same is true of speed tables, which are marketed as convenience and function as an increase in hourly cost.

Card counting, properly explained #

Counting deserves an honest description rather than either mystique or dismissal.

Why it works in principle

Blackjack is one of the few casino games where past cards affect future odds, because cards are removed from the shoe as they are dealt. A shoe rich in tens and aces favours the player: naturals are more frequent, doubles succeed more often, and the dealer — obliged to draw on stiff totals — busts more.

A counter tracks a rough proxy for that composition, assigning small positive and negative values to low and high cards, and raises stakes when the remaining shoe is favourable. The gain is small, a fraction of a percent, and it only appears over very large numbers of hands.

It is not cheating and it is not illegal. It is arithmetic applied to public information. Venues are also entitled to decline custom, which is why it tends to end in exclusion rather than prosecution.

Why it does not work online

• Software blackjack reshuffles every hand. The virtual shoe is reset before each deal, so no count persists. There is nothing to track.
• Live dealer tables shuffle early. Cards are real, but the shoe is not dealt deep enough for a count to become reliable before it is replaced.
• Stake variation is visible. Counting requires raising stakes sharply when conditions are good — the most detectable behaviour there is, in an environment that records everything.

Any product sold as an online counting system is selling something that does not function against a per-hand shuffle. Treat it accordingly.

Bankroll and session management #

Blackjack’s low edge is often taken to mean low risk. It means low cost; the variance is entirely separate, and it is what actually empties a bankroll in an evening.

At roughly 0.5%, £10 a hand and 60 hands an hour is £600 staked for an expected cost of about £3. That is the average. Swings of tens or hundreds of pounds either way inside that hour are completely ordinary, and doubling and splitting increase them further by putting more money out on individual hands.

• Keep stakes flat. Basic strategy assumes a constant bet. Raising after losses reintroduces the Martingale problem into a game that had solved it.
• Size to the swings, not the average. A bankroll set at a few hands’ worth will not survive an ordinary losing run, whatever the edge.
• Decide the session cost in advance. Roughly £3 an hour at the stakes above is the price of the entertainment. Set a loss limit that reflects it.
• Stop on time, not on result. “One more hand to get even” is chasing, and the edge does not care how far down you are.

The mistakes that cost the most #

1. Playing a 6:5 table. The single most expensive decision available, made before the first card.
2. Taking insurance. A roughly 7% side bet dressed as protection.
3. Standing on 16 against a 10 because it “feels safer”. It is not safer; it loses about four times in five.
4. Never doubling or splitting. Gives away most of what basic strategy is worth.
5. Splitting tens. Breaking up the second-best hand in the game to chase two ordinary ones.
6. Playing side bets every hand. Several per cent, attached to a game you chose for 0.5%.
7. Copying the rest of the table. Other players’ decisions do not affect your hand, and most of them are playing badly.
8. Abandoning the chart during a losing run. Precisely when it matters, and precisely when people stop.

If a session stops being entertainment, the free and confidential support on our safer gambling page is open 24 hours a day.

Insurance, worked through #

Insurance is declined in one line in most guides. It is worth doing the arithmetic once, because the same reasoning applies to every side bet you will ever be offered.

The dealer shows an ace and offers insurance up to half your stake, paying 2:1 if their hole card is a ten-value card. Presented as protection, it is a standalone bet on one question: is that hidden card a ten?

In a standard deck, sixteen of fifty-two cards are ten-value — four tens plus four each of jacks, queens and kings. That is roughly 30.8%. Excluding the ace already showing, the hole card is a ten a little under a third of the time.

So on thirteen £10 insurance bets you would expect to win about four and lose about nine. The four wins return £20 profit each, £80. The nine losses cost £90. You are down £10 on £130 staked — a house edge of roughly 7%, against the 0.5% you sat down for.

“Even money” is the same bet. When you hold a blackjack and the dealer shows an ace, you are offered a guaranteed even-money payout instead of the 3:2 that is at risk. Accept it and you take 2 units instead of a 3:2 payout you will keep roughly two-thirds of the time. The maths is identical to insurance, and so is the answer.

The general principle is worth carrying: any bet offered in addition to the main game exists because it is more profitable for the operator than the main game. That is the reason it is on the table.

Side bets, and what each actually costs #

Blackjack’s low edge is unprofitable per hand by casino standards. Side bets restore the margin, and they are placed alongside the main bet where they look like part of the same game.

Perfect Pairs

Pays if your first two cards form a pair, with larger payouts for suited pairs. The house edge varies with the paytable and deck count but typically runs several per cent — an order of magnitude above the main game.

21+3

Combines your two cards with the dealer’s up card and pays on poker hands: flush, straight, three of a kind, straight flush. Again several per cent, again varying by paytable.

The arithmetic that matters

Consider £10 a hand on the main game at 0.5%, plus £5 on a side bet at 5%. The main hand costs about 5p in expectation. The side bet costs about 25p. You have made the cheap game five times more expensive by adding a smaller bet to it.

Over 60 hands an hour, that is roughly £3 for the blackjack and £15 for the side bet. The side bet is now the game you are playing, and blackjack is what you do between side bets.

There is nothing wrong with playing one knowingly for the swings. The problem is that it is presented as a garnish on a low-edge game when it is the opposite, and almost no site scoring blackjack “0.5%” mentions that most players at the table are not paying 0.5%.

Composition-dependent plays #

Basic strategy is built on your total and the dealer’s up card. A small number of hands are played differently depending on which cards make up the total. These are marginal and worth knowing mainly because they explain why the chart sometimes seems to contradict itself.

• 16 against a 10. The chart says hit. But a 16 made of three or more cards — say 5, 5, 6 — already contains low cards that would have helped you. With those removed from the shoe, standing becomes correct.
• 12 against a 4. Stand normally. Composition can shift this in single-deck games, where removing specific cards has a much larger effect.
• Doubling on 11 against an ace. Depends on whether the dealer checks for blackjack before you act. Where they do, doubling becomes stronger, because the hands where they already hold a natural have been removed.

None of these is worth much — fractions of a percent, appearing rarely. Play the chart. The exceptions are here so that a reader who notices the tension knows it is real rather than an error.

The dealer’s procedure, and why it is fixed #

The dealer has no decisions. That is the structural fact the whole game rests on, and it is worth stating plainly because players often assume the dealer is playing against them tactically.

They must draw until reaching a fixed total — 17 in the standard game — and must then stop, regardless of what you hold. A dealer sitting on 17 while you show 20 cannot draw to try to beat you. A dealer on 16 must draw even when you are visibly busted.

Two consequences follow. First, the house edge does not come from dealer skill; it comes entirely from you acting first and being able to bust before they play. Second, the dealer’s up card is genuinely informative, because their behaviour from that card is entirely predictable — which is what the bust-rate table earlier on this page describes.

Hitting versus standing on soft 17

Where the dealer must hit a soft 17 rather than stand, the edge moves against you by roughly 0.2%. It is a small number that appears in every hand, and it is one of the first things to check in the table rules — usually printed on the felt or in the game’s information panel.

Live tables: procedure and etiquette #

Live dealer blackjack adds a human element and a set of conventions. None of it changes the odds, but knowing it removes friction.

• Bets close on a timer. Miss the window and you sit the round out. Do not rush a stake to make it — a hurried bet at the wrong size costs more than a missed hand.
• Decisions also have a timer. Time out and the software plays a default, commonly standing. That may not be the play the chart wants.
• Other players cannot hurt you. Each hand is settled against the dealer independently. The player who “took the dealer’s bust card” changed which cards came out, not your expectation.
• Bet behind means inheriting a stranger’s decisions. You get their result and none of their choices. If you would not let them play your hand, do not.
• Chat is public and the dealer is working. They cannot influence outcomes, and asking them to is misplaced.

Practising without staking money #

Basic strategy is a memory task and memorising it is genuinely quick — a few sessions rather than weeks.

• Use demo play. Free-play blackjack runs the same rules and the same generator. Play with the chart open until you stop consulting it.
• Learn in the order that matters: the hard totals first, since they cover most hands, then pairs, then soft totals. Soft hands are the rarest and the most misplayed, so they benefit most from deliberate practice.
• Test yourself on the uncomfortable ones. 16 against 10, soft 18 against 9, and 12 against 2 are where errors cluster, because all three feel wrong.
• Do not practise stake variation. There is nothing to learn there, and the habits it builds are the expensive ones.

One caution about demo play: it is useful for learning decisions and misleading about everything else. Playing without money removes the pressure that makes people deviate from the chart, and a demo balance that runs out costs nothing. The skill transfers; the temperament does not.

A losing session, and why the chart still held #

Basic strategy is easiest to believe after a winning run and hardest during the run that follows. A concrete example is more useful than reassurance.

Fifty hands at £10, flat stakes, 3:2 payouts, dealer standing on soft 17. Expected cost across £500 of stakes is about £2.50. Here is a session that finished £120 down.

Eleven of the fifty hands were the uncomfortable ones — a stiff total of 12 to 16 against a dealer 7 or higher. The chart says hit all of them. Nine busted. On its own that reads like the strategy failing, and it is exactly the point at which most players abandon it.

But the alternative was not “not busting”. Standing on those hands wins only when the dealer busts, and against a 7 or higher they bust about a quarter of the time. Hitting lost nine of eleven. Standing would have lost roughly eight of eleven, and won less when it did win, because a made 20 pays the same as a scraped 17.

Six hands were doubles. Four won, two lost — an ordinary outcome that netted about £40 and quietly offset a third of the session’s loss. Three were splits, including 8,8 against a 10, which lost both halves and felt like the worst decision of the night. It was still correct: sixteen against a ten is the worst hand in the game, and two eights are merely bad.

The session lost £120 against an expected £2.50. That gap is variance, and it is normal. Fifty hands is a tiny sample; the edge is a statement about tens of thousands. What basic strategy guarantees is not a winning night — it is that you did not add to the cost of a losing one.

The failure mode to watch for is the conclusion people draw at this point: that the chart “does not work here”, followed by standing on stiff hands, declining doubles, and taking insurance to feel protected. Each of those decisions raises the edge. The session was going to lose either way; those choices are how it loses more.

How blackjack compares to the rest of the floor #

Worth stating in one place, because blackjack’s advantage is real and frequently overstated.

The honest reading is that blackjack is the cheapest game available if you meet its conditions, and roughly a 2% game if you do not. A player at a 6:5 table taking insurance and playing 21+3 is paying more than someone betting Banker at baccarat and thinking about nothing at all.

That is a genuinely useful conclusion: if you do not want to memorise a chart, baccarat’s Banker bet at about 1.06% asks nothing of you and beats mediocre blackjack comfortably. Choosing the game that suits how you actually want to play is worth more than choosing the one with the best headline number.