How to Win at Blackjack: The Best Strategy, in Order of Importance

The honest answer first

The best blackjack strategy is basic blackjack strategy played without error at a table with decent rules. That gets the house edge down to about 0.5% on a typical UK six-deck game. Nothing short of card counting turns that edge positive, and counting is hard to apply online. Everything else on this page is about not giving that 0.5% back.

That figure needs its rules stated: six decks, dealer stands on soft 17, double on any two cards, double after split allowed, no surrender, no hole card, blackjack pays 3:2. Change the rules and the number moves, which is why table selection is step one below.

A house edge is a long-run average. Over an evening you can win or lose many times the expected amount. No strategy guarantees a win on any hand, session or month. What a good strategy does is make sure that, over thousands of hands, you lose as little as the game allows.

The five steps below are ranked by how much money each one saves. The first two do almost all the work.

1. Pick the right table

Before you sit down, or before you open a game lobby, check the rules. The house edge of about 0.5% assumes the UK standard set above. Each departure from it has a price, and the table below puts a number on each one. All figures assume basic strategy.

Two things stand out. A 6:5 payout on blackjack costs more than every other bad rule combined, so that is the first thing to look for. And the "good" rules do not add up to a positive edge: a six-deck game with every favourable rule is still around 0.33% against you.

Decision: if the game pays 6:5, do not play it. If the dealer hits soft 17 and another game at the same stake stands, pick the one that stands. Single-deck games are only worth it at 3:2 with normal doubling; a single-deck 6:5 game is worse than a standard shoe game.

Online, the rules are in the game information panel. Read it. Two games from the same provider with the same name can have different soft 17 and surrender rules.

2. Play basic strategy on every hand

Basic strategy is the mathematically best decision for every player hand against every dealer up-card, worked out from the composition of the deck. It is not an opinion. The full chart for every rule set is on this site, and the UK version with no hole card adjustments is on the UK blackjack strategy guide.

The gap between basic strategy and the way most people play is the single biggest number on this page:

That 1.5 point gap is made entirely of small, individual decisions: standing on 16 against a 10 because "I'll probably bust", hitting 12 against a 4, not doubling 11 against a 10, refusing to split 8s against a 9. Each of those feels sensible and each one costs money.

Why one deviation a shoe matters: a six-deck shoe dealt to a single player gives roughly 40 to 50 hands before the shuffle. If you get one of those wrong every shoe, you are not playing basic strategy at 0.5%, you are somewhere on the way to 2%. At £10 a hand and 60 hands an hour, the difference between 0.5% and 2% is £3 an hour against £12 an hour in expected loss. The cost of a mistake is not the one hand; it is that the mistake repeats every time the situation comes up.

The hands people get wrong most often, with the correct play from the reference chart (six decks, S17, DAS, hole card game):

Decision: do not play for money until you can get through 100 hands on the blackjack strategy trainer without an error. If you are unsure of a hand at the table, the blackjack hand calculator gives the answer for the exact rules you are playing.

3. Bet flat and size to your bankroll

Betting the same amount every hand is the best blackjack betting strategy for anyone who is not counting cards. That is not a lifestyle preference; it follows from how expected loss is calculated:

Expected loss = house edge x total amount wagered.

A progression system (double after a loss, raise after a win, any of them) changes which hands carry the bigger stakes. It does not change the house edge on any hand, and the dealer's cards do not know what you bet. Over time, the total wagered is what sets the expected loss, and a progression only changes how bumpy the ride is on the way. The betting strategy page works through each system with ten-hand examples.

What flat betting lets you do is choose your stake so that normal variance does not empty your session budget early. A reasonable rule is 50 to 100 units per session, where a unit is your standard bet.

Why so many units? Because a run of five or six losing hands in a row is ordinary, and doubles and splits put two or more units on a single hand. Playing £10 hands on a £50 budget means a normal bad patch finishes the session in minutes.

Decision: decide the session bankroll first, then divide by 50 to 100 to get the stake. Not the other way round.

4. Skip side bets and insurance

Side bets are where the casino recovers what basic strategy takes away. The main game at 0.5% is one of the cheapest bets in the building. The side bets on the same table are not.

Insurance is worth a sentence on its own because it is offered on every hand where the dealer shows an ace, and the word "insurance" makes it sound prudent. It is a bet that the dealer's hole card is a ten. It pays 2 to 1, but in a six-deck game the true odds against a ten in the hole are roughly 2.2 to 1, which is where the 7% comes from. "Even money" on your own blackjack is the same bet under a different name and has the same edge. Unless you are counting and the true count is +3 or higher, decline it every time.

Decision: if a £1 side bet at 5% sits next to a £10 main bet at 0.5%, the side bet has an expected loss of 5p and the main bet 5p. You have doubled your expected loss for a tenth of the stake. Skip them.

5. Set a session limit and stop

Because the edge is a percentage of everything wagered, the way to reduce expected loss is to wager less: lower stakes, fewer hands, shorter sessions. The table shows what an hour costs on average at different stakes and speeds, playing basic strategy (0.5%) and playing casually (2%). Online and live dealer games often deal 100 or more hands an hour to a single player; a busy land-based table might manage 60.

Two points. First, the actual result in any hour will almost never be the expected figure; being £100 up or £100 down at £10 stakes is normal. Second, that is exactly why a limit matters: if you only stop when you are losing, you will stop more often on a loss than the average suggests. A time limit, a loss limit and a win target, all set before the first hand, are the only things in this section that are fully in your control.

Decision: write down three numbers before you play: the most you will lose, the point at which you will bank a win, and when you will stop regardless. Then stop at whichever comes first. The safer gambling page covers tools for enforcing this online.

What does not work

Each of these is popular, and each has a one-line reason it fails.

• Hunches. The cards do not know what you feel, and your feeling has no information in it that the chart does not already use.
• Riding streaks. Each hand is dealt from the same shoe with roughly the same composition; three wins in a row do not change the probability of the fourth.
• "The dealer is due to bust." The dealer's bust probability depends only on the up-card and the cards left in the shoe, not on the last five hands.
• Progression betting systems. Expected loss equals edge times total wagered, and no sequence of stakes changes the edge; see the betting strategy page for the ten-hand examples.
• Mimicking the dealer (hit to 17, never double or split). This gives up every advantage the player has and costs several per cent.
• Never busting (standing on every 12 or more). You hand the dealer every hand they would have made, and the edge climbs well past the casual-play 2%.
• Taking insurance on a good hand. It is the same 7% bet whatever you are holding.

The only method with a positive expectation

Card counting is the only legal way to play blackjack with the edge on your side, and it is worth saying plainly what it involves so you can judge whether it applies to you.

With Hi-Lo you add 1 for each 2 to 6, nothing for 7 to 9, and subtract 1 for each ten-value card and ace. Divide the running count by the decks remaining to get a true count. Each point of true count is worth roughly 0.5% to the player, and the edge turns positive at around +1 to +2 depending on the rules. You then bet more when the count is high, typically with a spread of 1 to 8 units in a six-deck shoe, and adjust a handful of plays (the Illustrious 18, such as standing on 16 against a 10 at a true count of 0 or above, and taking insurance at +3). The card counting page and the advanced strategy page cover the method.

What it needs: a shoe game with penetration of 75% or better (50% is not workable), a bet spread the table will tolerate, a bankroll large enough to survive the swings that a 1 to 8 spread creates, and the ability to keep the count accurately while holding a conversation.

Why it rarely works online: RNG games shuffle every hand, so there is no count to keep. Live dealer games either use continuous shuffling or cut the shoe at around 50%, which removes the deep-shoe hands where the edge appears. A continuous shuffler lowers the house edge by about 0.02% but it also deals more hands an hour and makes counting pointless. The honest summary is that for online play, basic strategy at a well-chosen table is the ceiling, and the rest of this page is how to reach it.

Frequently asked questions

Can you win at blackjack consistently?

Not with basic strategy alone. It brings the house edge to about 0.5% on UK six-deck rules, which means a small expected loss over any long stretch, with short-term wins and losses on either side of it. Consistent winning requires an edge, and the only legal source of one is card counting in a game that allows it.

What is the best blackjack strategy?

Basic strategy, played without deviation, at a table that pays 3:2 on blackjack and lets the dealer stand on soft 17. Flat bets sized to your bankroll and no side bets complete it. That combination is the best blackjack strategy available to anyone who is not counting cards.

Is online blackjack easier to win at?

The edge can be slightly lower online because rules are often better and you can read them before playing. Against that, online games deal far more hands an hour, often 100 or more, which raises the amount wagered and so the expected loss per hour. The online blackjack strategy guide covers the differences.

Is it better to play one hand or two?

Playing two hands at once doubles the amount wagered per round, so it doubles the expected loss per round. It also smooths variance slightly because the two hands share the dealer's result. If you want to play for longer on the same budget, one hand is the better choice.

How much can you win at blackjack?

In any single session, several times your stake in either direction is normal. Over the long run, with basic strategy at £10 a hand and 60 hands an hour, the expected result is a loss of about £3 an hour. There is no stake, system or table at which the long-run expectation is positive without counting.

Does the number of players at the table matter?

Not to the house edge. Other players' decisions do not change your expected result, even when they "take the dealer's bust card", because they are equally likely to take a card that would have helped the dealer. What a full table does change is speed: fewer hands an hour, which lowers your expected loss for the time spent.

Back to Blackjack strategy: basic strategy explained, with the full chart