Blackjack Basic Strategy, Hand by Hand

What is blackjack basic strategy?

Hold a hand of 16 against a dealer showing a 10. Hitting risks busting on any card above a five. Standing means losing to every dealer total from 17 upward. Both feel wrong, and that is the point: one of them is simply less wrong, and the difference can be measured to several decimal places.

That measurement is what basic strategy is. Take every player hand that can exist, pair it with every dealer upcard, and compute the expected value of hitting, standing, doubling, splitting and surrendering by following each branch through every possible sequence of remaining cards. Choose whichever number is highest. Repeat until the grid is full. The work was first done with mechanical calculators in the 1950s and has been repeated with computers many times since, always with the same result, because the rules have not changed.

This makes blackjack unusual on a casino floor. Roulette systems do not alter expected value, because there are no decisions to optimise. Baccarat road maps record history without predicting anything. Basic strategy is the genuine exception on this site, and it is worth stating plainly what kind of exception it is. It is not a way of winning. It is the cheapest way of playing, and the gap between the cheapest way and the intuitive way is large enough to matter over an evening. The house edge comparison puts the resulting figure alongside the other games on the floor.

The chart is read by finding your hand down the side and the dealer card across the top

Three tables follow: hard totals, soft totals and pairs. Look up your total in the left column, run across to the dealer upcard, and the cell tells you what to do. Pairs are checked first, then soft hands, then hard hands.

The tables below assume the common live dealer configuration: six or eight decks, dealer stands on soft 17, doubling permitted after a split, and split aces receiving one card each. That is the standard shape of a multi-deck table, and the variations section later covers what moves when a table deals differently.

Hard totals split down the middle at the dealer seven

The vertical line between the dealer 6 and the dealer 7 is the most important feature of the whole chart. A dealer showing 2 through 6 must keep drawing from a weak start and busts roughly four times in ten from a 5 or a 6. Against those cards you stand on anything from 12 upward and let them break. A dealer showing 7 through ace reaches a standing total comfortably, so a stiff hand of 12 to 16 has to be improved even though improving it often fails.

Totals of 9, 10 and 11 are the doubling band. You cannot bust on the next card from 9 or 10, and 11 turns into 21 whenever a ten-value card lands, which is nearly a third of the pack. Doubling in those spots is where basic strategy makes most of its money back.

The 12 against a dealer 2 or 3 cell surprises people. It is close, and it goes to hitting because a dealer 2 or 3 busts less often than the rest of the weak range. If a table does not offer surrender, every Rh in the chart becomes a plain hit.

Soft hands cannot bust, which is why they are played so differently

A soft hand holds an ace still counting as eleven, so the next card cannot break it. Draw a nine to an ace and a six and the hand becomes 16 rather than a bust, because the ace quietly drops to one. That safety net is why soft totals double aggressively against weak dealer cards and why they keep drawing in places a hard hand would stop.

Soft 18 is the row people misplay most. Ace and seven looks like a decent hand and against a dealer 9, 10 or ace it is not: 18 loses to 19, 20 and 21, and the hand can be improved with no bust risk. Against a dealer 3 through 6 it doubles, and against 2, 7 or 8 it stands. Three different answers for one total, which is exactly why a chart exists.

Note also that a Ds cell means double where the rules permit and stand otherwise, never hit. Doubling on soft totals is often unavailable at tables that restrict doubling to hard 9, 10 and 11, in which case standing is the fallback for soft 18 and hitting is the fallback for the lower soft rows. If you are still learning the terminology, the how to play blackjack guide covers what soft and hard mean at the table.

Pairs are checked before anything else because splitting changes the hand you own

Two rows never vary. Aces are always split, because a pair of aces is a hand of 12 that becomes two hands each starting with the best card in the pack. Eights are always split, because 16 is the worst total in blackjack and two hands of eight are both better prospects, even against a dealer 10 where you are simply choosing the least bad option.

Two rows are never split. Tens stand, always, because 20 wins the large majority of the hands it plays and breaking it up to chase two mediocre hands costs money every time. Fives are treated as a hard 10 and doubled, because splitting them produces two hands starting on five, which is the worst possible start.

The 9,9 row is the one with a pattern worth memorising rather than reading: split against everything except 7, 10 and ace. Standing against a dealer 7 looks odd until you notice that a dealer 7 most often makes 17, and your 18 already wins that hand. Note that these pair decisions assume doubling after a split is permitted. Where it is not, the 4,4 row becomes a plain hit and the low pairs split against a narrower range.

Deviating from the chart costs a predictable amount, hand after hand

The temptation to override the chart usually arrives in one of three forms: a hunch about the shoe, a desire to protect a stack after a bad run, or the sense that a decision has been failing lately. All three are expensive, and the cost is measurable.

A player following the chart exactly on liberal rules faces a house edge of roughly 0.5%. A player making sensible-looking decisions by instinct typically gives up somewhere between 1.5% and 2.5%. That is three to five times the price for the same hands at the same table. The difference does not come from exotic situations either. It comes from a small set of common ones.

• Standing on stiff totals against strong upcards. Refusing to hit 15 or 16 against a dealer 9 or 10 feels prudent and is the single most costly habit in casual play.
• Not doubling. Skipping the double on 11 or on soft totals gives up the moments where extra money goes in with the odds in your favour.
• Splitting tens. Two hands of ten look like twice the opportunity. They are a downgrade from a hand of 20.
• Failing to split eights against a strong card. Doubling your exposure on a losing proposition feels wrong and is still the cheaper route.
• Mirroring the dealer. Playing your hand as though you must reach 17 ignores that you act first and pay for busting first.

None of these decisions wins every time it is correct. Splitting eights against a dealer 10 loses more often than it wins, and it is still right, because it loses less than the alternative. Judging a decision by the outcome of one hand is the fastest route away from the chart.

Insurance is declined every time, whatever you happen to be holding

When the dealer shows an ace you are offered insurance, a side wager of up to half your stake that the hole card is a ten-value card. It pays 2 to 1. Behind that lies the reason to refuse it: in a multi-deck shoe rather less than a third of the remaining cards are ten-value, so a 2 to 1 payout is short of the true price. The resulting edge against the player is around 7%, which is worse than the main game by an order of magnitude.

The version offered when you hold a natural is called even money. Taking it guarantees a 1 to 1 return instead of the 3 to 2 you would collect if the dealer has no natural. Framed as certainty, it is the same losing bet with different packaging, and basic strategy declines it too.

Insurance is worth understanding rather than simply avoiding, because it is where the counting story and the honest story meet. In a physical game with a tracked shoe there are moments where the remaining pack is rich enough in tens to justify the bet. That situation cannot be identified at an online live table, which is covered on the blackjack guide along with why shoe tracking has no practical application there. Under basic strategy the answer is unconditional: no.

Rule variations move a handful of cells, not the shape of the chart

The tables above assume six or eight decks with the dealer standing on soft 17. Live tables vary, and the adjustments are small.

• Dealer draws on soft 17. Double 11 against an ace instead of hitting, double soft 19 against a 6, and surrender 17 against an ace where it is offered. The chart is otherwise unchanged, and the rule costs about 0.2%.
• No doubling after a split. Split the low pairs against a narrower range of dealer cards, and stop splitting 4,4 entirely.
• Doubling restricted to 9, 10 and 11. Every D on a soft total becomes a hit, and every Ds becomes a stand.
• No surrender. Every Rh becomes a plain hit.
• Fewer decks. Single and double deck games have their own charts with a number of differences, but they are rare at live dealer tables and the multi-deck version above is the practical one.

What no variation changes is the payout on a natural. A table paying 6 to 5 rather than 3 to 2 adds roughly 1.4% to the house edge, which swamps every cell adjustment on this page and is not recoverable through better play. Check that line first, then the rest. The exact conditions attached to doubling, splitting and surrender are set out on the blackjack rules page.

What the chart does and what it very deliberately does not do

Correct play brings the house edge down to roughly half a percent on liberal rules. That is a genuinely low number for a casino game, and it means a modest session at a modest stake costs far less than the same session at almost any other table. That is the honest case for learning it, and it is a good one.

Here is the other half. Roughly 0.5% is above zero. Over a long enough sequence of hands the expected result of perfect basic strategy is still a loss, and there is no arrangement of correct decisions that changes the sign. Basic strategy minimises the edge. It does not remove it, and nothing on this page should be read as suggesting the game can be turned around.

Anyone presenting the chart as a route to profit has either misunderstood the arithmetic or is relying on you to. Progressive betting layered on top changes nothing either: staking more after a loss alters the shape of your results without touching their expected value, in exactly the way it fails to work on roulette betting systems.

Play blackjack because the decisions are interesting and the price is low, with a budget you have already decided to spend. Set that number before you sit rather than during a session, which the responsible gaming page covers in practical terms.

Learn it in the order the value arrives

Nobody memorises the full grid in one sitting, and nobody needs to. Around a dozen rules capture most of the gain, and they are the ones that come up constantly: always split aces and eights, never split tens or fives, double 11 against everything except an ace, stand on hard 12 to 16 against a dealer 2 to 6 and hit them against 7 upward, stand on hard 17 or more, and decline insurance every time.

Learn those first and the remaining cells add a fraction of a percent between them. Keep the chart open beside the table while you play, which is normal practice and costs nothing. The full grid becomes automatic faster than most people expect, mainly because the same dozen hands keep arriving. If a decision is going to be rushed by a live table clock, it is better to have decided it in advance than to reason it out in eleven seconds.