Blackjack Basic Strategy 2026 – Hard, Soft and Pair Charts, Rule Effects and the EV Behind Every Decision

Why basic strategy crushes intuition – the 2 % vs 0.5 % math

Edward O. Thorp put the first computer-derived basic strategy on paper in 1962 in <em>Beat the Dealer</em>. He ran an IBM 704 against every possible hand combination and built the table that every casino book has copied since. The point of the table is simple: for every total you can hold against every dealer upcard, only one of {hit, stand, double, split, surrender} maximises your expected return. Any other decision gives EV back to the house.

The total cost of bad decisions adds up faster than most players realise. The casino industry has measured it on the floor for decades: a player making "casino-floor average" decisions – splitting tens, refusing to double 11 against an ace, taking insurance on every blackjack – faces a house edge of 1.8–2.4 %. The same player switched to a printed basic strategy card faces 0.42–0.55 %. That is a 4x reduction in expected loss per hour at the table.

House edge with basic strategy across common rule sets. Source: Schlesinger, <em>Blackjack Attack</em> (3rd ed.), Wattenberger CVCX runs.

Basic strategy is not card counting. It does not assume any knowledge of the upcoming cards – every decision is the EV-optimal play given only your own hand, the dealer's upcard and the rules of the table. That makes it the cheapest, fastest edge you can claim back from any casino, on any blackjack table, in any country. There is no banned-from-the-table risk and no math homework at the table itself once the chart is memorised.

For a deeper look at how the house edge is built into every casino game, see our companion guide on RTP and house edge.

The rules that move EV – S17, H17, DAS, decks, surrender, resplit

Before you sit down, check the table sign. The same game card with different rules can be 0.50 % cheaper or more expensive to play. These are the six rule swings that actually matter, in descending order of impact.

Rule EV adjustments measured against the 6-deck S17 DAS baseline. Source: Schlesinger (2018), Wong (1994), Wizard of Odds rule-effect tables.

Hard, soft and pair hands – why they need three different charts

A blackjack hand sits in one of three states, and each state changes the math of every decision. Memorising which state you are in is the prerequisite to using the chart correctly.

• Hard hand. No ace, or an ace forced to count as 1 because counting it as 11 would bust the hand. The bust risk is fixed by the count alone. Example: 10 + 6 = hard 16. The next card immediately matters.
• Soft hand. Contains an ace counted as 11 without busting. You cannot bust on the next card, which is why soft hands let you double far more aggressively. Example: A + 6 = soft 17.
• Pair. Two cards of equal rank dealt on the first action. You can choose to play it as a single 12, 14, 16, 18 or 20 – or to split it into two hands. The optimal action is almost never the obvious one. Tens, for example, are never split.

The three charts below cover the 6-deck S17 DAS late-surrender game – the most common rule set in regulated European online and live blackjack rooms. If your table is different (single deck, H17, no surrender), the chart shifts on a handful of cells. We list those exceptions after the tables.

Hard totals – the chart that ends the splitting-tens debate

Read this table by finding your hand total on the left, then crossing to the column that matches the dealer's visible card. The cell tells you the EV-optimal action. Anything below hard 8 is always a hit. Anything 17 or higher is always a stand. The interesting decisions live between 9 and 16.

Hard totals, 4–8 deck, S17, DAS, late surrender. H17 swap: hit 11 vs A becomes double, surrender 15 vs A and 17 vs A.

Two cells deserve special attention. Hard 12 against a dealer 2 or 3 is a hit, not a stand – the dealer's bust rate on a 2 or 3 is only around 35 %, and you have eight cards in the deck that improve your 12 without busting. Hard 16 against a 10 is the worst hand in blackjack; surrender if the table allows it, hit if it does not.

Soft totals – why A,7 is not the safe hand it feels like

Soft hands give you the safety of never busting on the next card, so the chart is heavy on doubles. The mistake we see at every table is players standing on A,7 (soft 18) against a dealer 9, 10 or ace. Soft 18 is a losing hand against any of those upcards – the dealer will out-draw you more often than you outdraw them, so the correct play is hit, not stand.

Soft totals, 4–8 deck, S17, DAS. H17 swap: double A,7 also against dealer 2 instead of stand; double A,8 against dealer 6.

Pair splitting – the chart that vetoes splitting tens forever

A held 20 (two tens) wins 83 % of the time. Two new hands starting with a 10 each win, against an average dealer upcard, around 55 % each. Splitting trades a strong winning hand for two mediocre ones. The chart says it bluntly: never split tens. The same logic, in the opposite direction, gives you the most aggressive split in the game – always split A,A and 8,8, no matter what the dealer shows.

Pair splits, 4–8 deck, S17, DAS, late surrender. Without DAS: do not split 2,2 / 3,3 / 4,4 / 6,6 against dealer 2 or 3 – hit instead.

Insurance and even money – never take it, here is the EV math

When the dealer shows an ace, the table offers you insurance: a half-stake side bet that the dealer's hole card is a 10. The bet pays 2:1 if correct. It looks tempting, but the math is brutal. In a fresh 6-deck shoe, exactly 16 of the remaining 51 unseen cards are tens. The break-even point would require 17 of 51. Take insurance at 16:51 and you give 7 % EV to the house on every insurance unit.

Insurance EV in a 6-deck shoe with no other visible cards. Negative for the player by ~6.86 % of the insurance stake.

Even money is the same bet wearing a different costume. When you hold a natural blackjack against a dealer ace, the dealer offers you a 1:1 payout instead of waiting to see if the dealer also has blackjack (in which case you push). Mathematically this is identical to taking maximum insurance on your blackjack. The break-even is identical, the EV is identical, and the right answer is the same: decline. Wait for the hole card. You will lose your blackjack to a dealer natural 30.8 % of the time and win 3:2 the other 69.2 %, which beats locked-in 1:1 by 0.13 units per blackjack.

Side bets – Perfect Pairs, 21+3, Lucky Lucky and friends

Side bets are the most expensive square inch on the blackjack table. They exist because they print money for the operator and because the dopamine of a 100:1 hit makes them feel survivable. Here is what they actually cost.

Source: Wizard of Odds side-bet calculator (Pristine Pay Table) and Galaxy Gaming math sheets. Hit frequencies include all winning combinations, not net-positive only.

A 6 % house edge on a side bet is twelve times the cost of the main game. Even the "cheap" Lucky Lucky at 2.66 % is six times more expensive. If you want lottery-ticket variance, buy a lottery ticket – it pays out a more honest 50 % of its face value. We never recommend side bets in our blackjack reviews.

Variants – European, Spanish 21, Double Exposure, Free Bet

Online lobbies often offer five or six "blackjack" variants. They share the name, not the math. Here is what actually changes – and how each variant shifts the basic-strategy chart.

Live blackjack vs RNG blackjack – same math, different speed

Online lobbies serve two flavours. RNG blackjack uses a software shuffler that returns each card from an independent draw – effectively an infinite shoe with no card removal between hands. Live blackjack streams a real dealer cutting a real 6 or 8-deck shoe at the studio, with cards removed from the shoe as hands play out.

• Basic strategy is identical. The chart for a 6-deck S17 DAS table works for both RNG and live. The infinite-shoe assumption of RNG matches the 6-deck math closely enough that the few cells that differ are not worth a second table.
• Card counting is dead in RNG. Without card removal between hands, the count never moves. RNG blackjack is a pure basic-strategy game. Live blackjack still has a count – but the deck penetration is typically only 50–60 % before reshuffle, so the practical edge is tiny.
• Tempo decides loss rate. RNG blackjack deals 200–300 hands per hour for a single player. Live blackjack deals 50–80 hands per hour with multiple players sharing the table. Same EV per hand, four times the EV per hour on RNG. Tempo, not luck, drains bankrolls.

Loss-per-hour math: a 100 € bettor on RNG blackjack at 0.45 % house edge and 250 hands/hour loses 112.50 € per hour on average. The same bettor at a live table at 70 hands/hour loses 31.50 € per hour. If you want to play long, play live. For a deeper look at sizing your spin against the math, see our bankroll management guide.

Five common mistakes that quietly cost real money

FAQ

Is basic strategy legal in every casino?

Yes. Basic strategy uses only the information that the dealer hands you – your cards and their upcard. Casinos may not love it, but it cannot be banned. Many live floors will let you use a printed chart at the table as long as it does not slow down play.

Will basic strategy make me a winning player long term?

No. The math still has a small house edge built in. Basic strategy minimises your expected loss to about 0.45 % of stake. You are a much cheaper loser, not a winner. Beating blackjack long term requires card counting plus appropriate bankroll, plus the freedom to be backed off – outside the scope of basic strategy.

How long does it take to memorise the full chart?

Most players reach perfect recall in 6–10 hours of focused drill, spread across 2–3 weeks. Use a free trainer (Wizard of Odds, Blackjack Apprenticeship app) and run 200 randomly generated hands per session. The hard chart is the longest to drill, the pair chart the shortest.

Does basic strategy change between online and live?

No. The chart is the same for RNG and live blackjack at the same rule set. What changes is your hands-per-hour rate (much higher online) and therefore your loss-per-hour rate. Pick rules first, table second – do not pay 1 % more house edge for a fancier UI.

Should I clear a casino bonus on blackjack?

Usually no. Most operators give blackjack a wagering weight of just 5–10 %, meaning every 1 € wagered counts as 0.05–0.10 € toward the requirement. The low house edge that makes blackjack attractive also makes it ineligible for full-weight wagering. Slots remain the math-of-choice for bonus clearing – see our bonus value calculator.

EJ

★ By Erik Jansson

Spelrecensent · Slots, bordsspel & betallösningar

Erik är casinorecensent med fokus på spelmekanik, slotvarians och betallösningar. Han har spelat och analyserat över 4 200 casinospel.