Basic strategy is the set of decisions — hit, stand, double, or split — that maximises your expected value on every possible starting hand against every possible dealer upcard. It's not a system or a guess: it's the exhaustively-computed optimal play for the specific ruleset you're playing. Memorising the chart for one variant takes a few hours; executing it correctly closes the house edge from ~5% to ~0.5%.
This guide covers the standard live-casino variant — 4 to 8 decks, dealer stands on soft 17 (S17), doubling after split allowed (DAS), no surrender. That's the rule combination you'll find at every Evolution, Stake, and Roobet live blackjack table. Single-deck and H17 variants need slightly different charts; we'll cover the deltas at the end.
Hard totals
A hard total is any hand without an active ace counted as 11. Hard hands can bust on the next card, so the central question is always: does drawing have a higher expected value than standing here?
Hard 17 and above
Always stand. The probability of improving by drawing is smaller than the probability of busting plus the times you bust the dealer by surviving. There is no upcard against which it's correct to hit a hard 17 or higher.
Hard 13-16: the "ugly" zone
Stand against dealer upcards 2-6 (dealer is likely to bust), hit against 7-A. The intuition: when the dealer is showing weak, they're likely to draw themselves into a bust at 16/17 — you survive by standing. When the dealer is showing strong, sitting on 15 against a dealer 10 loses to every non-busted dealer outcome. You have to take the risk.
16 against a 10 is the most painful spot in blackjack — the odds are bad either way (you lose either ~77% by standing or ~77% by hitting). Hit is marginally less awful. If the table allowed surrender, that's the spot to take it.
Hard 12
Stand against 4-6, hit against everything else (2, 3, and 7-A). The reason 2 and 3 flip back to hit despite being weak upcards: the dealer is much less likely to bust starting from a 2 or 3 (~35%) than from a 4-6 (~40%+). Your 12 doesn't beat enough non-bust dealer outcomes to coast.
Hard 11
Double against everything 2-10. Against a dealer ace, just hit. (In H17 variants you'd double against the ace too.) Hard 11 is the textbook doubling spot: you can't bust on the next card, and a 10-value draws to 21.
Hard 10
Double against 2-9. Hit against 10 or ace. The cutoff at 10 is precise: a dealer showing 10 has the strongest non-blackjack starting hand; doubling into them costs EV.
Hard 9
Double against 3-6. Hit against 2, 7-A. 9 against the dealer's 2 is borderline (some old charts say double), but modern computer analysis prefers hit by a thin margin.
Hard 5-8
Always hit. These totals are too low to stand or double — you improve more than you risk by drawing.
Soft totals
A soft total contains an ace counted as 11. The key property: you can't bust on the next card, because the ace gracefully drops to 1 if needed. That changes every decision. Hitting soft hands is much cheaper than hitting hard ones.
Soft 20 (A,9) and soft 19 (A,8)
Almost always stand. The one exception: double soft 19 against a dealer 6 in S17 — a tiny edge from doubling into the dealer's likely bust. Don't lose sleep if you forget.
Soft 18 (A,7)
This is the trickiest hand in the chart, and the one beginners most often get wrong. Against dealer 2-6, double if you can, otherwise stand. Against 7 or 8, stand — your 18 already beats a 17/18 dealer. Against 9, 10, or ace, hit. The dealer is too likely to land 19+; sitting on 18 loses too often.
Soft 17 (A,6)
Never stand on soft 17. You can't bust on the next card, so there's no downside. Double against dealer 3-6 (capitalise on the bust-prone dealer), hit everywhere else.
Soft 13-16
Double against 4-6 (5-6 only for soft 13/14). Hit everywhere else. The doubling spots are where the dealer is most likely to bust and your hand is weak enough that one more card meaningfully helps.
Pairs
Splitting turns one decision into two, with proportional risk. The math gets messy fast, but the chart distills to a small number of rules.
Always split
- Aces. Two starting-aces play far better separately. The catch: most casinos give each ace exactly one card, no resplits, no doubling. Even with the restriction, splitting is correct.
- 8s. 16 is the worst hand in blackjack. Splitting into two starting-8s is significantly better than playing a combined 16.
Never split
- 5s. A pair of 5s is a hard 10 — treat it as such and double against 2-9, hit against 10/A. Splitting 5s into two starting-5 hands is a major EV leak.
- 10s. 20 is a near-winning hand. Breaking two 10s into two starting-10s usually loses EV (one good hand becomes two decent hands).
Conditional pairs
- 9s. Split against 2-9 except7 (where 9-9=18 beats the dealer's likely 17). Stand against 7, 10, A.
- 7s. Split against 2-7. Hit against 8-A (your 14 is too weak to stand, splitting just doubles your exposure).
- 6s. Split against 2-6 (DAS rules). Hit against 7-A.
- 4s.Split only against 5 or 6 (DAS rules). Hit against everything else — two starting-4 hands aren't a strong base.
- 3s and 2s. Split against 2-7, hit everywhere else. The DAS rule matters here — without DAS the chart shrinks to 4-7.
The most common leaks
Most amateur play doesn't deviate randomly — the same five or six mistakes account for the bulk of the lost EV. Eliminate these and you're basically playing strategy:
- Standing on 12 against a dealer 2 or 3 (the dealer isn't bust-prone enough at 2/3 to justify it).
- Standing on soft 17 (you can't bust — always hit).
- Taking insurance against a dealer ace (~7.4% house edge on the bet).
- Refusing to split 8s against a 10 (16 vs 10 loses ~77%; two starting-8s loses far less).
- Splitting 10s (turns a 20 into two so-so hands).
- Doubling hard 11 against a dealer 10 but not against an 8 or 9 (the latter are better doubles).
Practising the chart
Reading a chart isn't the same as playing it. The hard part is the speed-of-recall reflex: live tables move at four or five hands per minute, and you need the right call on each without thinking. The blackjack simulator deals random hands and tells you instantly whether your decision matched the chart — including the long-run EV cost of any divergence. A few hundred hands and the chart is in your fingers.
If you prefer a chart you can stare at, the full interactive strategy chartis on the trainer's strategy page. Hover any cell for a one-sentence explanation of why that's the right call. The same page also includes side-bet odds sorted by house edge, so you can see at a glance which optional bets are merely bad and which are catastrophic.
Rule variations
The chart above is for the 6-deck S17 DAS no-surrender variant — the live-casino default. A few common variations shift things:
- H17 (dealer hits soft 17): the dealer is slightly more likely to bust on soft 17 hits but also slightly more likely to improve to 18+. Net effect: house edge ~0.2% worse for the player. A handful of chart cells flip — double soft 19 against dealer 6, double hard 11 against ace, surrender 15 vs ace.
- Single deck: house edge ~0.5% better for the player. Most cells unchanged; a few pair-splits and doubles shift.
- Surrender available: roughly 0.07% house edge better for the player. Surrender hard 16 vs 9/10/A and hard 15 vs 10. (Almost never offered on online live tables.)
- No DAS: tightens the split rules — fewer pair splits, particularly of low pairs (2s, 3s, 4s, 6s).
What the chart isn't
Basic strategy minimises long-run loss. It does not eliminate the house edge — that 0.5% is still there, and you will have losing sessions. It also doesn't care about bankroll or progressive betting; it's a per-decision chart, not a betting system. Doubling your bet after a loss (Martingale, etc.) is uncorrelated with basic strategy and adds variance without changing the underlying EV.
Card counting goes a step beyond basic strategy: by tracking the ratio of high to low cards remaining in the shoe, a counter adjusts both their bet sizes and a small number of chart cells to push expected value slightly positive. Counting works in principle on live tables; in practice, online live shoes are often reshuffled mid-shoe to defeat it, and the operators actively flag suspected counters.
From here
Now that you have the chart: practise it. The simulator is the fastest way to drill basic strategy into reflex. If you want the math behind the chart — where the 0.5% house-edge figure comes from, how variance plays out over real session lengths — the odds & house edge guide walks through the numbers.
