Six decks, the dealer stands on all 17s, double after split, no surrender — the rules of our table. Print it, learn it, then test yourself in the trainer.
Basic strategy is the move with the best long-run result for every hand you can hold against every card the dealer can show. Played perfectly at our table, it leaves the house an edge of 0.40% — on average you lose 40 debens for every 10,000 you bet. Find your hand on the left, the dealer’s card on top, and play the letter.
Look at the pairs table first, then soft totals, then hard totals. A pair you don’t split is played as a total: 5,5 is hard 10, and 10,10 is 20.
If the full chart looks like a lot, these seven rules get most hands right and fit in your head:
What is left — the small pairs and the soft doubles — is where the chart earns its keep, and where the trainer helps most.
For each of the 55 possible two-card hands and each of the 10 dealer cards, a program took a six-deck shoe, removed those three cards and worked out the average result of standing, hitting, doubling and splitting. It followed every card the player and the dealer could draw, with the dealer standing on 17 and holding no blackjack, and after the first move it kept choosing the best play. Hands with the same total were then combined by how often they occur: 10-6 comes up more often than 9-7, so it weighs more in the “16” row.
The result is the chart above. It matches the standard published charts for six decks, dealer stands on soft 17, double after split and no surrender, cell for cell. Then we tested it from the other side. A simulator played 50 million rounds with this chart through a shuffled six-deck shoe, reshuffled at the same card as in the game. The house kept 0.43% ± 0.03% of the initial bets; the exact calculation says 0.40%. With a fresh shoe dealt every round, the simulation gives 0.41% ± 0.03%, so the small gap comes from chance and the cut card, not from the chart.
Only ten-value cards bust a 12 — four ranks out of thirteen. A dealer showing 2 busts only 35% of the time, with a 3 it is 37%, so standing and hoping isn’t enough. Per 100 debens bet, hitting 12 against a 2 loses 25.3 on average and standing loses 29.3. Against a 3 the gap narrows: 23.2 against 25.2. Against a 4 the dealer busts 40% of the time, and standing wins by a hair: 20.9 against 21.2.
Both moves are bad. Standing on 16 wins only when the dealer busts, which with a ten up happens 23% of the time; hitting busts you on any card from 6 up. Per 100 debens, hitting loses 53.5 and standing 54.0 — the closest call on the chart. With three or more cards the balance often tips the other way: with 7-5-4 or 4-4-4-4 against a ten, standing is better, because the small cards that could have saved the 16 are already in your hand.
Soft 18 is the only row that uses all three moves. Against 3–6 it doubles: against a 6, doubling earns +38.3 per 100 and standing +28.0. Against 2, 7 and 8 it stands, because 18 already beats or ties what those cards usually make. Against 9, 10 and ace it hits: there 18 is the underdog, and one card can’t bust a soft hand. Against a ten, hitting loses 14.3 per 100 and standing 18.0.
A dealer showing 7 most often finishes on 17 — 37% of the time — so your 18 is already ahead. Standing earns +40.0 per 100, splitting +36.9. Against an 8 the picture flips: split +23.0, stand +9.9. Against a 9 splitting loses less, −8.1 against −18.5. Against a ten or an ace, keep the 18.
Not all mistakes are equal. What matters is how much a wrong move costs and how often the hand comes up. Here is the price of common habits, as a share of everything you bet, on top of the 0.40% the house already has:
| Habit | How often the hand comes up | Extra cost |
|---|---|---|
| Standing on 12–16 against 7 to ace | 21% of hands | 2.60% |
| Hitting 13–16 against 2–6 | 11% of hands | 1.56% |
| Never doubling 11 | 4.3% of hands | 0.74% |
| Never doubling 10 | 2.2% of hands | 0.40% |
| Splitting tens against 4–6 | 2.2% of hands | 0.36% |
| Taking insurance every time | 7.7% of hands | 0.28% |
| Standing on soft 17 | 1.1% of hands | 0.25% |
| Splitting fives | 0.35% of hands | 0.19% |
| Hitting a pair of aces instead of splitting | 0.54% of hands | 0.18% |
| Standing on 16 against a ten | 1.7% of hands | 0.01% |
The last row is there on purpose. People argue about 16 against a ten more than about anything else in blackjack, and it is worth a hundredth of a percent. The first two rows — standing on every stiff hand, or hitting every stiff hand — cost hundreds of times more.
To see what the chart is worth in practice, we let the same simulator play 50 million rounds in each of several styles. Same shoe, same rules, one player at the table:
| How the player plays | House edge |
|---|---|
| Basic strategy (this chart) | 0.43% |
| The chart, plus insurance and even money | 0.74% |
| The chart, but never splitting | 0.99% |
| The chart, but never doubling | 1.95% |
| The chart without doubles or splits | 2.34% |
| “Never bust”: stand on hard 12 or more, doubles and splits by the chart | 3.94% |
| “Mimic the dealer”: hit to 16, stand on 17, never double or split | 5.66% |
| “Never bust”, never double or split | 6.06% |
At 100 debens a hand, 1,000 hands cost a basic-strategy player about 430 debens on average. Mimicking the dealer costs about 5,700. The cards are the same; only the decisions differ.
This chart is for our rules, and we checked how it would change under others by running the same calculation again. If the dealer hits soft 17, three cells change: double 11 against an ace, double soft 18 against a 2 and soft 19 against a 6. With late surrender, you would surrender 16 against a 9, 10 or ace and 15 against a ten. Without doubling after a split, 2s and 3s are no longer split against a 2 or 3, 6s against a 2, and 4s not at all.
And none of the fine points matter as much as the payout. If blackjack pays 6 to 5, the house edge rises by 1.36%, more than three times the whole edge of our table. Check the payout first, the chart second.
Mostly, but not exactly. The number of decks, whether the dealer hits soft 17, doubling after splits and surrender each change a few cells. This chart is for six decks, dealer stands on soft 17, double after split, no surrender.
Yes. Keep this page open next to the game or print it out. The debens are free, and nothing in the game forbids it.
No. It gives the best average result, and the average is still a small loss: 0.40% of what you bet. Over one evening anything can happen, in either direction.
A dealer showing 2 busts only about 35% of the time, while only ten-value cards can bust your 12. Hitting loses 25.3 per 100 debens bet, standing 29.3.
Hard 12–16 against the dealer’s card. These hands come up about once in three deals, and the two common errors there cost 1.6–2.6% of everything you bet.
Deal a hand, pick a move, see whether the chart agrees and what each move is worth.
Read more →How to tell them apart, why soft hands are played boldly, and the soft 17 and 18 traps.
Read more →Hard 9, 10 and 11, the soft doubles, the doubles that cost money, and what doubling is worth.
Read more →Every pair against every dealer card, our split rules and what each split is worth.
Read more →Our 0.40%, how it was calculated, what every rule changes and how simple styles compare.
Read more →
Sit down at the Scribes’ Table in the Hall of Thoth with the chart beside you. A hand costs from 100 free debens.
Play free
