A basic strategy chart asks you two questions: what is your total, and what is the dealer showing. It never asks which cards made your total. A 12 is a 12, whether it arrived as ten and two, seven and five, or nine and three.
That simplification is deliberate, and it is almost always harmless. Almost. There is a small family of hands where the specific cards matter enough to flip the correct play, and the strategy that accounts for them has a name: composition-dependent strategy. It is a real refinement, it is genuinely optimal, and it is worth so little in a six-deck shoe that most players should read this article for the understanding and then keep playing the chart.
Total-dependent versus composition-dependent
The chart you have seen is a total-dependent strategy. It collapses every hand into a total and a hand type (hard, soft, or pair) and prescribes one play for each combination against each dealer upcard. It is a compact object, roughly a page, and that compactness is the entire reason it is teachable.
A composition-dependent strategy does not collapse. It evaluates the specific cards you hold, because those cards are no longer available in the rest of the deck. Every card dealt changes what remains, and what remains determines your odds of drawing well and the dealer's odds of busting.
Both are still basic strategy in the strict sense: both assume you have no memory of previous rounds and are working only from what is visible right now. That is what separates this idea from counting, which we will come back to.
The canonical example: 12 against a 4 in single deck
Standard basic strategy says stand on hard 12 against a dealer 4. It is a marginal decision to begin with, one of the closest calls on the chart, and marginal decisions are exactly where a small nudge changes the answer.
Now consider a 12 made of a ten and a two, in a single-deck game. You are holding one of the sixteen ten-value cards in the deck. Removing it from the pool has two effects that both point the same direction: your chance of drawing a ten and busting is slightly lower, and the dealer's chance of drawing well is slightly different. In a single deck, that is enough. A ten and a two favors hitting, even though the chart says stand.
The same 12 assembled from a seven and a five leaves all sixteen tens in the deck and does not flip. So in a single-deck game the correct instruction is not "stand on 12 against 4" but "stand on 12 against 4 unless your 12 contains a ten."
This is the standard illustration because it is clean. The mechanism is visible in one sentence: the card in your hand is a card that is not in the deck.
Multi-card hands are the other big family
The second common case is not about which two cards you hold but about how many cards you hold. A multi-card 16 against a dealer 10 plays differently from a two-card 16.
If your 16 is a five, a five and a six, you are holding three small cards. Those are exactly the cards that would have helped you draw to a 17 or better without busting, and they are now gone. That changes the arithmetic of hitting. It is the reason a 16 built from many small cards is a stronger candidate for standing than a fresh ten and six.
Hard 16 against a 10 is already the most argued-about hand in the game, and we treat it at length in our piece on 16 against a dealer 10. The composition-dependent layer is a footnote to that discussion rather than a replacement for it: the base answer against a two-card 16 does not change.
21 TrainerFree on the App Store. The strategy engine implements composition-dependent play, with an explanation for every hand.
Get the app
What it is actually worth
Here is where enthusiasm should meet arithmetic. The gain from playing a full composition-dependent strategy instead of the ordinary total-dependent chart has been computed, and it is small. As Wizard of Odds calculates:
| Number of decks | Approximate gain over total-dependent basic strategy |
|---|---|
| 1 | about 0.036% |
| 2 | roughly 0.011% to 0.014% |
| 6 | about 0.003% |
| 8 | about 0.002% |
Wizard of Odds frames the single-deck figure in a way that makes it concrete: roughly one extra winning bet every 2,800 hands. That is the best case, in the game where the effect is strongest. At six decks the number has fallen by an order of magnitude, and at eight decks it is around two thousandths of a percent, which is closer to zero than to anything you would notice in a lifetime of play.
The reason for the decline is obvious once stated. The effect depends on how much one card changes the composition of the remaining pool. In a 52-card deck, holding a ten removes one of sixteen tens. In an eight-deck shoe of 416 cards, it removes one of 128. The signal is diluted eight times over.
How this compares with the mistakes you are actually making
Put the numbers next to the other levers available to you and the priority becomes clear.
- A 6:5 blackjack payout instead of 3:2 costs roughly 1.4 percentage points. That single rule is several hundred times more expensive than the entire composition-dependent refinement is valuable.
- A dealer who hits soft 17 costs about 0.22%, and European no hole card about 0.11%.
- Double after split gains you about 0.14%.
- Ordinary basic strategy errors on common hands run in the tenths of a percent each, and most players make several.
Against that backdrop, chasing 0.003% at a six-deck table while misplaying soft 18 is not a strategy, it is a hobby. Our house edge calculator guide shows how the rule effects stack, and the stack is where your attention belongs first.
When it is worth learning
There is a defensible case for composition-dependent play, and it is narrow:
- Single deck, and to a lesser degree double deck. This is where the effect lives. If you have found a good single-deck game with a 3:2 payout, the exceptions are worth knowing, and the comparison in single deck versus multi-deck explains why such games are rare and worth recognizing.
- You already have the base chart automatic. Not mostly right. Automatic, at the rule set you play, including the awkward soft doubles and the pair-splitting boundaries.
- You are heading toward counting anyway. The habit of thinking about the remaining deck rather than just your total is the mental move that index deviations are built on. Learning it here, where the stakes are tiny, is decent preparation.
If none of those describe you, the honest recommendation is to enjoy the idea and ignore the practice.
It is not counting, and it does not remove risk
Two clarifications worth stating plainly. First, this is not card counting. Composition-dependent strategy uses only the cards on the table in the current round, requires no memory of previous hands, and is fully correct off the top of a freshly shuffled deck. Counting tracks cards across many rounds and is a far larger undertaking with a far larger payoff.
Second, none of this creates an edge. Basic strategy in either form reduces the house edge, it does not reverse it, and a refinement worth thousandths of a percent changes nothing about the variance that dominates any real session. Play a perfect composition-dependent game and you will still have losing nights, routinely.
What the idea does offer is a better mental model. Once you understand that the chart is a compression of a much larger calculation, you stop treating it as a set of arbitrary rules and start seeing it as what it is: the best one-page approximation of a game that is really about what is left in the deck.
Get the base chart automatic first270+ hand scenarios, charts that redraw for your rules, and a Weakness Drill that finds your worst hands.
Download
Frequently asked questions
What is composition-dependent strategy in blackjack?
It is basic strategy that reads the actual cards in your hand rather than only their total. A hand of ten and two is a 12, and so is seven and five, but they are not identical: the first has already removed a ten-value card from the deck. Composition-dependent strategy accounts for that removal, and occasionally the best play differs from what the standard total-based chart says.
How much is composition-dependent strategy worth?
Very little, and it shrinks fast with deck count. As Wizard of Odds calculates, the gain over total-dependent basic strategy is about 0.036% in single deck, roughly 0.011% to 0.014% at two decks, about 0.003% at six decks and about 0.002% at eight decks. The single-deck figure is framed there as roughly one extra winning bet every 2,800 hands.
Should I hit 12 against a 4 if my 12 is a ten and a two?
In a single-deck game, yes: that is the canonical composition-dependent exception. You normally stand on 12 against a dealer 4, but a 12 built from a ten and a two favors hitting, because one of the ten-value cards that would bust you is already in your hand. In six or eight decks the effect is diluted to the point of irrelevance, so stand.
Is composition-dependent strategy the same as card counting?
No. Composition-dependent strategy uses only the cards visible in the current hand, so it needs no memory of previous rounds and works from the first hand off a fresh shoe. Card counting tracks cards across many rounds to estimate what remains in the shoe. Counting is far more powerful, and far more work. The two ideas share a premise but not a method.
Should beginners learn composition-dependent exceptions?
No. The gain from these exceptions is a rounding error next to the cost of ordinary basic strategy mistakes, which run in the tenths of a percent per misplayed decision. Get the standard chart automatic first, at the rules you actually play. Composition-dependent refinements are worth adding only for single-deck and double-deck games, and only after that.