ABC Tools

Blackjack Basic Strategy and House Edge, Explained

After reading this you will know where basic strategy comes from, how to read the three charts, and how each rule on the table shifts the house edge by a number you can check.

What basic strategy actually is

Basic strategy is the single decision for every hand that maximizes your expected value, assuming you know only your two cards and the dealer's upcard. It is not a betting system and it does not track cards. It answers one question: given this exact situation, which action loses the least (or wins the most) money on average?

Here is the hook. You hold a hard 16 against a dealer 10. It feels terrible either way. Standing wins only when the dealer busts, which happens about 23% of the time from a 10. Hitting busts you most of the time, but the times you survive you often win. Run both through the exact model and the numbers come out close: standing returns about -0.540 per unit, hitting about -0.535. Hitting is the better action by a whisker of 0.005. You lose either way, but the chart tells you which loss is smaller.

The Explorer computes this decision for every hand and every upcard, for the exact rules you select. The result is three heatmaps: hard totals, soft totals, and pairs.

When to use it, and when not

Use the Explorer to build the correct chart for the table in front of you. Casinos vary the rules, and the correct play changes with them. A 16 against a 10 might say hit at one table and surrender at another. A generic chart printed on a card assumes one rule set and can be wrong for yours.

Do not treat the output as a way to beat the house. Basic strategy minimizes the edge; it does not erase it. Even at the best common rules the house keeps roughly 0.4% of every dollar you wager over the long run. Playing perfectly means you lose slowly, not that you win.

Basic strategy is fixed-composition advice. It ignores what has already come out of the shoe. If you want the running edge of your rules without the charts, use the Blackjack House Edge tool. To see the honest cost of a session in dollars, try the Expected Loss Calculator.

The math: exact recursive enumeration

The core of the model is the dealer's outcome distribution. From a given upcard, the dealer draws by a fixed rule (stand on 17, or hit soft 17). Every possible sequence of dealer cards is enumerated, and its probability is the product of drawing each card from the remaining shoe.

P(\text{dealer final} = f) = \sum_{\text{paths} \to f} \; \prod_{i} \frac{n_{c_i}}{N_i}

Here f is a dealer end total (17, 18, 19, 20, 21, or bust), each path is a specific run of dealer cards, n_{c_i} is the count of card c_i left in the shoe, and N_i is the total cards remaining at that step. Your two cards and the upcard are removed first, which is why the tool asks for the deck count.

Once you have the dealer distribution, the EV of standing on total t is direct:

EV_{\text{stand}} = P(\text{dealer bust}) + \sum_{f \le 21} \big[ \mathbf{1}(t \gt f) - \mathbf{1}(t \lt f) \big] \, P(f)

You win the full bet when the dealer busts or ends below your total, lose it when the dealer ends higher, and push when equal. Hitting, doubling, and splitting are handled recursively: each draws a card, then evaluates the best action from the new situation, weighting by the probability of each card. The action with the highest EV is the recommended play, and that number is what you see when you hover a cell.

Worked example: the default 6-deck table

Reproducing the demo

Load the demo: 6 decks, dealer stands on soft 17, double on any two cards, double after split allowed, no surrender, blackjack pays 3:2. These are common Las Vegas Strip rules for a mid-limit shoe game.

  1. The tool removes your cards and the upcard, then enumerates every dealer draw for each upcard from a 6-deck shoe (312 cards).
  2. It computes stand, hit, double, and split EV for all 340 hand-versus-upcard cells.
  3. It picks the best action per cell and sums the results weighted by how often each situation arises, giving the overall house edge.

The result for these rules is a house edge of about 0.46%. That means for every 100 units you wager across a long session, expect to lose about 0.46 units on average. On a 10 per hand game at 80 hands an hour, that is roughly 3.66 per hour of expected loss before variance.

A few decisions from this exact chart worth memorizing: always split aces and eights, never split tens or fives, double hard 11 against everything but an ace, and stand on hard 12 against a dealer 4, 5, or 6 (the dealer's most bust-prone cards).

How each rule moves the edge

Rules do not all matter equally. The chart below shows the approximate change in house edge for each rule change from the 6-deck baseline. Positive means worse for you.

The 6:5 payout dwarfs everything else. A single rule swap adds nearly 1.4% to the edge, while surrender actually helps you.

The lesson is stark. Moving blackjack from 3:2 to 6:5 costs about 1.36%, more than all the other rules combined. A 3:2 game pays 15 on a 10 bet; a 6:5 game pays 12. That missing 3 on every natural, which arrives about once per 21 hands, is where the extra edge lives.

If a table pays 6:5, walk to a different table. No amount of correct play recovers a rule that expensive. A 6:5 single-deck game is worse for you than a 3:2 eight-deck game, even though single deck sounds better.

Try the rules yourself

The baseline 6-deck, S17, DAS, 3:2 game has a house edge near 0.46%. Switching to 6:5 pushes it to about 1.82%. Adding hit-soft-17 to the baseline raises it to about 0.67%. Late surrender lowers the baseline to about 0.39%.

Reading the three charts

Each cell is color-coded by best action. Hover to see the EV of every action ranked, so you can see how close the decision was. This margin matters for your intuition.

Hard totals
Hands with no ace, or an ace forced to count as 1. Read your total down the side, the dealer upcard across the top.
Soft totals
Hands with an ace counting as 11. Soft 18 (ace-seven) is the famous trap: you double against 3 through 6, stand against 2, 7 and 8, and hit against 9, 10 and ace, because 18 is not strong enough there.
Pairs
Whether to split. The split EV assumes one split and no resplits, and split aces get one card each, per the tool's model.

When two actions are within about 0.01 of each other, the choice barely affects your bankroll. A 12 against a 4 is close: stand about -0.208, hit about -0.213. When the gap is large, discipline pays. Splitting eights against a 10 returns roughly -0.485 versus about -0.535 for hitting the hard 16, a real 0.05 saving.

Common mistakes

The most expensive error is not a strategy slip; it is sitting at a 6:5 table. After that, the frequent leaks are these. Standing on soft 18 against a 9 or 10 costs you EV every time. Not doubling 11 against an ace at S17 tables leaves money on the table. Taking insurance is a separate bad bet: it pays 2:1 but a blackjack fills only about 30.8% of the time in a full shoe, so it carries roughly a 7.4% house edge on its own. Skip it always.

Another mistake is using a chart built for the wrong rules. A hit-soft-17 table changes several borderline plays, including doubling hard 11 against an ace and soft 19 against a 6. Always regenerate the chart for your exact table.

Related tools

To compare blackjack against other games on the same house-edge scale, see Roulette Odds and Craps Odds. Roulette's single-zero wheel holds 2.70%, nearly six times the best blackjack edge. For a different shape of risk, the Slot Session Calculator shows how session outcomes spread, and Lottery Odds covers match-tier probabilities.

Frequently asked questions

Does basic strategy let me beat the casino?

No. It minimizes the house edge, typically to around 0.4% at good rules, but the edge stays positive for the house. You lose slowly rather than quickly. Only counting cards can turn the edge, and basic strategy does not count.

Why does the correct play change with the rules?

Every rule shifts the EV of some actions. When the dealer hits soft 17, the dealer makes more strong hands, which pushes several borderline decisions toward more aggressive play. The Explorer recomputes the whole chart so the recommended action is always correct for your selections.

Should I ever take insurance or even money?

No. Insurance wins only when the dealer's hole card is a ten, which happens about 30.8% of the time from a full shoe. At a 2:1 payout you would need it to hit 33.3% to break even, so it carries about a 7.4% house edge. Even money on your own blackjack is the same bet in disguise.

Is a single-deck game always better?

Not if the payout is 6:5. Fewer decks lower the edge by only about 0.62% compared to eight decks, while a 6:5 payout raises it by about 1.36%. Many single-deck games pay 6:5 precisely to erase the deck advantage and then some. Check the payout first.

How accurate are the EV numbers?

They come from exact enumeration of every dealer outcome in a fixed-composition shoe, the same model behind published basic-strategy charts. The split model uses one split with no resplits, so real edges with resplitting allowed are a hair better than shown.

Try it yourself in the tool

Related tools

Card Counting Trainer

Learn Hi-Lo card counting: speed drills, true-count quizzes and a bankroll simulator with your bet spread.

Lottery Odds

Jackpot odds and every match tier of a k-of-N lottery.

Roulette Odds

Win probability, payout and house edge for every bet.

Blackjack House Edge

House edge from the exact table rules you are playing.

Craps Odds

Exact win probability and house edge for the main bets.

Expected Loss Calculator

The honest cost of a session at any casino game.