Played correctly under a standard rule set, blackjack costs about 0.51%, which makes it one of the best games on a casino floor. Played at the wrong table, it costs three times that.
Each figure below comes from running the full expected-value calculation twice: once with the rule and once without. Nothing here is a quoted delta.
| Rule | Effect on house edge | Resulting edge |
|---|---|---|
| Blackjack pays even money | +2.25 pp | 2.767% |
| Blackjack pays 6:5 | +1.35 pp | 1.865% |
| Dealer hits soft 17 | +0.22 pp | 0.731% |
| No double after split | +0.14 pp | 0.653% |
| Double on 9-11 only | +0.09 pp | 0.601% |
| Late surrender allowed | -0.09 pp | 0.426% |
The order matters more than the numbers
Read the blackjack payout first. A 6:5 table costs 1.35 percentage points, more than every other rule on this page combined. Nothing else is close, and no combination of good rules rescues it.
Then the dealer rule. Hitting soft 17 costs 0.22. Prefer S17, but never walk past a 3:2 payout to get it.
Then the small print. Double after split, doubling freedom and surrender are worth a tenth of a point or so each. Real, but decided after the first two.
Then play it correctly
All of the above assumes correct basic strategy. Without it the rules barely matter: playing by feel can cost more than a rule on this page, though no engine can price it.
Our basic strategy chart gives the computed action for every hand, and the dealer upcard odds behind it explain why the chart says what it says.
Every page in this section
- Basic strategy chart: the computed action for every hand against every dealer upcard.
- Dealer upcard odds: bust probability for each upcard, and why it drives the chart.
- Splitting pairs, which pairs to split, and the two rules that are always right.
- Soft hands: why a hand with an ace counting eleven plays completely differently.
- Doubling down: when to press a bet, and what a restricted table costs.
- Double after split: the small rule that decides whether to split fours.
- Surrender: the four hands worth folding for half a bet.
- 6:5 blackjack: the single most expensive rule in the game.
- H17 against S17: the dealer rule printed on every felt.
- Insurance; the side bet that is never correct without counting.
Two things that are not decisions
Insurance and even money both lose money in every deck size off the top of the shoe. They turn profitable only in a ten-rich shoe that you would have to be counting cards to recognise. Without a count, decline both, every time.
Frequently Asked Questions
- What is the house edge in blackjack?
- About 0.51% with correct basic strategy under standard rules. The rules at your table can take it above 1.5%.
- What is the single worst blackjack rule?
- A 6:5 payout on a natural, at +1.35 percentage points. It costs more than every other common rule variation combined.
- Does blackjack have the lowest house edge in the casino?
- It is among the lowest for a game of skill, but craps free odds carry a 0% edge and full-pay video poker can approach 100% return.
- Do I need to count cards?
- No. Basic strategy alone produces the 0.51% figure. Counting is a separate discipline and is not needed to play the game well.
- Where do these numbers come from?
- An expected-value engine that plays the game out by recursion and measures each rule by re-running the whole calculation with it changed.
Corrected on 17 October 2026: the engine now values a resplit as two hands, not one, so every figure on this page moved by up to 0.01 points: the standard game is 0.5117% (was 0.5153%) and losing double after split costs 0.14 points (was 0.13). The resulting edges in the table are now the engine's own output rather than the base plus a rounded change.
Corrected on 2 November 2026: this page said the model sits a few hundredths of a percent from an exact shoe analysis. It is about a tenth of a point: Wizard of Odds' simulations give about 0.41% for six decks under these rules, against the model's 0.5117%.
Every figure on this page is computed by an expected-value engine that plays the game out: it derives the dealer's final-total distribution by recursion, evaluates standing, hitting, doubling and splitting at each decision, and takes the best. Rule costs are measured by re-running the whole calculation with the rule changed, not by quoting a published delta. The model assumes independent card draws, which prices a six-deck game about a tenth of a point worse than a shoe does; the rule differences it produces match published values. Rules vary by table, so read the felt.
On seventeen. The dealer's S17 or H17 rule, your hard 17 and your soft 17 are three different questions that share a number: standing beats hitting on a hard 17 against all ten upcards, and hitting beats standing on a soft 17 against all ten.