How the blackjack house edge is calculated

The house edge in blackjack is the average amount a basic-strategy player loses per round, expressed as a fraction of the initial bet. For the site’s rule profile (six decks, dealer stands on soft 17, double any two cards, double after split, no surrender, US peek) it is about 0.4%: the Wizard of Odds gives 0.4062% and the blackjackinfo.com strategy engine 0.44%. The number is small because the rules hand back most of the structural advantage the dealer gets from acting last.

Two definitions: house edge and element of risk

The Wizard of Odds defines house edge as “the ratio of the average loss to the initial bet” and element of risk as “the average loss divided by total money bet”. In most casino games they coincide. In blackjack they do not, because doubling and splitting put extra money on the table after the initial bet. A basic-strategy player wagers roughly 1.1 units per round on average across the hands that are doubled or split, so dividing the same average loss by a bigger denominator gives a smaller percentage. The Wizard quotes one Atlantic City rule set at 0.43% house edge and 0.38% element of risk.

For budgeting, house edge is the figure to use with your initial stake: expected loss ≈ rounds × initial bet × house edge. The bankroll and variance guide applies that formula. Do not multiply by total money wagered as well; the doubles and splits are already priced in.

Where the house’s advantage comes from

The dealer plays a fixed strategy that any player could copy, and the dealer has no option to double, split, surrender or stand on 16. On paper that looks like a disadvantage. The house wins because of order of play: you act first, and if you bust you lose your stake at once, before the dealer has drawn a card. When both sides bust the house still collects.

A rough illustration, not a published figure: the six-deck S17 dealer busts 29.6% of the time after a peek (Wizard of Odds dealer-odds table). If a player mimicked the dealer exactly, hitting to 17 with no other choices, the player would also bust around 28% to 30% of the time, and the double-bust region of roughly 0.29 × 0.29 ≈ 8% of rounds would go to the house. That is the size of the structural hole the player starts in.

How the rules give most of it back

Everything the player gets that the dealer does not is a partial refund. The table below separates the main components. The first is exact arithmetic for six decks; the rest are the Wizard of Odds rule-variation prices (eight-deck baseline), which measure what each option is worth by removing it.

Standing on stiffs against small upcards is the largest single contribution after naturals, but it cannot be stated as one number because it is inseparable from the strategy itself: the whole basic strategy chart is the set of choices that minimise the loss on each starting hand. Add the refunds to the structural hole and the residual is the house edge of a few tenths of one percent.

Per-hand outcome frequencies

Another way to see the edge is through outcome frequencies. For a liberal six-deck S17 game with double after split, late surrender and resplit aces, the Wizard of Odds simulation of about ten billion hands reports:

The player loses more hands than they win by nearly seven percentage points, yet the money lost is only 0.28% of the initial stake. The gap is bridged by payout asymmetry: wins include 3:2 naturals and doubled hands, while a sizeable share of losses are surrendered half-units. The lesson is that win frequency says little about expectation; the betting systems article explains why no bet-sizing scheme can change the 0.28%.

The baseline for the site’s profile

The site’s chart and trainer assume six decks, S17, double any two cards, double after split, no surrender, US peek, split to four hands, one card to split aces. Two independent published figures bracket that profile:

The Wizard figures assume a fresh shoe for each hand; the blackjackinfo estimate is rounded and may use a different assumption on resplits. Quoting “about 0.4%” for the profile is honest; quoting 0.4062% implies a precision that depends on details no table card states. For each hand the site’s strategy chart reflects these rules, and the rule variations article gives the adjustments if your table differs.

Why it is “about 0.4%” and not a universal constant

The house edge is a function of rules, strategy and method:

• Rules. H17 adds about 0.22%, 6:5 naturals add 1.39%, late surrender against a 10 gives back 0.07%, single deck gives back 0.48% relative to eight decks (Wizard of Odds rule-variation table).
• Strategy. Any deviation from basic strategy raises the figure. Composition-dependent play lowers it by 0.0031% in six decks, which is why it is not worth learning for shoe games.
• Method. The Wizard’s house-edge calculator distinguishes optimal (composition-dependent, shuffle every hand), basic strategy with a cut card, and basic strategy with a continuous shuffler. The cut-card effect slightly favours the house because more cards are dealt from ten-poor shoes.

None of these change the sign. Basic strategy minimises the house edge; it does not eliminate it, and no betting pattern alters it. Only card counting, which changes the bet according to the remaining shoe, moves the expectation above zero, as the card counting guide explains.

Related reading

• Probability of being dealt a blackjack
• Standard deviation and risk of ruin
• How rule variations change the house edge
• Bankroll and variance