The Blackjack House Edge, Rule by Rule

What "house edge" means

The house edge is the casino's expected profit, expressed as a percentage of each bet wagered. A 0.5% edge means that for every $100 you put through the game, the casino expects to keep 50 cents on average — over the long run, across millions of hands. It is not what you lose per hand, and it is not what you lose per session. It is the slope of the line you slide down if you play forever.

Mathematically, the edge is the negative of the expected value (EV) of a one-unit bet. If a hand has an EV of −0.005 units, the house edge for that hand is 0.5%. The total house edge for a blackjack game is the bet-weighted average EV across every possible starting hand and dealer upcard, assuming the player plays each one according to basic strategy.

Two practical consequences follow from that definition. First, the edge is computed assuming optimal play — if you misplay, your personal edge is worse than the table's published number. Second, the edge says nothing about variance. You can win twenty hands in a row at a 2% edge, or lose twenty in a row at a 0.3% edge. Edge is the long-run trend; variance is the ride.

The starting point: an untrained player

A player with no strategy — someone who mimics the dealer ("hit until 17"), never doubles, never splits — gives the house roughly 5.5% on a standard game. A player operating on instinct, splitting pairs occasionally and standing on stiff hands sometimes, lands somewhere around 2.0% to 2.5%. That 2-3% range is the realistic edge against most recreational players, and it is what the casino is quietly counting on.

Perfect basic strategy compresses that down to the half-percent neighborhood. For the precise numbers, the canonical source is Wizard of Odds, whose calculator allows the entire rule set to be specified explicitly. The figures in this article are aligned with those values, with the caveat that exact edge depends on a handful of secondary rules (resplit limits, double-after-split, surrender variant) that move things by hundredths of a percent.

Standard rule sets and their edge

The phrase "0.5% house edge" gets repeated everywhere, but it actually only applies to one specific game: six decks, dealer stands on soft 17, double on any two cards, double after split allowed, late surrender, 3:2 blackjack payout. Change any of those rules and the number moves.

That bottom row is the trap. A 6:5 game is mathematically four times worse than a standard 3:2 game and is now standard on most Las Vegas Strip low-limit tables. We will return to it.

How each rule modifier moves the edge

The rules at a blackjack table do not contribute equally. A short list of common modifiers and their approximate impact:

Read this table additively. A six-deck H17 game with no surrender and no DAS adds roughly 0.22 + 0.14 + 0.07 = 0.43% on top of the standard 0.5%, putting you near 0.93% — almost double the edge for what looks like a minor difference in posted rules. Always read the felt before sitting down.

The 6:5 trap

A standard blackjack pays 3:2: a $10 bet wins $15 on a natural 21. A 6:5 game pays only $12 on the same bet. Since a natural occurs about once every 21 hands, that single change costs the player roughly 1.39% of edge — more than every other rule modifier in the casino combined.

The 6:5 game exists because casinos discovered that most players read "single deck" as good and do not check the payout. The single-deck advantage is real (about 0.35%); the 6:5 penalty (1.39%) dwarfs it by a factor of four. Look at the corner of the felt — if it says "Blackjack pays 6 to 5", the table is a tax.

Card counting and the edge floor

Basic strategy is the floor of how low the house edge can go without using outside information. Card counting goes further — by tracking the ratio of high cards to low cards remaining in the shoe, an accurate counter can identify moments when the player has an edge, and bet more in those moments.

On a typical six-deck game with good rules and reasonable penetration, a well-trained counter using a simple system like Hi-Lo can achieve a long-run edge of roughly 0.5% to 1.0% in their favor. This is the only legal way to be a long-term winner at blackjack, and casinos respond by backing off, shuffling early, or banning known counters.

Card counting is outside the scope of basic strategy entirely. Every other section of this article assumes you are not counting — that you are flat-betting the same amount every hand, which is what the published house edge numbers describe.

Edge versus variance

The house edge tells you what happens on average. It does not tell you what happens to you. Blackjack's standard deviation per hand is roughly 1.15 units, meaning a typical 100-hand session has a standard deviation around 11.5 units. On a 0.5%-edge game flat-betting one unit per hand, your expected loss over 100 hands is 0.5 units — half a unit. The standard deviation is 23 times larger than the expectation.

In practical terms: you can play 1,000 hands at perfect basic strategy on a good game and finish $400 ahead, or $400 behind, without the house edge being either confirmed or refuted. Edge becomes visible only over thousands or tens of thousands of hands. Most recreational players will never play enough blackjack to see the long run.

This is why basic strategy can feel like it does not "work" in any given session. It does work — but at the scale of seasons, not sessions.

What edge means for your bankroll

The practical formula every player should know:

Expected loss per session = house edge × hands played × average bet.

At a typical casino pace of 80 hands per hour, a $25 bettor playing for four hours on a 0.5% edge game expects to lose 0.005 × 80 × 4 × $25 = $40. On a 6:5 game (2% edge) the same session expects to lose $160. The math is brutally consistent. Pick the right table and your entertainment costs four times less for the same time at the same stakes.