A casino does not create house edge by choosing how much it wants to win tonight. The edge is built into the relationship between possible outcomes, their probabilities, the amounts paid, and the rules used to settle them.
For a fixed set of rules, house edge is a mathematical property of the wager. Game speed, table occupancy, and average bet affect how quickly the casino earns theoretical revenue, but they do not change the percentage edge of an individual decision.
That distinction is the starting point for understanding how a game is designed and selected.
Begin with expected value
For a wager with several possible outcomes:
Player EV = Σ (probability of outcome i × net player payoff for outcome i)
The symbol Σ means add the value for every possible outcome. Probability of outcome i is the chance of that result. Net player payoff is the profit when the player wins or the negative stake when the player loses.
House edge is then:
House edge = −Player EV ÷ initial wager
A negative player EV becomes a positive casino edge. If a $1 wager has EV of −$0.027, the house edge is about 2.7%.
This formula applies cleanly to games of pure chance. For decision games, the calculation must also specify the player's strategy. A blackjack edge quoted for correct basic strategy is not the edge produced by random hitting, standing, doubling, and splitting.
The simplest lever: pay less than true odds
A straight-up wager on single-zero roulette wins on one of 37 pockets. Fair odds would produce 36 units of profit for a one-unit stake because there are 36 losing outcomes for every winning outcome.
The casino pays 35 to 1.
EV = (1/37 × 35) + (36/37 × −1)
EV = −1/37 ≈ −0.027027
House edge ≈ 2.70%
The wheel supplies the probability. The paytable supplies the return. The one-unit gap between fair profit and actual profit creates the edge.
Add 00 while keeping the 35-to-1 payout and the calculation becomes:
EV = (1/38 × 35) + (37/38 × −1) = −2/38
The edge rises to about 5.26%. Nothing about the dealer's speed caused that change. The outcome set changed while the payout remained compressed.
Many games use rule asymmetry instead of one obvious short payout
Blackjack is more complex. Players act before the dealer and can bust before the dealer completes the hand. That positional asymmetry creates casino value. Other rules then move the edge:
- 3:2 or 6:5 payment for a natural blackjack;
- dealer standing or hitting on soft 17;
- number of decks;
- doubling restrictions;
- double after split;
- resplitting rules;
- surrender availability;
- dealer peek or no-hole-card procedure.
The edge is the combined result of the full rule set and a specified player strategy. Changing one line of the rules can alter expected value without changing the game's name.
Baccarat uses a different mechanism. Banker wins slightly more often than Player under the drawing rules, so the traditional game charges commission on winning Banker wagers. Commission-free versions replace that charge with a special push or reduced-payment rule. The operator is not removing the edge; it is changing where the pricing appears.
Pai Gow Poker may use a commission, copy-hand advantage, banker priority, or commission-free push condition. Carnival games frequently combine a base wager with optional side bets whose paytables create much larger margins.
The design process has to state its assumptions
A defensible game-math model identifies:
- The complete outcome space or a valid simulation method.
- The probability of each outcome under the dealing or random process.
- Every payout, collection, push, commission, and rounding rule.
- The permitted player decisions.
- The strategy assumed for those decisions.
- Any banker rotation, jackpot contribution, or progressive component.
- The unit used for the edge calculation.
Without those assumptions, a single published percentage can be misleading.
For example, “the blackjack edge is 0.5%” is incomplete. Which deck count? Which payout? Which rules? Which strategy? “The side bet pays 100 to 1” is also incomplete. How often does it win, and what are the other awards?
A paytable should be evaluated as a probability-weighted whole, not by its most attractive top prize.
Pushes, commissions, and conditional rules
Pushes return the initial wager and contribute zero net payoff for that outcome. They can slow the rate at which wagers resolve without eliminating the edge on resolved action.
A commission reduces a winning payoff. For a $100 Banker win with 5% commission, the net profit is $95 rather than $100. The probability model stays the same, but the payout term in expected value changes.
Conditional rules can move value only in specified states. Examples include:
- half return on an even-money roulette wager when zero appears;
- surrender of half a blackjack wager before the hand is completed;
- a commission-free Pai Gow push when the dealer makes a named hand;
- reduced payment for a particular baccarat Banker result.
The exact frequency of the triggering state and the amount returned determine the effect. A rule that sounds generous may apply too rarely to offset a harsher rule elsewhere.
From math submission to an approved game
A game supplier typically documents rules, paytables, equipment, mathematical analysis, and operational procedure. Independent testing or regulatory review may verify that the implementation matches the submitted model. The casino then chooses among versions it is permitted to offer; it does not normally rewrite the probabilities or payouts during live play.
The Nevada Gaming Control Board maintains an approved-games rules library showing how many named variants and side-bet configurations exist. The list demonstrates why “same game” is not enough information: the approved rule document and paytable matter.
Within a property, game selection involves mathematics, regulation, competition, customer tolerance, equipment, training, surveillance, and floor strategy. A high-edge wager that players reject can be less commercially useful than a lower-edge game with strong demand and sustainable play.
What casinos can change without changing the edge
Several operating decisions affect revenue but not the mathematical percentage of a fixed wager:
| Operating lever | Changes house edge? | Changes expected dollars per hour? |
|---|---|---|
| Higher average bet | No | Yes |
| More decisions per hour | No | Yes |
| More occupied seats | No per wager | Yes for the table |
| Longer playing time | No | Yes |
| Different approved paytable | Usually | Yes |
| Different game rule | Often | Yes |
| More side-bet participation | Changes blended player cost | Yes |
Theoretical win is commonly modeled as:
Theoretical win = average wager × decisions per hour × hours × house edge
Suppose two roulette tables use the same single-zero rules and average $25 per decision. Table A completes 35 decisions per hour, while Table B completes 50.
Table A hourly theory = $25 × 35 × 2.70% ≈ $23.63
Table B hourly theory = $25 × 50 × 2.70% ≈ $33.75
The edge stayed 2.70%. Pace increased the amount of action exposed to it. This is why dealer speed and revenue belongs in an operating analysis, not in the definition of house edge.
Blended edge is not one bet's edge
A player may place a $25 base wager with a low edge and a $5 side bet with a high edge. The session's combined expected loss is the sum of the two wager-level expectations:
Combined expected loss = base wager × base edge + side bet × side-bet edge
If the base wager edge is 1% and the side-bet edge is 10%:
Expected loss per round = $25 × 0.01 + $5 × 0.10 = $0.75
The blended edge on $30 total action is:
$0.75 ÷ $30 = 2.5%
The player may describe the session as playing the 1% game, but the optional chip raises the actual price of the full package.
How an operator should review a proposed rule change
A responsible review asks more than whether the headline edge increased.
- Is the mathematical report complete and reproducible?
- Does the rule create new dealer errors or dispute points?
- Can the payout be handled accurately at live speed?
- Is the change clearly disclosed to players?
- Does it interact with side bets or progressives?
- Is the volatility acceptable for table limits and bankroll exposure?
- Will the market accept the price?
- Is the exact version approved in the jurisdiction?
After launch, actual win should be compared with theory over suitable volume. A short losing period does not prove the edge was set incorrectly, and a large short-term win does not prove the game is overpriced. Variance remains separate from expectation.
For the commercial layer, read how casinos price games. For reporting, use table win, drop, and hold explained. The player-facing definitions are house edge, expected value, and theoretical loss.
House edge is built by rules and payoffs. Casino revenue is built by applying that edge to action. Confusing those two steps leads to poor game comparisons, poor floor decisions, and exaggerated claims about what one night's result means.