Casino house edge explained

A 4% edge means a $4 expected loss per $100 wagered

House edge is the casino’s expected gain as a percentage of the wager. On a game with a 4% edge, the player’s expected net result is a loss of 4 cents for every dollar bet.

That is not a fee deducted neatly from each spin. Individual outcomes can be losses, break-even results or wins. The edge comes from the probabilities and payouts across all those possibilities.

It also applies to total wagers, including money you win and bet again. A $100 starting balance is not the same as $100 of play.

Turn the percentage into dollars

For a fixed game and bet size, the basic calculation is straightforward:

Expected loss = total wagered × house edge

Suppose you make 500 bets of $2 on a game with a 4% edge. You wager $1,000 in total. Multiply by 0.04 and the expected loss is $40. This is the average across repetitions of that amount of play, not a prediction that your balance will fall by exactly $40.

Playing speed changes the hourly cost

At $2 per bet and a 4% edge, 250 bets create a $20 expected loss. Make 500 bets in the same time and it becomes $40. The game has not changed; you have bought twice as much play.

A time estimate uses your assumed pace, not a universal “bets per hour” figure. It also assumes you can fund every bet. If you stop or run out of money, the volume of play changes.

Expected loss is useful for comparing costs. It is not a safe spending limit. A session can lose much more, including the entire amount you choose to risk.

Where the edge comes from: one roulette bet

Consider a $1 bet on a single number on a standard single-zero roulette wheel. There are 37 equally likely pockets. A winning bet pays 35 to 1: $35 of profit plus your $1 stake back.

(1/37 × $35) + (36/37 × −$1) = −$1/37

The average loss is about $0.027 per $1 bet: a 2.70% house edge. A fair profit payout for a one-in-37 chance would be 36 to 1. Paying 35 to 1 creates the edge.

This is expected value (EV): multiply each net result by its probability, then add the results. A negative player EV means a house advantage. For this bet, RTP is 36/37, or about 97.30%, because the winning total return includes the original stake.

House edge and RTP therefore add to 100% when both use the same wager basis. See the RTP guide for the difference between expected return and a session’s measured return.

Compare the bet and the rules—not just the game name

“Roulette” or “video poker” is not enough information to attach a reliable percentage. The wheel, paytable, commission and player decisions can change the answer.

Figures are rounded and exclude promotions. The game comparisons are supported by Wizard of Odds’ house-edge analysis; the video-poker return assumes its specified 9/6 paytable and optimal strategy.

Blackjack needs a rules-specific calculation: deck count, blackjack payout, dealer behavior and available player actions all matter. A single “blackjack edge” hides those differences.

Also watch the denominator. Some analyses divide expected loss by the initial wager; others divide by all money put into action, including extra bets. A combined craps bet with odds should not be compared casually with the base Pass Line figure.

A winning session does not disprove the edge

Expected value describes an average. Variance describes the spread around it. A low-edge game can still produce a sharp loss, and a high-edge game can produce a winning session.

As independent, identically distributed bets accumulate, average results tend toward their expectation. That is not a promise that a particular session will recover, or that every casino will profit over every period. There is no universal “10,000 bets is enough” rule.

For independent bets with unchanged rules, a losing streak does not improve the next bet. Games dealt from a finite shoe need a different qualification: remaining cards can change the probabilities, but the losing streak itself is not a reason a win is due.

What you can change—and what a betting system cannot

Choose a different price

Better rules, a better paytable or a different bet can reduce the edge. In decision games, using the right strategy can also improve expected return. In random slots, timing a button press or switching after a losing streak does not improve the math.

Choose how much play to buy

Smaller bets and fewer bets reduce expected dollar loss when the percentage stays the same. A spending limit can constrain what you put at risk. Neither changes the underlying house edge.

Do not confuse a progression with an advantage

Doubling after losses changes the pattern of results, not the return per dollar on an unchanged independent bet. Several small recoveries can be offset by a loss large enough to hit your bankroll or the table limit. A betting system cannot remove negative expected value.

Value rewards separately

A reward may offset some cost, but its value depends on eligibility, restrictions and what you would actually pay for it. Playing extra to earn a perk also creates extra wagers. Compare both sides before calling it a saving.

Questions readers ask

Sources & calculation notes