Casino Odds & House Edge Guide
What "the odds" actually means at a casino
Every casino bet has two sets of odds attached to it, and almost every misunderstanding about gambling math traces back to conflating the two. The first is the true odds — the actual mathematical probability of an outcome, determined by counting cards, numbers, or dice combinations. The second is the payout odds — what the casino actually pays you when you win. These two numbers are never equal on a standard casino bet, and the gap between them, expressed as a percentage of your wager, is the house edge. Understanding that single relationship is the foundation for almost everything else in this guide, and for every other guide in this category: casino strategy is, in large part, the practice of choosing bets where that gap is smaller rather than larger, and managing your money so the gap has less opportunity to compound against you.
This guide exists to make that relationship — and the related concepts of RTP, variance, volatility, expected value, and hit frequency — precise and usable. These terms get used loosely and sometimes interchangeably in casual conversation about gambling, but they describe genuinely different things, and mixing them up leads to bad decisions. By the end of this guide you should be able to look at any casino bet and know exactly what its odds mean, what its cost is, and how to compare it fairly against any other bet on the floor.
True odds vs. payout odds
Take a single-number bet on European roulette as the clearest possible example. There are 37 pockets on the wheel (numbers 1 through 36 plus a single zero), so the true odds against your number landing are 36-to-1 — for every 37 spins, on average, your number hits once and misses 36 times. If the casino paid out at true odds, a winning $1 bet would return $36 in profit, and the game would have no house edge at all; over enough spins, wins and losses would balance out exactly.
Casinos don't pay true odds. A winning straight-up bet on European roulette pays 35-to-1, not 36-to-1. That one-unit shortfall, spread across the true probability of winning, is where the house edge comes from. The math works out to a house edge of 1/37, or about 2.70% — meaning that over a large number of spins, the casino keeps roughly 2.70% of everything wagered on that bet. This pattern — true odds slightly better than payout odds — is how virtually every bet in every casino game is structured. The size of that gap is what separates a comparatively cheap bet from an expensive one, and it's the subject of the rest of this guide. For a deeper, standalone breakdown of this exact mechanic, see how casino odds work.
Some bets make the gap much easier to see than others. Craps odds bets are the one true exception on a standard casino floor — they pay at exactly true odds, with zero house edge, which is why experienced craps players treat them as the best value bet available (though they can only be placed alongside a pass line or don't pass bet, which does carry its own edge). Nearly everything else you'll encounter, from a blackjack hand to a slot spin, embeds a payout gap somewhere in its structure, even when that gap isn't stated as plainly as "35-to-1 instead of 36-to-1."
House edge, defined precisely
House edge is the long-run percentage of every dollar wagered that a casino expects to keep, averaged across a very large number of bets. It is not the amount you'll lose in a single sitting, and it is not a fee taken out of your bankroll upfront — it's a statistical average that only becomes visible over a large enough sample of bets. A 2% house edge doesn't mean you lose 2% of your buy-in and walk away; it means that if you and thousands of players like you each wagered $10,000 total on that bet over time, the casino would keep roughly $200 of every $10,000 wagered, on average, across that entire population.
The formula behind house edge is straightforward once you have the true odds and the payout odds in hand:
House edge = (Probability of losing × Amount lost) − (Probability of winning × Amount won), expressed as a percentage of the amount wagered.
Using the single-number roulette example: the probability of losing is 36/37, losing $1 each time; the probability of winning is 1/37, winning $35. Multiply through: (36/37 × $1) − (1/37 × $35) = $0.973 − $0.946 = $0.027. That $0.027 on a $1 bet is a 2.70% house edge — the exact figure quoted above, arrived at independently through the payout math rather than the pocket count. Both routes to the number agree, which is the kind of internal consistency that confirms the math is being applied correctly.
House edge varies enormously by game and even by specific bet within a single game. A pass line bet in craps carries a 1.41% house edge; a "Big 6" or "Big 8" bet on the very same table carries 9.09% — more than six times as expensive, despite sitting a few feet apart on the same layout. This is why "the house edge of craps" or "the house edge of roulette" as single numbers can be misleading shorthand — the honest answer to "what's the house edge of this game" is almost always "it depends which bet." Casino games with the lowest house edge and casino games with the highest house edge work through this bet-by-bet variation in detail.
RTP: the same math, described from the other direction
Return to Player (RTP) is house edge's mirror image. Where house edge describes what the casino keeps, RTP describes what's returned to players, on average, over the long run. The two numbers always sum to 100%: a game with a 2.70% house edge has a 97.30% RTP, and a game with a 96% RTP has a 4% house edge. Neither figure tells you anything the other doesn't — they're the same underlying statistic, and which one gets used is mostly a matter of industry convention.
RTP is the term you'll almost always see attached to slots and video poker, typically published directly in the game's information screen or paytable, because those games are built around a single blended return figure across all their possible outcomes. House edge is the term more commonly used for table games, where the edge is usually calculated bet-by-bet rather than game-by-game, since a single table like roulette or craps offers many different bets with very different edges attached. Typical online slot RTPs range from about 94% to 98%, meaning the built-in house edge for slots generally runs from roughly 2% up to 6% or more, though the specific figure varies title by title and can occasionally run higher on lower-quality games. For the mechanics of exactly how a slot's RTP is calculated and where to find it, see slot RTP explained; the general relationship between the two terms — including a few situations where the naming convention flips — is covered in full in RTP vs. house edge explained.
Expected value: the concept that ties it together
Expected value (EV) is the dollar-figure expression of house edge, calculated for a specific bet size rather than as an abstract percentage. If a bet carries a 2.70% house edge and you wager $100 on it, your expected value is −$2.70 — not a guaranteed loss of exactly $2.70, but the average outcome if that same $100 bet were repeated a very large number of times. EV is what turns an abstract percentage into a concrete, comparable number: it's how you can say precisely that a $20 bet at a 1.41% house edge (craps pass line, expected cost $0.28) is meaningfully cheaper than a $20 bet at a 5.26% house edge (American roulette even-money bet, expected cost $1.05), holding bet size constant.
EV is also the concept that explains why literally every standard casino wager has a negative expected value from the player's side — that's what "the house edge" means, restated as a dollar figure instead of a percentage. There is no legitimate standard casino bet with positive EV for the player; if there were, it wouldn't remain on the floor for long. The narrow exceptions — certain video poker pay tables under perfect strategy, or specific promotional offers with a positive-EV structure — are covered later in this guide and in can you beat the house?, and they're exceptions precisely because they deviate from how standard bets are priced. The full mechanics of calculating EV, with several more worked examples across different game types, are covered in expected value in casino games explained.
Variance: why individual results don't match the average
House edge, RTP, and EV all describe long-run averages. None of them predict what happens in your next hour of play, and that's not a flaw in the math — it's a separate concept entirely, called variance. Variance measures how much actual results swing above and below the expected average over any given stretch of play, and every casino game has it, even games with a very small house edge.
This is the single most misunderstood relationship in casino math: a low house edge does not mean low variance, and a low-variance session doesn't mean a low house edge. Blackjack has one of the lowest house edges of any standard casino game (around 0.5% under perfect basic strategy) but produces plenty of short-term variance — losing five or six hands in a row is a common, unremarkable event, not a sign that something is statistically wrong. Conversely, a single-number bet on roulette has high variance (you either win 35-to-1 or lose the whole bet, with nothing in between) sitting on top of a fixed 2.70% house edge that doesn't change based on how that variance plays out. The full explanation of how variance behaves over different sample sizes, with concrete win/loss-streak probabilities, is in variance in casino gambling explained.
Volatility: variance's practical cousin, mostly used for slots
Volatility describes essentially the same underlying statistical concept as variance but is used more specifically in the context of how a single game's pay structure is designed — most commonly slots. A low-volatility slot pays out more frequently in smaller amounts, producing a smoother, more predictable session; a high-volatility slot pays out less often but with the potential for much larger individual wins, producing a choppier session with longer dry spells punctuated by occasional big hits. Two slots can carry an identical 96% RTP while feeling completely different to play, purely because of how differently their volatility is designed — one might return that 96% through frequent small wins, the other through rare enormous ones.
Volatility doesn't affect a game's long-run house edge at all; it only affects the shape and rhythm of how that edge gets expressed across a session. This makes it a genuinely useful thing to match to your own bankroll and preferences rather than a mathematical variable to try to optimize — a small bankroll paired with a high-volatility game is a combination that can run out quickly through no fault of the underlying math, simply because of how far results can swing before the long-run average has a chance to assert itself. The full breakdown of volatility bands and how they're typically communicated by providers is in casino volatility explained, and the slots-specific mechanics are covered in slot volatility explained.
Hit frequency: how often a bet actually pays something
Hit frequency is a narrower, more specific statistic than RTP or volatility: the percentage of individual bets or spins that return any payout at all, regardless of size. It's most commonly discussed in the context of slots and video poker, where it's often published alongside or near the RTP figure. A slot with a 40% hit frequency returns some kind of win, on average, about four times in every ten spins; the vast majority of those wins may be small, partial returns rather than meaningful profit, but they still count toward hit frequency even when they don't move the needle on your bankroll.
Hit frequency and RTP are frequently confused because both use "how much comes back" as their basic idea, but they answer different questions. RTP tells you the average dollar amount returned over the long run; hit frequency tells you how often you'll see any return at all, independent of size. A game can have high hit frequency and modest RTP (frequent tiny wins that barely offset losses) or low hit frequency and strong RTP (rare wins that are individually large enough to carry the average). Neither number alone tells you whether a game is a good value — you need RTP for that — but hit frequency is genuinely useful for predicting what a session will feel like moment to moment. The full explanation, along with how it interacts with volatility, is in hit frequency explained.
House edge across major casino games: a comparison
Putting real figures side by side is the fastest way to see how much these numbers actually vary, and why "which game should I play" has a genuinely different answer depending on which bet within that game you're comparing.
| Game / Bet | Approximate House Edge |
|---|---|
| Craps — Odds bet (any) | 0% |
| Blackjack — perfect basic strategy, good rules | ~0.5% |
| Craps — Don't pass / don't come | ~1.36% |
| Craps — Pass line / come | ~1.41% |
| Video poker — full-pay Jacks or Better, perfect strategy | ~0.46% |
| Baccarat — Banker | ~1.06% |
| Baccarat — Player | ~1.24% |
| Craps — Place 6 or 8 | ~1.52% |
| European roulette — any standard bet | ~2.70% |
| French roulette — even-money bets (la partage) | ~1.35% |
| Craps — Place 5 or 9 | ~4.0% |
| American roulette — any standard bet | ~5.26% |
| Slots — typical range | ~2%–6%+ |
| Baccarat — Tie (8-to-1 payout) | ~14.4% |
| Craps — Any seven | ~16.67% |
These figures are approximations that assume standard rules and, where applicable, correct player strategy — actual numbers shift slightly with table-specific rule variations, payout tables, and specific paytable versions. For the exact bet-by-bet breakdowns behind each game, see blackjack house edge explained, roulette house edge explained, craps house edge explained, baccarat house edge explained, slot RTP explained, and video poker odds and RTP explained.
The lowest and highest house edge bets on the floor
At the cheap end, three categories consistently stand out: craps odds bets (0% by definition), blackjack played with correct basic strategy (around 0.5%), and full-pay video poker variants played with perfect strategy (some Jacks or Better and Deuces Wild pay tables sit under 0.5%, and a small number of specific full-pay Deuces Wild tables even cross into positive territory before accounting for the skill required to realize it). These are the bets serious value-focused players gravitate toward, because the mathematical cost of playing them is about as low as a standard casino game gets. The full ranked list, with context on why each one lands where it does, is in casino games with the lowest house edge.
At the expensive end, single-roll and proposition-style bets dominate: baccarat's Tie bet (~14.4%), craps' any-seven and hardways bets (11%–16.67%), and most slot side-bet or bonus-buy features, along with the majority of table-game side bets generally, which commonly run in the 5%–15%+ range. These bets share a common structural trait — a large payout on a comparatively rare outcome — which is exactly the kind of structure that tends to carry a steep built-in cost. That doesn't make them irrational to play occasionally (the appeal of a large, low-probability payout is a legitimate form of entertainment value), but it does mean treating them as a repeated core strategy is one of the more expensive habits available at a casino. The complete list is in casino games with the highest house edge.
How house edge compounds: speed of play matters
A fixed house edge doesn't act on your bankroll all at once — it acts on every individual bet, which means the total number of bets you place has a direct, multiplying effect on your total expected cost, independent of the edge percentage itself. This is why a game's pace of play is a genuinely important, and often overlooked, strategic factor.
Consider two players each starting with $200 and flat betting $10 per round at a 2% house edge. One plays blackjack at a typical live-dealer pace of roughly 40 hands per hour; the other plays a fast-paced online slot at roughly 300 spins per hour. Both have the same house edge and the same bet size, but the slot player's total wagered volume — and therefore expected cost — accumulates more than seven times faster per hour of play. Over a two-hour session, the blackjack player's expected cost is roughly $32 (80 hands × $10 × 2%); the slot player's is roughly $240 (600 spins × $10 × 2%) — a dramatic difference driven entirely by pace, with the house edge percentage held constant. This is one of the clearest practical reasons that session length and bet sizing, covered in the casino bankroll management guide, matter as much as the house edge figure itself.
What this math doesn't tell you
It's worth being direct about the limits of everything above. House edge, RTP, and EV describe averages over a large number of trials — they are genuinely accurate descriptions of long-run behavior, but they say nothing certain about any individual session. A player can sit down at a 0.5% house edge blackjack table and lose their entire bankroll in twenty minutes through ordinary bad variance, and a player can sit down at a 5.26% house edge roulette table and walk away up several hundred dollars through ordinary good variance. Neither outcome contradicts the math; both are entirely consistent with it, because variance is baked into the same statistical framework that produces the average in the first place.
This is also why no adjustment to bet sizing or bet timing — the entire category of progressive betting systems covered in the casino betting strategies guide — can change any of the figures discussed in this guide. House edge is a property of the bet itself: its true odds versus its payout odds. Betting $10 instead of $5, or doubling after a loss, or waiting for a "cold" table, changes nothing about that underlying ratio. The full mathematical explanation of why is in why casino strategies don't beat the house.
Using this math to actually choose a game
The practical payoff of understanding all of the above is a genuinely useful decision framework: compare the house edge of the specific bets you're considering, factor in how fast each game moves (and therefore how quickly that edge compounds), and weigh that against the volatility and hit frequency profile that matches how you actually want a session to feel. A player with a modest bankroll who wants a long, steady session is generally better served by a low house edge, low-to-medium volatility game played at a measured pace; a player looking for the chance at a large win with a smaller amount of money accepts a different trade-off, usually involving higher volatility and, often, a steeper house edge as the price of that chance. Neither preference is a mistake — they're different, equally legitimate ways to spend a gambling budget, once the actual cost of each choice is understood clearly. How to choose a casino game builds this framework out in full, and how casino odds work goes deeper into the true-odds-versus-payout-odds mechanic that underlies every figure in this guide.
Common misconceptions about odds and house edge
A few misunderstandings about this math come up often enough to be worth addressing directly, because each one leads to a genuinely different — and worse — decision than the underlying numbers actually support.
"A game is 'due' for a win after a losing streak." This is the gambler's fallacy, and it's a misunderstanding of independence rather than of house edge specifically. Independent events — a roulette spin, a dice roll, a slot spin — carry no memory of previous outcomes. A roulette wheel that has landed on black ten times in a row has exactly the same probability of landing on black on the eleventh spin as it did on the first, because the wheel doesn't track its own history. House edge describes an average outcome across a very large number of trials; it says nothing about any individual trial being more or less likely based on what came before it.
"A higher house edge always means a worse game to play." House edge measures cost, not entertainment value or enjoyment, and treating it as the only relevant variable misses half the picture. A player who enjoys the pace and format of a higher house-edge game and budgets accordingly isn't making a mathematical error by choosing to play it — they're making an informed trade-off between cost and preference, which is a perfectly reasonable basis for a decision as long as the actual cost is understood going in.
"RTP is guaranteed over any session I personally play." RTP is a long-run average calculated (for slots and video poker, typically via simulation or exact combinatorial calculation) across a number of rounds far larger than any individual player will ever personally complete — often tens or hundreds of millions of spins. Your specific session, whether it's 50 spins or 5,000, will virtually never match the published RTP exactly, and there's no mechanism that "corrects" a session back toward the long-run average as it goes.
"Progressive betting can shift the house edge in my favor." Bet sizing changes the distribution and volatility of your results — sometimes producing a higher chance of a small net win alongside a smaller chance of a larger net loss, or the reverse — but it cannot alter the true-odds-versus-payout-odds ratio that defines the house edge of the underlying bet. This is covered in complete detail, system by system, in the casino betting strategies guide.
"A casino can secretly change the odds mid-session." For regulated, licensed online and physical casinos, game math is certified by independent testing labs and locked at that certified configuration; RNG outcomes and payout tables aren't dynamically adjusted based on a specific player's session, win streak, or account history. This is a common myth worth being clear about, since it often gets used to explain ordinary variance as if it were manipulation.
Frequently asked questions
What's the difference between house edge and RTP? They're the same statistic viewed from opposite sides and always sum to 100%. House edge is what the casino keeps on average; RTP is what's returned to players on average. A 2.70% house edge is the same thing as a 97.30% RTP.
Is a lower house edge always the better choice? For minimizing your expected cost of play, yes — a lower house edge means less expected loss per dollar wagered over time. It doesn't guarantee a better outcome in any individual session, since variance can outweigh a small house-edge difference in the short run.
Does house edge change based on how much I bet? No. House edge is a fixed property of the bet's odds structure, not the wager size. A $5 bet and a $500 bet at the same table carry the identical house edge percentage; only the dollar amount of expected cost scales with the bet size.
Why do some games list RTP and others list house edge? It's mostly a convention split by game type — slots and video poker typically publish RTP directly, while table games are more commonly discussed by house edge, often broken down bet by bet since a single table can offer many different wagers with different edges.
What does a 96% RTP actually mean for my money? Averaged over a very large number of spins or hands, the game returns about 96 cents of every dollar wagered. It does not mean you personally get back 96% of any specific deposit — short-term results vary widely due to variance.
Can variance make a "bad" game beat a "good" game in the short run? Yes, routinely. A higher house-edge game can easily outperform a lower house-edge game over a single session purely through variance. House edge only reliably asserts itself over a large number of bets.
What's the single lowest house edge bet available at a standard casino? Craps odds bets carry a mathematically exact 0% house edge, though they can only be placed as a supplement to a pass line or don't pass bet, which does carry its own edge.
Does a game's volatility affect its house edge? No — they're independent. Two games or two bets within the same game can share an identical house edge while having very different volatility profiles, producing very different session experiences despite costing the same on average.
Is it possible for a casino game to have a positive expected value for the player? Under standard rules and payouts, no — every mainstream bet is designed with a house edge. The rare exceptions involve advantage-play techniques on specific games under specific conditions, covered in can you beat the house?
How do I use house edge to compare two completely different games, like slots and craps? Convert both to a comparable unit — either house edge percentage or expected cost per hour, factoring in typical bets-per-hour for each game type — rather than comparing raw payout tables, which aren't directly comparable across different game structures.
