RTP, House Edge and Volatility Explained
A slot that says 96% RTP does not give you back 96% of your money. It gives back 96% of everything wagered on it, averaged over millions of spins, and that difference matters more than any other number in gambling. Here's what RTP, house edge and volatility measure, how to turn them into a dollar cost, and where they stop being useful.
What RTP actually measures
Return to player (RTP) is the share of all money staked on a game that the game pays back as winnings over the long run. A 96% RTP slot is built so that, across its whole lifetime of play, every $100 wagered returns $96 in wins on average. The other $4 stays with the house.
The two words that matter are "wagered" and "long run". RTP isn't calculated per session or per player. It's a theoretical figure, worked out from the game's maths (the paytable and the probability of every outcome) and then checked by testing labs over millions of simulated rounds. Your own results over one evening can land almost anywhere.
RTP also counts every bet, including bets made with money you just won. Deposit $100, spin $1 at a time until it's gone, and you have usually wagered several hundred dollars, because wins get fed back into new spins. That's why "96% back" never means you walk away with $96 of your $100.
House edge: the same number from the other side
House edge is simply what RTP leaves for the casino.
House edge = 100% − RTP
A 96% RTP slot has a 4% house edge. European roulette at 97.30% RTP has a 2.70% edge. A dice game at 99% RTP has a 1% edge.
Table players tend to talk in house edge and slot players in RTP, but they describe one thing. A smaller edge means a slower average loss, and that's all it means. No betting pattern, progression or staking plan changes it. The only exception is games where your decisions matter, like blackjack and video poker. Their published RTP assumes perfect play, so if you play badly the real edge goes up, never down.
Expected loss: turning the edge into dollars
This is the formula worth remembering, because every bonus, cashback and leaderboard decision you make builds on it.
Expected loss = total amount wagered × house edge
Example: 1,000 spins at $1 on a 96% slot is $1,000 wagered. Expected loss = $1,000 × 4% = $40. The same $1,000 wagered on 99% dice has an expected loss of $10. On American roulette (5.26% edge) it's about $53.
The formula uses total wagered, not your deposit. Faster play and bigger bets turn over more money per hour, so the expected cost per hour rises even if the RTP stays the same. A slot played at 600 spins an hour with $2 spins turns over $1,200 an hour. At a 4% edge that's roughly $48 an hour on average. For budgeting, that hourly figure tells you far more than the RTP percentage on its own.
The same maths explains rakeback. It hands back a slice of the edge you paid on each bet, so it lowers your effective house edge. It doesn't add anything on top of a fair game, because the game was never fair to begin with.
Volatility, variance and hit frequency
RTP tells you the average. Volatility, also called variance, tells you how far results swing around that average. Two slots can both be 96% and feel nothing alike.
- Low volatility: frequent small wins and few big ones. Your balance drifts down slowly. Pleasant for long sessions, dull if you want a big hit.
- Medium volatility: a mix of both. Most mainstream slots sit here.
- High volatility: long dry spells, with much of the RTP packed into rare features and multipliers of 1,000x or more. Most sessions end down, and a small number end with a large win.
Hit frequency is the percentage of spins that return anything at all. A slot with a 25% hit frequency pays something on roughly one spin in four, but plenty of those "wins" are smaller than the stake, so your balance still falls. Hit frequency and volatility are different things. A game can hit often and still be highly volatile if the meaningful money sits in the bonus round.
Providers show volatility on their own scales (stars, bolts, 1 to 5) and many list a max win. There's no shared industry standard, so compare games within one provider rather than across providers.
Why the same slot can have several RTPs
Many big providers ship each game in several RTP versions, for example 96.5%, 95.5%, 94%, 92% and sometimes 88% or lower. Graphics, features and volatility look identical. The casino chooses which version to run. Two casinos can offer the "same" slot with four or more percentage points between them, and going from a 4% edge to an 8% edge doubles your expected loss.
Checking takes a minute:
- Open the game, then its info, paytable or "?" screen.
- Look for "theoretical RTP", "return to player" or "expected payout". It's often near the bottom of the rules text.
- If the game has a bonus-buy option, look for a separate RTP for the buy. It can differ slightly from the base game.
- Ignore RTP figures on review sites or the provider's own page. Those usually show the top version, which isn't necessarily the one your casino runs.
If the info screen hides the number or just says "RTP may vary", assume it's on a lower setting. We'd pick another game, or another casino.
RTP of common casino games
Typical figures with standard rules and, for skill games, correct play. Exact numbers depend on the version and rules at your casino, so treat these as a guide and check in-game.
| Game | Typical RTP | House edge | Notes |
|---|---|---|---|
| Online slots | 94–97% (some versions 88–92%) | 3–6% (up to 12%) | Same title can run at different settings |
| Blackjack, basic strategy | about 99.5% | about 0.5% | Rules matter: 6:5 blackjack payouts add about 1.4% to the edge |
| European roulette (single zero) | 97.30% | 2.70% | Same edge on almost every bet |
| American roulette (double zero) | 94.74% | 5.26% | The five-number bet is worse, at 7.89% |
| Baccarat, banker bet | 98.94% | 1.06% | After 5% commission. Player bet 1.24%, tie at 8:1 about 14.4% |
| Video poker (Jacks or Better 9/6) | 99.54% | 0.46% | Only with perfect strategy |
| Dice originals | usually 99% | usually 1% | Edge is the same whatever target you pick |
| Crash originals | 97–99% | 1–3% | Cash-out point changes volatility, not the edge |
| Live game shows | 95–96.5% | 3.5–5% | Varies by bet spot |
Two things stand out. The low-edge games are exactly the ones casinos exclude or count at 5–20% when you're clearing a bonus, which is covered in our guide to wagering requirements. And in-house originals can often be checked round by round with provably fair verification. That proves a result wasn't changed after your bet. It says nothing about how big the edge is.
What RTP can't tell you about one session
RTP is a long-run number and a single session is very short. A slot needs hundreds of thousands of spins before your actual return starts to settle near the stated figure, and high-volatility games need more. Over 300 spins, anything from losing the lot to a big multiple of your balance is possible.
Worked example: you bring $200 and wager a total of $500 in an evening on a high-volatility 96% slot. Expected loss is $500 × 4% = $20. The realistic outcomes look nothing like $20: losing the full $200 is common, finishing roughly level happens, and hitting a feature to finish $1,500 up is rare. The $20 is the average across all players and all evenings, not a forecast for yours.
A few things follow from that:
- A higher RTP doesn't noticeably raise your chance of a winning night. It makes your losses smaller on average over months.
- A slot is never "due". Each spin is independent and the random number generator doesn't remember your losses.
- Your own results don't reveal the RTP. A big win doesn't mean the game is loose, and a bad week doesn't mean it's rigged.
- Over a single session, volatility affects your bankroll more than RTP does. Size your bets so you can ride out the swings of the game you picked. Our bankroll management guide shows how.
Using RTP in practice
Play the highest-RTP version you can find of games you actually enjoy, check the info screen rather than trusting the marketing, and estimate your monthly cost as total wagered × house edge. If that number is more than you're comfortable spending on entertainment, bet smaller or play less. No system shrinks it.
Rebates are the only thing that moves the long-run maths, and only partly. At a 4% edge, $10,000 wagered costs about $400 on average. A 10% cashback on net loss would return around $40 of that if your month lands near the average. Gambbly's Loyalty Cashback pays 10% of the previous month's net loss at partner casinos (15% for VIP members), in USDC, with no wagering requirement and on top of whatever the casino itself gives. It softens losses. It doesn't make gambling profitable, and nothing else does either.
See Gambbly cashback in the app
FAQ
Is 96% RTP good?
For online slots, 96% or a little above is roughly the top setting most mainstream games offer. Anything under 94% is worth avoiding if a higher version of the game exists at another casino.
Does a slot's RTP change while I'm playing?
No. A certified game's RTP is fixed by its maths and doesn't react to you or to recent results. The casino can only choose which certified version to offer.
What is the difference between RTP and hit frequency?
RTP is the average share of stakes paid back over the long run. Hit frequency is how often a spin returns anything at all, and many of those hits pay less than the stake.
Which casino games have the highest RTP?
With correct play, blackjack and good video poker paytables run at around 99.5%, and many crypto dice games are 99%. If you're playing with a bonus, check whether these games count toward wagering.
Can I calculate how much I will lose?
You can calculate the average: total wagered multiplied by the house edge. Your real result will vary a lot around that, especially on high-volatility slots.