RTP and house edge: what 96% actually costs you
RTP and house edge are one number stated two ways. A 96% RTP game has a 4% house edge, and it costs you an expected €4 for every €100 you stake — not for every €100 you deposit. Because a deposit gets re-staked many times over, total expected cost routinely exceeds the deposit itself. Variance hides this entirely across a session, so a session tells you nothing; the arithmetic only asserts itself over thousands of bets. And because a bonus wagering requirement multiplies your turnover, that same 4% is what makes a 35x requirement cost more than the bonus it unlocks.
What does 96% RTP actually mean per unit staked?
Return to player is the share of everything staked on a game that comes back as winnings across its full theoretical cycle. At 96% RTP, every €100 staked returns €96 on average and €4 does not come back. The word doing the work is staked. RTP is measured against turnover, not against your deposit, and those two numbers are almost never the same. Deposit €50, spin it through €1,000 of bets over an evening, and you have staked €1,000, not €50 — so the expected cost of that evening is €1,000 x 0.04 = €40, which is 80% of what you put in. The formula is simply expected cost = turnover x (1 minus RTP). One more detail that surprises people: published RTP is calculated over the whole game cycle, including free-spin rounds, feature buys and any progressive contribution. The base game on its own returns less than the headline figure, because part of the 96% is parked in outcomes most sessions never reach.
Is house edge just RTP stated backwards?
Yes, exactly that. House edge is 100% minus RTP, so 96% RTP is a 4% edge, 94% RTP is a 6% edge, and 99% RTP is a 1% edge. They are not two properties to weigh against each other; they are one property with two names, and you can convert freely between them. What makes the edge the more useful form is that it plugs straight into a cost calculation. Expected cost equals turnover multiplied by the edge, so you can price an hour of play before you sit down. Suppose you stake €1 a spin and spin once every six seconds: that is 600 spins, so €600 of turnover an hour, and at a 4% edge an expected €24 an hour. Move to a 94% RTP title and the identical hour costs an expected €36. Note which variables you control. You cannot change the edge, but stake size and speed are entirely yours, and halving either halves the hourly cost.
Why does a real session look nothing like 96%?
Because variance swamps the edge over short runs, and by an enormous margin. The cleanest case to check by hand is a single-zero roulette bet on red: 18 of 37 numbers win, so the edge is 1 minus 36/37, which is 1/37, or 2.70%. Bet €1 a spin for 100 spins and your expected loss is €2.70 — while the standard deviation of that result is about €10, because each bet swings a euro either way and the spread grows with the square root of the number of bets. The noise is roughly four times the signal. On those numbers you finish the 100 spins ahead a little over a third of the time. This is why a winning night proves nothing about a game and a brutal one proves nothing either. Slots are wilder still than roulette, because most of their return is concentrated in rare large wins, so the typical outcome sits well below the average while a few outcomes sit far above it.
How long is the long run, in actual spins?
Long enough that most players never reach it. Expected loss grows in proportion to the number of bets, but the standard deviation grows only with the square root of that number, so the ratio between them improves with the square root of volume. Using the roulette figures above, the expected loss equals one standard deviation when 0.027n equals the square root of n — that is, at n of about 1,370 spins, where both come to roughly €37. Even at that point a one-standard-deviation band is wide, so the edge is merely visible rather than dominant; it takes several times that volume before the outcome reliably resembles the average. For a slot the equivalent threshold is much further out again, because the per-spin spread is several times the stake rather than roughly equal to it. The practical reading is that the house edge is a reliable fact about the operator's book across millions of spins, and not a schedule your own account will follow.
Why does the house edge decide whether a bonus is worth taking?
Because a wagering requirement converts a bonus into an obligation to generate turnover, and turnover is exactly what the edge charges you for. Write D for the deposit, m for the match as a fraction, W for the wagering multiplier and h for the house edge as a fraction, so the bonus is B = mD. Required turnover is W x B when wagering applies to the bonus alone, or W x (D + B) when it applies to deposit plus bonus. The expected cost of producing that turnover is turnover x h, and the realistic value of the offer is B minus that cost. Run the market-standard case: a €100 deposit, a 100% match, 35x on deposit plus bonus. The base is €200, so the turnover owed is 35 x €200 = €7,000. At a 4% edge that is an expected €7,000 x 0.04 = €280 to clear a €100 bonus, giving a realistic value of €100 minus €280, or minus €180. The offer costs more than it pays. Our full table of these is at break-even table, and bonus calculator runs the same maths on your own numbers.
What is the highest wagering requirement that can ever be worth taking?
At a 4% edge, 25x — and that is a hard ceiling, not a rule of thumb. Set the clearing cost equal to the bonus and solve. On a bonus-only base, hWB = B gives W = 1/h, which at h = 0.04 is 25x, and note that the match size cancels out entirely: a 500% match at 40x is exactly as poor a deal as a 50% match at 40x. On a deposit-plus-bonus base, hW(D + B) = mD gives W = m/(h(1 + m)), which is 12.5x at a 100% match and 16.7x at a 200% match. That looks like bigger matches rescue the offer, but m/(1 + m) is always less than 1, so this break-even can never reach 1/h. At a 4% edge, no deposit-plus-bonus offer above 25x is worth taking at any match size whatsoever. Of the six crypto casinos whose own published terms state both a multiplier and its base, all six sit above their own break-even: KatsuBet at 45x on deposit plus bonus, BitStarz at 40x bonus-only, mBit at 40x bonus-only, 7Bit at 35x on deposit plus bonus, Mirax at 35x on deposit plus bonus and Vavada at 35x bonus-only.
Why does game weighting force you onto higher-edge games?
Because the contribution percentage divides your effective edge in a way that more than cancels any advantage a low-edge game offers. The rule is short: effective edge equals the game's edge divided by its contribution weight. Take the €7,000 requirement from the worked example above. On slots weighted at 100%, you stake €7,000 and, at a 4% edge, expect to pay €280 to clear it. Switch to a table game weighted at 10% and every €100 staked credits only €10 towards the requirement, so you must stake €70,000 to accrue €7,000 of credit — and even at a 1% edge that costs an expected €70,000 x 0.01 = €700. The low-edge game costs two and a half times more, because a 1% edge at 10% weighting behaves like a 10% edge. That is precisely what weighting is designed to do. Read the exclusion list and max-bet clause in the same pass: under most published terms, playing an excluded game or breaching the bet cap while wagering voids the bonus and anything won with it, whatever the balance has reached.
Where can you check a game's RTP, and when should you distrust it?
Check it inside the game client itself — the information panel, paytable or rules screen — and cross-check it against the provider's own specification for that title. Two things make the in-client figure the one that counts. First, several studios ship the same game in multiple RTP configurations, so a title with a 96% build may also exist as a materially lower one, and the operator chooses which it runs. The version in front of you is the version you are playing, whatever the review sites say the game returns. Second, an operator's advertised site-wide average RTP is not a number you can act on: it is a blend across a catalogue you will not play evenly, and there is no way to reconstruct it. If a game does not disclose its RTP anywhere in the client, treat the marketing figure as unverified and price it as unknown. Our editorial ranking at global ranking weighs licensing and published bonus terms rather than advertised RTP, because terms can be quoted from an operator's own pages and an RTP claim generally cannot.
Does a lower house edge mean you can win?
No. A lower edge is a slower loss, not a gain, and no staking system, streak-reading or provably-fair proof changes the sign. Watch what recycling does to a bankroll. Stake €100 on a 96% RTP game, re-stake whatever comes back, and repeat ten times: the total turnover is €100 x (1 minus 0.96 to the tenth) divided by 0.04, which is €837.92, and the expected cost is €837.92 x 0.04 = €33.52, leaving about €66. Do the same on a 1% edge game and 0.99 to the tenth is 0.9044, so you expect about €90 left. Ninety is better than sixty-six. Neither is a hundred, and neither trend points upward. The only levers that genuinely change your outcome are how much you stake, how fast, and when you stop — which is why a house edge should be read as the price of the entertainment rather than a hurdle to be beaten. If you want the casino whose terms suit how you actually play, casino matcher matches on terms rather than on headline offers.
FAQ
What does 96% RTP actually mean?
It means the game returns €96 on average for every €100 staked on it, across its full theoretical cycle of millions of spins. The remaining €4 is the house edge. The critical detail is that it applies to turnover, not to your deposit: if you deposit €50 and spin it through €1,000 of bets, you have staked €1,000, so the expected cost is €1,000 x 0.04 = €40, or 80% of your deposit. Published RTP also includes free-spin rounds and jackpot contributions, so the base game alone returns less than the headline figure.
Is a 4% house edge the same as losing 4% of my deposit?
No, and the difference is large. The 4% applies to everything you stake, not to what you deposited, and a deposit is normally staked many times over. Stake a €100 deposit, re-stake what comes back, and repeat ten times, and your total turnover is €837.92 — calculated as €100 x (1 minus 0.96 to the tenth power) divided by 0.04. At a 4% edge that carries an expected cost of €33.52, leaving roughly €66. Play longer, or at bigger stakes, and the expected cost keeps climbing towards and past the deposit itself.
Why did I lose most of my money on a 96% RTP game?
Two reasons, both entirely normal. First, RTP is charged against turnover: re-staking your balance repeatedly means the 4% is deducted again on every cycle, so a €100 balance spun through €2,000 of bets carries an expected €80 cost. Second, variance dominates any short run. For an even-money roulette bet at 100 spins, the expected loss is €2.70 while the standard deviation is around €10 — the swing is roughly four times the edge. Slots swing far harder still, because most of their return sits in rare large wins that most sessions never hit.
Does a higher RTP game mean I will win more often?
No. A higher RTP lowers the long-run cost of play, which means losing more slowly rather than winning. A 99% RTP game carries a 1% edge and a 96% RTP game carries a 4% edge, so a €100 bankroll re-staked ten times leaves an expected €90 on the first and about €66 on the second. Both figures are below €100 and both point downwards. Over any single session, variance decides the outcome entirely, so you can finish ahead on a low-RTP game and well behind on a high-RTP one.
What wagering requirement is worth accepting on a 96% RTP slot?
Nothing above 25x, and usually far less. The break-even multiplier on a bonus-only base is 1 divided by the house edge, which at 4% is 25x, regardless of how large the match is. On a deposit-plus-bonus base it is m/(h(1 + m)), which is 12.5x at a 100% match and 16.7x at a 200% match — and because m/(1 + m) is always below 1, it can never reach 25x at any match size. The market standard of 35x to 45x therefore sits above break-even in every case. A €100 deposit with a 100% match at 35x on deposit plus bonus owes €7,000 of turnover, costing an expected €280 to clear a €100 bonus.
Why do table games only count 10% towards wagering requirements?
Because weighting is what keeps you on the games with the higher house edge. Effective edge equals the game's edge divided by its contribution weight, so a 1% edge at 10% contribution behaves like a 10% edge. Clearing a €7,000 requirement on slots weighted at 100% means staking €7,000, costing an expected €280 at a 4% edge. Clearing the same €7,000 on a table game weighted at 10% means staking €70,000 to earn the credit, costing an expected €700 even at a 1% edge — two and a half times more for the supposedly better game.
How many bets does it take before RTP shows up in my results?
Far more than a typical session. Expected loss grows in proportion to the number of bets while the standard deviation grows only with the square root, so the edge overtakes the noise slowly. On an even-money single-zero roulette bet at a 2.70% edge, the expected loss equals one standard deviation at around 1,370 spins, where both are about €37 on €1 bets — and it takes several times that volume before results reliably resemble the average. Slots need far more again, because the per-spin spread is several times the stake rather than roughly equal to it.