Understanding Rtp And House Edge

RTP is a promise about the long run, not a prediction about your session

Every slot, table game and provably fair original at a crypto casino publishes a Return To Player percentage. Sweet Bonanza shows 96.48%. Gates of Olympus shows 96.50%. Stake's Dice and Mines show 99%. BGaming Crash shows 99%. 1429 Uncharted Seas from Thunderkick shows 98.60% — the highest of any mainstream slot. The percentages are real, audited and verifiable, but they describe a game's behaviour over hundreds of millions of spins, not over a single session. The casino's mathematical edge is 100% minus the RTP, and that edge applies to every dollar wagered, not to every deposit. A player who deposits $100 and wagers it ten times before withdrawing has, in expected-value terms, exposed $1,000 to the house edge — not $100. Understanding the math behind RTP, house edge, expected loss per session and the slot-versus-table-game differences is the single most useful intellectual investment a casino player can make. This guide explains the formulas, the worked examples, the games with the cleanest math, and the most common mistakes that compound the house edge unnecessarily.

What RTP and house edge actually mean

RTP, Return To Player, is the percentage of total stakes that a game returns to players in aggregate over its lifetime. A 96% RTP slot returns $96 for every $100 wagered, averaged across all players. The 4% missing is the house edge — the casino's expected income from the game. The number is derived from the game's paytable and probability distribution, calculated by the game developer and verified by independent labs (iTech Labs, eCOGRA, GLI) for slots, or computed openly for provably fair originals.

The key word is aggregate. RTP says nothing about a single player's outcome on a single session. Two players who each wager $1,000 on the same 96% RTP slot will have wildly different results — one may finish $300 up, one $700 down, and the average across thousands of similar pairs converges on the $40 house take per $1,000 wagered.

The variance around the RTP defines the experience. Low-volatility slots (Mega Joker at 99% RTP with very low variance) hit small wins frequently and trend toward the published RTP quickly. High-volatility slots (Nolimit City's Tombstone Slaughter at 96% with a 500,000x max win) deliver almost nothing most of the time and occasionally pay astronomical multipliers. Both have similar long-run RTP but completely different short-run distributions.

How to compute expected loss step by step

1. Identify the game's RTP. Listed in the game's info panel or rules. For Sweet Bonanza, 96.48%. For Stake Dice, 99%.
2. Compute the house edge. 100% minus RTP. Sweet Bonanza: 3.52% house edge. Stake Dice: 1% house edge.
3. Compute total wagered per session. Average bet × number of bets, not the deposit amount. A $1 bet for 200 spins = $200 wagered. A $100 deposit played for that long had $200 of exposure.
4. Multiply house edge by total wagered. Sweet Bonanza at $200 wagered = $7.04 expected loss. Stake Dice at $200 wagered = $2 expected loss. The expected loss is the long-run average; your actual session result varies around this number.
5. Standard deviation. Variance scales differently for different games. Roughly, slot session standard deviation is 5-15x the expected loss for high-volatility games. A $7 expected loss on Sweet Bonanza has a session standard deviation around $35-$100. Most sessions finish within one standard deviation of expectation, but the tails are wide.

Practical examples — three games side by side

Sweet Bonanza (Pragmatic Play), 96.48% RTP. House edge 3.52%. A $1 bet for 1,000 spins = $1,000 wagered. Expected loss = $35.20. Session standard deviation around $200. The 25,000x max payout on Sweet Bonanza 1000 creates the long-tail variance that makes 1,000-spin sessions sometimes finish in the green by hundreds and sometimes down hundreds.

Stake Dice, 99% RTP. House edge 1%. A $1 bet for 1,000 dice rolls = $1,000 wagered. Expected loss = $10. Session standard deviation depends on the roll probability — a 50% win probability strategy has lower variance than a 1% strategy. The 1% edge is among the lowest in the entire casino industry; for pure expected-value, Stake Dice is structurally close to fair.

European Roulette, 97.30% RTP. House edge 2.70% (single zero). A $5 bet for 200 spins = $1,000 wagered. Expected loss = $27. Standard deviation around $80-$150 depending on bet type (outside bets lower variance, single-number higher). The American roulette wheel with double zero has 5.26% edge — nearly double — for the same gameplay, which is why European tables are always preferred where available.

Blackjack basic strategy, 99.5% RTP. House edge 0.5%. A $25 hand for 100 hands = $2,500 wagered. Expected loss = $12.50. Among the cleanest table-game math, but requires correct basic strategy on every decision. Deviation from basic strategy can push the edge to 2-3% even at the same game.

Mega Moolah progressive slot, 88.12% RTP. House edge 11.88% in the base game (the progressive jackpot pool absorbs the rest). One of the lowest base-game RTPs in the major slot catalogue. The progressive jackpot offsets some of the expected loss — €19.4 million record win — but the variance is enormous and the chance of hitting is roughly one in 50 million spins. A $1 bet for 100 spins = $100 wagered with $11.88 expected loss, ignoring the jackpot tail.

Cash or Crash (Evolution live), 99.59% RTP. The highest-RTP live game show on any major platform. House edge 0.41%. The optimal strategy involves a specific cashout pattern; deviations push the effective RTP lower.

BGaming Crash, 99% RTP. House edge 1%. The crash strategy choice (cashout multiplier) does not change the expected return; the long-run is fixed at 1%. The standard deviation is large because of the 1,000,000x max multiplier tail.

Game weighting, contribution and how bonuses change the math

Casino bonus terms typically apply different "game weighting" — slots usually 100%, blackjack 5-25%, roulette 5-20%, live dealer 0-20%. The weighting changes how much of each dollar wagered counts toward bonus clearance. It also indirectly drives players away from the lowest-edge games during bonus play, because clearing 35x WR on slots at 100% weighting takes far fewer dollars than clearing on blackjack at 5% weighting.

The combined math during bonus play: a $100 bonus with 35x WR on slots at 5% house edge requires $3,500 of slot wagering with expected loss of $175 during clearance. The same WR on blackjack at 0.5% house edge requires $3,500 of contribution but $70,000 of actual blackjack wagering, with expected loss $350. The slot path looks worse but the blackjack path actually loses more in absolute terms because of the weighting amplification.

The structural insight is that low-edge table games are good for non-bonus play and bad for clearing bonuses. High-volatility slots are bad for non-bonus EV but acceptable for clearing because they produce large win swings that can pay off the WR via variance.

Common mistakes and red flags

• Confusing RTP with win rate. RTP is the long-run average return. The probability of finishing a session up depends on volatility, not directly on RTP.
• Ignoring max bet during bonus. Most welcome bonuses cap bets at $5-$10. Exceeding once voids the bonus. The bet cap forces longer clearance times, increasing expected loss.
• Switching to "lucky" games. Streak-based intuition does not change the underlying RTP. Every spin on a 96.5% slot has 3.5% expected loss whether the previous spin was a win or a loss.
• Falling for "RTP variants". Some operators allow providers to ship multiple RTP settings of the same game — for example, Nolimit City games can be configured at 84%, 88%, 94%, or 97% RTP. The operator picks; the player rarely sees it. Always check the displayed RTP in the game info panel before playing.
• Treating progressive jackpots as fair value. Jackpots are sometimes booked separately from the base-game RTP. Mega Moolah's 88.12% base RTP underestimates the total return when the progressive is included, but only over millions of spins.
• Ignoring blackjack rule variants. Single deck pays 99.7%; six-deck pays 99.4%; some Asian variant rules push to 99.8%. The rules variant matters more than the table label suggests.
• Comparing slots by RTP alone. 96.5% Sweet Bonanza and 96.5% Gates of Olympus have similar long-run returns but vastly different variance profiles. RTP plus volatility together describe the game.

FAQ

What is a "good" RTP? Above 96% for slots is solid; above 97% is excellent. Stake Originals at 99% are exceptional. Blackjack basic strategy at 99.5% and Cash or Crash at 99.59% are among the cleanest games available.

Does RTP change between sessions? No, the RTP is fixed by the game's paytable. Operators displaying "today's RTP" or similar are showing the empirical short-run result, which fluctuates but converges on the long-run RTP.

Can I beat a 99% RTP game? Yes in the short run; no in the long run. Variance allows winning sessions; the house edge is mathematically certain to take its share over enough wagers.

Is 1% house edge actually winnable? Over thousands of hours of play, the 1% certain loss compounds. A $100,000 cumulative wager at 1% edge has an expected loss of $1,000. The standard deviation around that is wide enough that some players walk away ahead, but the median outcome is the expected loss.

What about RTP-boosted casino events? Some operators run "RTP weekends" where specific slots are configured at higher than usual RTP. Stake and BC.Game both run these. The bump is typically 1-2% and only on selected games — useful but small relative to the variance.

Updated 22 May 2026.