RTP and house edge describe the same reality from opposite sides. If a crash game has an RTP of 97%, its house edge is 3%. Over the long run, the game returns 97% of all bets placed and keeps 3% as the operator's margin. This article explains what those numbers mean, what they don't mean, and why the "long run" matters more than the headline percentage.
What is RTP?
RTP stands for Return to Player. It is a percentage that describes how much of all money wagered on a game is returned to players over the long run.
If a crash game has an RTP of 97%, it means: for every $1,000,000 wagered across all players over millions of rounds, the game returns $970,000 and keeps $30,000 as operator revenue.
RTP is always expressed as a percentage. It is always less than 100% (otherwise the operator would lose money). And it is always a long-run average, not a promise about any individual player's experience.
Every crash game has an RTP. It is determined by the mathematical formula that converts random seeds into crash multipliers. The formula includes a parameter, the house edge, that ensures the game retains a specific percentage of all bets. In a provably fair game, this formula is public and rounds can be checked. In a certified-RNG game such as Spaceman, a testing laboratory checks the generator, and players cannot check a round themselves.
What is house edge, and how does it relate to RTP?
House edge is the complement of RTP. They always sum to 100%.
| RTP | House Edge | Meaning |
|---|---|---|
| 99% | 1% | Operator keeps $1 per $100 wagered |
| 97% | 3% | Operator keeps $3 per $100 wagered |
| 95% | 5% | Operator keeps $5 per $100 wagered |
| 90% | 10% | Operator keeps $10 per $100 wagered |
The gambling industry prefers "RTP" in player-facing materials because 97% sounds generous. Mathematicians and analysts often prefer "house edge" because it directly quantifies the operator's take.
Neither term is more honest than the other. They are two perspectives on the same number. But when evaluating a game, house edge is the more useful framing because it tells you directly what the game costs you per unit wagered.
Crash game house edges compared to other formats:
| Game | RTP | House Edge | Source |
|---|---|---|---|
| Aviator | 97% | 3% | Provider claim |
| Spaceman | 95.5% | 4.5% | Provider claim |
| Color Game, classic perya rules | 92.13% | 7.87% | Exact maths |
| European roulette | 97.3% | 2.7% | Exact maths |
| Typical online slot | 92-96% | 4-8% | Range of provider claims |
Crash games as a category state competitive RTPs compared to most gambling formats, and a game stated at 99% has an exceptionally low house edge of 1%. But the comparison is misleading if you do not account for speed, which is covered in Crash Game Economics.
Does a 97% RTP mean I'll get 97% of my money back?
No. This is the single most dangerous misunderstanding about RTP.
RTP describes the long-run average across all players and millions of rounds. Your individual experience in a single session will almost certainly differ, often dramatically, from the theoretical average.
Here is why: variance.
In a crash game with 97% RTP, the average round returns $0.97 per $1.00 bet. But individual rounds return either $0 (if you lose) or your bet times the multiplier (if you win). The distribution of outcomes is wildly uneven:
- About 50% of rounds crash below 2.0×, returning less than double or nothing.
- About 1% of rounds reach 100× or higher, returning enormous multiples.
- Most players experience many small losses punctuated by occasional large wins.
In a 100-round session, your actual return could be anywhere from $0 (losing every round) to several thousand percent (catching one or two high multipliers). The 97% average is what emerges when you aggregate all of those sessions across all players across all time.
A concrete example: Imagine 1,000 players each playing 100 rounds at $10 per round. Total wagered: $1,000,000. The game returns $970,000 total. But some players end up at $2,000 (doubled their money), some end up at $0 (lost everything), and the average across all 1,000 is $970. The average player gets 97% back. No individual player is average.
This is why RTP alone is insufficient for evaluating whether a game is "good" or "bad" for you. What matters is: what is your personal loss limit, how long do you play per session, and do you stop when you reach your limit?
How long is the "long run" in a crash game?
The "long run" is the number of rounds required for observed results to converge close to the theoretical RTP. This depends on the game's variance.
For a crash game with 97% RTP, cashing out at 2x every round, the 95% range of the observed return is roughly:
- After 100 rounds: 77% to 117%. Almost any result is possible.
- After 1,000 rounds: 91% to 103%. Closer, but still wide.
- After 10,000 rounds: 95% to 99%. Getting close.
- After 100,000 rounds: 96.4% to 97.6%. Converging on the theoretical value.
Illustrative numbers, not lab measurements. Higher cash-out targets swing more and need more rounds.
This is why the Clash Watchdog lab publishes nothing for a game until it has counted at least 10,000 rounds, and why every figure carries its sample size, its time and a 95% interval. See Evidence Tiers and Understanding RTP.
For an individual player, the practical implication is: your personal session is never the long run. A stated RTP describes the game's design, not your experience. Even when it is accurate, it does not protect you from losing money through normal variance.
How does Clash Watchdog measure real RTP?
The lab compares a game's measured return with the provider's claim. The process:
- Record rounds. Rounds are recorded one by one from official provider demos and public live result streams, by a collector separate from this website.
- Compute the observed return. For a crash game, at fixed cash-out targets such as 1.5x, 2x, 3x, 5x and 10x: the share of rounds reaching the target, times the target.
- Attach the uncertainty. Each figure carries its sample size, the time of the count and a 95% interval.
- Show it beside the claim. Once a game passes 10,000 rounds, its page shows the claim and the measurement side by side, labelled, never blended.
If the claim falls outside the measured interval, that is a finding about the counted rounds, reported with its sample size and re-checked on fresh rounds. It is not a verdict on intent.
| Game | Rounds, as of | Provider claim | Measured at 2x (95% CI) | Against the claim | Answers |
|---|---|---|---|---|---|
| Aviator | 59,276 · 21 Oct 04:36 UTC | 97% | 96.3% (95.5% to 97.1%) | Consistent with the claim | Answers |
| Big Bass Crash | Table paused | 95.5% | No figure yet | No figure yet | Answers |
| Crash X | 68,985 · 21 Oct 04:38 UTC | 96% | 94.2% (93.4% to 94.9%) | Below the claim, under watch | Answers |
| High Flyer | 54,421 · 21 Oct 04:37 UTC | 97% | 96.6% (95.8% to 97.5%) | Consistent with the claim | Answers |
| JetX | 63,342 · 21 Oct 04:39 UTC | 96.2% to 98.9% | 97.4% (96.7% to 98.2%) | Consistent with the claim (range, highest at the lowest cash-out) | Answers |
| Lucky Jet | 62,383 · 21 Oct 04:37 UTC | No published claim | 97.2% (96.4% to 98.0%) | No published claim | Answers |
| Space XY | 66,954 · 21 Oct 04:38 UTC | 96.88% to 98.92% | 97.4% (96.6% to 98.1%) | Consistent with the claim (range, depends on the cash-out strategy) | Answers |
| Spaceman | Table paused | 95.5% | No figure yet | No figure yet | Answers |
Rows are in alphabetical order, not ranked. Where two 95% intervals overlap, the games cannot be ranked on that figure.
The whole interval sits below the claim. With many games compared at once, about one interval in 20 misses by chance, so this is a watch, not a finding.
Past rounds describe behaviour. They do not predict the next round.
Computed from the lab's own records, not typed in; each row gives the time of its count. Each game's page has every figure with its sample size, time and interval.
The key insight: RTP is a testable claim, not a marketing statement. When a provider says "our RTP is 97%," that is a statement that a large count of rounds can test. The is it rigged? check explains how.
Is a higher RTP always better?
From the player's perspective, yes, a higher RTP means a lower cost per unit wagered. A 99% RTP game costs you $1 per $100 wagered; a 95% RTP game costs you $5 per $100 wagered.
But RTP is not the only factor that matters:
Speed matters. What matters is cost per hour of play: rounds per hour times stake times house edge. A 99% game with 15-second rounds, 240 an hour, costs 240 × 1% = 2.4 stakes an hour on average. A 97% game with 60-second rounds, 60 an hour, costs 60 × 3% = 1.8 stakes an hour. A higher RTP can still be the more expensive hour.
Variance matters. Two games can have the same RTP but very different variance profiles. A high-variance game produces longer losing streaks and larger occasional wins; a low-variance game produces more consistent, smaller outcomes. Your tolerance for losing streaks, and your bankroll's ability to survive them, matters more than a 1% difference in RTP.
Evidence matters. A stated RTP is a claim. A measured return with a large sample and a narrow interval is evidence. The number only matters if it is true, which is why the lab counts rounds. See the games list for each game's status.
The more useful question is not "which game has the highest RTP?" It is "what will an hour cost me at my pace and stake, and is that claim backed by a count of rounds?" No RTP makes any round predictable: history describes behaviour. It does not predict the next round.
Was this page useful?
Anonymous: no account and no cookie.