6 min read
RTP describes a model, not your next session
Return to player, usually shortened to RTP, expresses the expected share of stakes returned under a game’s mathematical model and stated conditions. It is not the chance of winning one round. It is not a refund promise. A player can lose an entire session budget even when the published percentage is high; another player can finish above the starting balance.
The UK Gambling Commission explains RTP as an average across substantial play, rather than a session entitlement. Its guidance is background on the concept, not a certification of this website or Master Crash. The percentage that matters for any particular game must come from that game’s current rules. We have not independently established an RTP for the linked Master Crash APK.
Count stakes, including money used again
Imagine a fictional wallet opened with 100 units. Ten settled entries of 10 units create 100 units of turnover. If returned units are then used for another ten entries of 10, turnover reaches 200 even though only 100 was initially put into the wallet. Turnover counts accepted stakes; it does not count only new deposits. Keeping those totals separate makes later calculations understandable.
Suppose those twenty entries return 180 units in total. The observed return ratio is 180 ÷ 200 × 100 = 90%. With no fees, bonuses, withdrawals or other adjustments, the net game result is −20 units and the closing balance is 80. Dividing 180 by the original 100-unit deposit would produce 180%, but that is not the return-on-stakes calculation being discussed here.
200 units were staked across the example entries and 180 were returned. The observed ratio is 90 percent; the net result is minus 20 units. This does not establish a theoretical RTP.
Swipe across to see all columns.
| Field | Fictional example | Meaning |
|---|---|---|
| Initial funds | 100 units | Money put into the example wallet |
| Total accepted stakes | 200 units | Includes units used again |
| Total returns | 180 units | Sum of settled returns |
| Observed return ratio | 90% | 180 ÷ 200 × 100 |
| Net game result | −20 units | 180 − 200; excludes other adjustments |
House edge uses the same assumptions as RTP
For a simple model in which RTP includes the full amount returned, the corresponding house edge is 100% minus RTP. A hypothetical 96% RTP therefore corresponds to a 4% edge. Across 500 units of turnover, the model’s expected loss is 20 units. That is an average implied by the assumptions, not a bill deducted at the end and not a maximum possible loss.
Do not apply that subtraction to a percentage labelled hit frequency, bonus contribution or a personal win rate. Those labels measure different things. Also check whether the quoted RTP assumes a particular decision rule or game setting. Without that context, two percentages may not describe comparable conditions. A screenshot containing “96%” but no field name is not enough to complete the calculation responsibly.
The same average can hide different outcomes
Consider two invented ten-entry records, each with 100 units staked. Record A returns 9 units on each entry, for 90 total. Record B returns 90 units once and zero on the other nine entries. Both observed ratios are 90%, yet the sequence and size of the returns are very different. Neither record describes a real game or estimates its future behaviour.
This is why the payout distribution matters alongside an average. BGaming’s explanation of RTP and volatility makes that distinction for slots; it does not establish the distribution of a Crash title. In Crash, do not infer the whole return model from the highest visible multiplier, a few cash outs or the colour of recent results. These observations leave most of the necessary data unspecified.
A personal sample is a record, not an audit
You can calculate an observed ratio from complete settled records, but label the period and scope. Include every accepted stake and the corresponding total return, including zero-return entries. Separate refunds and cancelled entries according to their recorded status. Combining selected winning screenshots while leaving out losses changes the dataset; it does not reveal a more accurate RTP.
A small sample below the advertised figure does not, by itself, prove manipulation. A sample above it does not prove the game is profitable. A proper performance assessment needs the applicable mathematical model, enough observations and an appropriate statistical method. Keeping records can clarify a disputed settlement, but extending play just to make a personal percentage approach a target adds exposure rather than resolving uncertainty.
Read the percentage with the rules beside it
When a game supplies an RTP figure, record the full title, provider, version if shown, date and exact wording. Open the rules rather than relying on a search snippet. If a number belongs to Aviator, JetX, Aviatrix or another provider’s game, keep that product name attached. Similar planes, curves or cash-out buttons do not make the underlying rules interchangeable.
For Master Crash, use the rules and file information available for the actual product rather than importing a competitor’s percentage. The linked calculation guide explains a single settled entry; the history guide explains which record to use. Those tasks are useful even when the overall RTP is not supplied, because a correct account of one entry should not depend on inventing an unverified statistic.
Do not average percentages from unequal stakes
Two invented record groups can have very different amounts behind their percentages. Group A contains 10 units of stakes and 10 returned, so its observed ratio is 100%. Group B contains 90 staked and 45 returned, giving 50%. Averaging those two percentages produces 75%, but it gives the small group the same weight as the group containing nine times as much stake.
Swipe across to see all columns.
| Group | Staked | Returned |
|---|---|---|
| A | 10 | 10 |
| B | 90 | 45 |
| Combined | 100 | 55 |
The combined observed ratio is 55 ÷ 100 × 100 = 55%. Add the underlying stakes and returns before dividing. This fixes the arithmetic for the specified records; it does not establish the game’s theoretical RTP or predict another session. First remove duplicate copies of the same entry, as explained in the round-history guide.
Sources and scope
- UK Gambling Commission — RTP and house edge explained (general background)
- UK Gambling Commission — calculating observed RTP and interpreting sample size
- BGaming — RTP and volatility in slots; conceptual background only
- Crash multiplier calculation
Product-specific controls remain subject to the installed app. Illustrations are explanatory and do not represent measured game results or genuine app screenshots.
How we edit and correct these guides
Quick questions
Does 96% RTP mean I get 96 of my 100 units back?
No. The hypothetical percentage describes a model across play, not a promised refund or session result. Your actual return can be much lower or higher.
Is RTP the same as the probability of a cash out?
No. A return ratio measures returned value relative to stakes. A probability describes how often a specified event occurs under stated rules.
What is Master Crash’s verified RTP?
An independently verified figure has not been supplied for the linked APK. The worked examples here are not Master Crash specifications.
