RTP describes a long-run expected return. House edge expresses the complementary expected share. Volatility describes how widely results vary. A single session cannot establish any of them on its own.
What you’ll take away
- A 97% target RTP does not mean every player receives 97 credits back from 100 credits staked.
- Compare stake-weighted payouts with stake-weighted turnover when measuring observed return.
- Payout caps and rounding can make realised return lower than an uncapped mathematical target.
Start with expected return
Return to player is a property of a game’s payout distribution under specified rules. If a fixed wager has possible gross payouts, multiply each payout by its probability, add the products and divide by the stake. Gross payout includes any returned stake; it is not the same as net profit.
For a simplified 1-credit wager with a 50% chance of returning 1.94 credits and a 50% chance of returning zero, the expected return is 0.97 credits. That is 97% RTP. The two actual outcomes remain 1.94 or zero; neither outcome is a 0.97-credit payout.
The complementary house edge is 3% for that same model. It is an expectation over turnover, not a guaranteed percentage of a player’s initial deposit. A balance can be staked repeatedly, so 100 credits deposited can produce more than 100 credits of turnover.
Read the target, then read the limits
The current Hangar registry publishes the following targets. These are implementation values, not independent lab certificates.
| Game | Target RTP | Complement |
|---|---|---|
| Updraft | 97% | 3.0% |
| Minefield | 97% | 3.0% |
| Dice Duel | 98.5% | 1.5% |
| Plinko | 99% | 1.0% |
A target belongs with its game rules. Minefield floors multipliers to hundredths and caps the payout. Plinko uses rounded paytable entries, so the exact weighted return depends on the selected table. A displayed headline alone does not describe every stake and setting.
The same expectation can feel very different
Consider two illustrative distributions before caps or rounding. Game A pays 1.94 credits half the time. Game B pays 97 credits one time in 100. Both have a 0.97-credit expected payout on a 1-credit stake, but Game B has much rarer and larger wins.
That difference is why expected return and variability should be discussed separately. The example does not assign a certified volatility rating to any Hangar game. In Plinko, changing the risk setting changes the paytable and distribution of outcomes; inspecting those probabilities is more informative than treating the word “high” as a universal rating.
Measure observed return without changing the denominator
Suppose a sample has 10,000 demo credits staked and 9,400 returned. Its observed RTP is 94%, and its gross gaming revenue is 600 credits. That sample does not prove the configured target has changed. Sampling variation, selected settings, payout caps and implementation errors are different explanations that require different checks.
Use the same game, period, currency and bet population for numerator and denominator. Do not average each player’s percentage without weighting by their stakes. Keep promotional payouts separately identifiable and document how they enter the calculation.
Sources & scope
Prepared from the Hangar project’s implementation and documentation. Source links point to the inspected revision, so the examples remain traceable as the project changes. These notes describe a software demo, not a certified real-money service.