On this page
RTP means return to player: a game’s theoretical average return under a stated mathematical model. Where the supplier offers several approved configurations, the casino chooses which version to run. Check the percentage in the actual game rather than assuming its highest reference figure. This NZ guide explains the calculation, the effect of different settings and why a short session cannot verify the model.
18+ information and comparisons. Casino links are affiliate links, and we may earn commission. Read our disclosure.
Casino offers and actual return settings
| Casino | Welcome offer and qualifying conditions | Casino link |
|---|---|---|
| Fair Go | Up to $1,000 pokies welcome bonus Code: WELCOME General deposit-bonus terms: $20 minimum; 30× (deposit + bonus); $10 maximum bonus bet, unless the offer states otherwise. Full offer terms — Fair Go | PLAY NOW |
| Ozwin | 400% up to $4,000 Code: WELCOME400 General deposit-bonus terms: $20 minimum; 35× (deposit + bonus); $10 maximum bonus bet, unless the offer states otherwise. Full offer terms — Ozwin | PLAY NOW |
| Ripper Casino | 100% up to $1,000 + 40 spins Code: RC400 Minimum deposit $20; 30× wagering on deposit plus bonus unless specified otherwise; maximum cashout 20× bonus. Bonus-round limit: $5 non-VIP / $10 VIP, including purchases. 40 Fortunate Buddha spins; offer expires 48 hours after redemption. Full offer terms — Ripper Casino | PLAY NOW |
| Uptown Pokies | Deposit $25, play with $125 + 50 spins Code: BANDIT125 The $125 includes your $25 deposit. Offer panel states 40× rollover and no maximum offer cashout. Confirm the wagering base and withdrawal limits in the full terms. Full offer terms — Uptown Pokies | PLAY NOW |
RTP sources and calculations checked 27 October 2026. These are general promotions, not verified RTP readings for every game in these NZ accounts. Confirm the complete title, active mode, installed configuration and accepted offer terms before comparing a particular game.
What RTP measures
RTP expresses expected prizes as a proportion of stakes over the mathematical model. It is not a personal refund rate on a deposit. The glossary separates stakes, prizes, net results and account balances.
A hypothetical 96% game has an expected return of $96 for each $100 of turnover. The corresponding theoretical edge is 4%.
Individual outcomes remain variable. A short session can return nothing, finish ahead or land somewhere between; none of those results alone identifies the installed percentage.
Keep the model attached to its conditions. The same name can cover different editions, approved configurations or modes, and a percentage without that context is incomplete.
Four figures that should not be swapped
- Theoretical RTP: the designed average return under the stated model. It describes an expectation, rather than a promised session refund.
- Measured return: recorded prizes divided by turnover for a defined sample. A small sample does not identify the installed mathematical model.
- House edge: 100% minus RTP for the relevant model. The expectation is different from an exact amount lost on each round.
- Hit frequency: how often the defined paying event occurs. A hit can return less than the complete stake and still produce a net loss.
Volatility and maximum potential are further comparison fields. Neither is another name for the return percentage.
The casino selects an available configuration
A supplier can produce several approved mathematical versions of one title. The operator chooses which available configuration to offer.
This is why copying a highest reference value from a studio page can give an inaccurate description of a casino’s actual game. Read the installed help or information screen.
Nolimit City’s original San Quentin sheet provides a concrete example, listing 96.03%, 94.11% and 90.11% builds. The sheet identifies SanQuentin, SanQuentinDX1 and SanQuentinDX3. Those references do not establish which build is installed in any particular NZ account. Our Nolimit City guide keeps title and mode information together.
A provider logo alone therefore cannot settle the question. Record the complete game name, actual return disclosure, active mode and any visible version information together.
Do not turn a list for one title into a universal option list for the entire supplier. Different releases can have different available models. The software directory distinguishes supplier information from a particular installed game.
A studio page and an operator page can differ
Gameburger’s original 9 Masks of Fire page publishes 96.24%. Grosvenor’s public listing for the named game publishes 94.06%.
These are explicitly source-specific figures, not live readings from the four featured NZ accounts. They demonstrate why the actual supplied version needs checking.
They also do not prove that a particular casino changed an existing player’s game. A difference between two public descriptions is evidence to investigate the edition and settings, not a complete account history.
Our 9 Masks of Fire review keeps the original separate from later named releases, with its source conditions retained.
Converting RTP to a theoretical edge
For the relevant model, subtract the return percentage from 100%. A 96.24% reference corresponds to a 3.76% edge; 94.06% corresponds to 5.94%.
On a hypothetical $1,000 of turnover, those expectations imply losses of $37.60 and $59.40. The difference is $21.80.
This compares mathematical expectation at equal turnover. It does not predict what either real session will lose, nor does it identify the chance of a particular feature. The high-RTP comparison guide explains checking the applicable game version.
Also distinguish percentage points from relative changes. Moving from 96% to 94% RTP is a two-percentage-point reduction in return. The theoretical edge moves from 4% to 6%, a 50% relative increase in that edge.
That relative comparison can explain why a small-looking RTP difference matters, but it still does not mean an individual session’s loss must rise by exactly 50%.
Use turnover, not just the deposit
Turnover is the total staked, including credits won and then staked again. A starting deposit and a turnover total answer different questions.
Suppose a hypothetical player deposits $20, stakes $10 in total and receives $8 in prizes. The first batch has lost $2 net. Staking another $10 raises turnover to $20; returns from that later batch must be recorded separately.
Further replay adds further turnover. The initial deposit does not keep reappearing as a new deposit merely because returned credits are used again.
For a constant 96% model and an assumed $200 of turnover, the theoretical expected loss is $8. Substituting a $20 starting deposit would instead calculate eighty cents and answer a different question.
Use the same units throughout: stakes and prizes in the same currency, percentages converted consistently, and a clearly defined period. Our banking overview explains why currency conversion and account transactions need separate records.
Different amounts of play need a weighted calculation
When comparing a defined mixture of models, calculate each part’s expected prizes from its own turnover and applicable RTP. Then add the prizes and divide by the combined turnover. A simple average gives every percentage equal weight, regardless of its turnover.
In an invented fixed-turnover example, $100 at 96% implies $96 of expected prizes, while $300 at 94% implies $282. Together that is $378 from $400 of stakes, or 94.5% theoretical return for this assumed mixture. Averaging 96% and 94% to get 95% would give the smaller batch too much weight.
This is a calculation under stated assumptions, not a measured session or a forecast that a particular balance will last through those stakes. Keep different editions, modes and currencies identified before combining records.
Calculating a measured session return
The arithmetic is recorded prizes divided by recorded stakes, multiplied by 100.
If a hypothetical session has $80 of stakes and $60 of prizes, its measured return is 75%. The gambling result is a $20 net loss for that sample.
That calculation does not establish a theoretical 75% model. A small result sample can sit far from a game’s long-run expectation.
Deposit, withdrawal and balance figures need careful reconciliation before using them as substitutes for stakes and prizes. Bonus credits, fees, currency conversion or other account transactions can affect balances without being game returns.
Our RTP calculator page explains the relevant arithmetic and input assumptions.
Why a few rounds cannot audit a game
A result record is useful evidence of what happened, but its size and prize distribution affect what can be concluded about the mathematics.
A game whose model includes rare large awards can show substantial sample variation. Missing one of those awards in a small sample is not by itself proof of a fault.
Likewise, one large win does not verify a high installed RTP. It is an outcome under the supplied model, not a disclosure of that model’s full probabilities.
There is no universal number of personal spins that proves every game’s percentage. An assessment needs the design, variability, relevant configuration and appropriate statistical evidence.
Retain the actual help disclosure and round references if investigating a concern. Do not replace them with an unsupported claim based on how a session felt.
RTP and volatility describe different things
Consider two invented $1 games. One returns $1 on 95% of independent draws and zero otherwise. The other returns $10 on 9.5% of independent draws and zero otherwise, also averaging $0.95. These are invented mathematical examples, not real-game prize tables.
Both have 95% theoretical return, yet their visible prize patterns can differ substantially. RTP alone cannot tell you how the return is distributed. The low-volatility and high-volatility guides explain prize patterns separately from the average.
A hit frequency can also count returns below the complete stake. Paying ten cents on a fifty-cent round is a non-zero award with a forty-cent net loss.
Use the volatility guide and mechanics guide to keep average return, frequency and maximum potential separate.
Mode-specific percentages need their conditions
A percentage can refer to ordinary play, purchased entry, a selected mode or a strategy-dependent model. Read the label rather than choosing the highest number displayed.
A higher return percentage for a purchase does not by itself make the purchase less costly in money. Compare the complete amount committed.
For a purely hypothetical example, 96% on a twenty-cent action implies an expected loss of 0.8 cents. A 97% model on a $20 action implies sixty cents.
The second percentage is higher while the expected monetary loss per action is much larger because the cost is larger. These are invented models, not quoted prices or returns for a particular game.
The comparison requires both the applicable percentage and the full stake. Neither a base-stake label nor a return figure is sufficient alone. The feature-buy guide explains the complete purchase price.
Do not mix components with alternative models
A supplier can describe how ordinary play, features and a jackpot contribute to a total model. Those components need not be alternative selectable RTP settings.
The San Quentin sheet lists a 58.74% main-game component and a 37.29% free-spins component, which sum to its 96.03% total. These components describe that breakdown; the 37.29% figure does not establish the return of purchasing free spins, or the component breakdown of the lower-RTP builds.
Do not add a guessed jackpot contribution to a published percentage without knowing what the published value already includes. That can double-count part of the return.
Read the supplier’s definition and the actual configured game information. If the scope is unclear, keep the uncertainty explicit rather than inventing a total.
RTP is not a jackpot ceiling
A fixed maximum, progressive meter and theoretical average answer different questions. A large ceiling alone does not establish a better return model.
Where a jackpot system permits operator-configured upper limits, the casino can choose a ceiling within those settings. Centrally governed network rules can work differently.
Keep a seed, current pool, upper limit, bonus cashout cap and withdrawal limit separate. A withdrawal-per-transfer condition does not become the game’s RTP or maximum prize.
Our jackpot explanation provides the relevant terminology.
A high RTP does not settle bonus eligibility
An accepted promotion can restrict titles, contribution rates, maximum total bets, purchases and cashout conditions. A game being available does not prove it qualifies. The welcome-bonus guide distinguishes the promotional award from the rules for using it.
Do not interpret a return percentage as a guarantee of completing a wagering target. The model does not promise a suitable sequence of results within a particular balance.
For a hypothetical $100 target with 10% contribution, $1 of eligible stake clears ten cents. At that contribution rate, completing a $100 counted target would require $1,000 of otherwise eligible stakes. The arithmetic concerns qualification progress, not the amount expected to be won.
Read the bonus terms and wagering calculation separately from the game’s RTP.
Finding the return disclosure on a phone
Look for the game’s help, rules, paytable or information control; its label and position vary. Read the complete disclosure alongside the mode and full stake, especially where a small screen separates the rules into several panels.
Record the title, percentage and mode together. A lobby thumbnail or shortened promotional card cannot replace the actual game information. Our mobile guide and browser-versus-app comparison explain access and smaller-screen controls without assuming every route uses the same build.
A useful RTP comparison record
Record the complete title, supplier, actual percentage, mode, account currency and date checked. Retain visible build information where available.
If two versions differ, compare like with like. Confirm whether one is a sequel, jackpot edition or special mode before attributing the difference solely to operator selection.
A demo or public catalogue provides background but does not verify another account’s installed configuration. Check the intended game directly and ask official support if its disclosure is unclear. Our free-pokies guide explains the scope of demo information, and how to play covers the paytable and stake controls.
Keep spending and time limits independent of “getting the RTP back”. There is no personal repayment schedule. Our New Zealand support information is available when play becomes difficult to control.
Primary sources and figure scope
Definitions draw on the UK Gambling Commission’s RTP explanation and performance-monitoring guidance, used for mathematical concepts rather than New Zealand law. Configuration examples use Nolimit City’s San Quentin sheet, Gameburger’s 9 Masks product page and the expressly attributed Grosvenor listing. Figures were checked 27 October 2026. Public references remain separate from unverified current NZ-account settings.
Pokies RTP questions
Who chooses the RTP when several approved versions exist?
The casino selects which supplier configuration it runs. Check the actual complete game’s return disclosure instead of assuming the highest reference value.
Does 96% RTP guarantee $96 back from a $100 deposit?
No. It concerns theoretical return on turnover, not a personal deposit refund. Turnover can include replayed credits, and a session can differ from the expectation.
How do I calculate theoretical house edge?
Subtract the applicable RTP from 100%. A hypothetical 96% model has a 4% edge, which is still an expectation rather than an exact loss each round.
How is measured session return calculated?
Divide recorded prizes by recorded stakes and multiply by 100. Keep the sample and units clear; a short result does not identify the theoretical model.
Can two games have the same RTP but different prize patterns?
Yes. The average can be equal while frequency, prize size and variability differ. Compare volatility and other conditions separately.
Does a higher percentage always mean a lower expected loss per action?
No. The full price also matters. A high-percentage model applied to a large purchase can imply more expected monetary loss than a lower-percentage model on a small stake.
Can I add a guessed jackpot contribution to a published RTP?
No. Establish what the published figure includes first. Adding an assumed component can double-count part of the return.
Can I keep playing until the game repays its RTP to me?
There is no personal repayment schedule. Theoretical expectation does not oblige a game to restore an individual’s losses, so keep time and spending limits independent of that idea.



