Roulette for New Zealanders: Bets, Rules and House Edge

Updated October 2026

Licensed

Available in US

Fast payouts

18+ Only

See how zero settles eligible bets across roulette variants and why betting systems cannot predict or guarantee a result.

How Roulette Bets Work When Zero Lands

Zero is the point at which an even-money outside bet can be treated differently from other wagers. On a French Roulette table, La Partage or En Prison may apply when zero appears, but only to eligible outside bets that pay even money, such as red or black. Neither rule extends to inside bets or to outside bets with different payouts. The distinction is about what happens to a qualifying stake at zero, not about changing the bet after it has been placed.

French Roulette has a single-zero wheel, like European Roulette. Its distinctive feature here is the availability of rules that alter the settlement of an even-money outside bet when that zero lands. La Partage is exclusive to French Roulette. Table rules therefore matter: the name of a game alone does not establish whether a particular rule is in use, and the treatment of the stake should be clear before betting begins.

La Partage: half returned

Under La Partage, zero does not take the entire eligible stake. Half is returned to the player, while the other half is collected. The bet does not win: the return is a partial recovery of the stake, not a profit or an even-money payout.

For example, if a qualifying bet is placed on red and zero lands, La Partage returns half of that bet and retains the remainder. The same treatment applies to another eligible even-money outside bet, such as odd or even. The colour or grouping selected does not change the rule; eligibility depends on the bet being an even-money outside wager.

This is the clearest case for distinguishing the stake from winnings. A winning even-money bet would produce a payout under the table’s stated terms. La Partage after zero instead leaves only part of the original stake in the player’s balance. Calling that a win would overstate the result. The calculation is simple: half of the stake is recovered, and half is lost.

La Partage does not convert every bet on a French table into a partially refunded wager. A straight-up number bet, for instance, is not an even-money outside bet and does not qualify under this rule. The same limitation applies to other wager types that do not meet the eligibility condition. A table’s French label is not a blanket refund policy.

En Prison: the stake remains at risk

En Prison handles the eligible stake differently. When zero lands, the even-money outside bet is carried forward rather than immediately settled as a partial return. The original stake stays in play for the next spin.

If that next spin resolves in favour of the original selection, the stake is returned, but no additional winnings are paid. In accounting terms, the result is recovery of the amount committed, not a profit. If the next spin loses, the stake is collected. The initial zero therefore postpones the final settlement; it does not erase the possibility of losing the bet.

Slot Strategies for New Zealanders: What Is Verified

Learn why blackjack charts do not provide slot-machine advice, and why denomination alone cannot establish…

That timing difference matters. With La Partage, half the qualifying stake is returned at once and the remainder is lost. With En Prison, the stake is held for another outcome: it may come back without winnings, or it may be lost. The two rules offer different forms of treatment after zero, so the table’s stated rule determines which calculation applies.

These conditions concern only the eligible even-money outside bets. They do not give an inside wager the same carry-forward treatment, and they do not make an unqualified bet eligible simply because it is placed on a French Roulette table. Where a table offers neither rule, neither settlement should be assumed.

Reading the bet in practical terms

The useful comparison is the amount still at risk after zero. La Partage returns half immediately. En Prison keeps the original stake pending the next spin. In both cases, the special treatment is restricted to even-money outside bets; other bet types do not receive these protections under the named rules.

That makes the choice of bet type relevant to the settlement, even before considering the number or colour selected. Someone comparing the best roulette bets should separate two questions: what outcome the wager covers, and what the table does with an eligible stake if zero appears. The second question cannot be answered from the bet name alone; it depends on the rules offered at that table.

No zero rule guarantees a winning result. A real-money bet can lose, including an eligible outside bet under En Prison if the following spin goes against it. Roulette play should not involve risking more than can be afforded to lose. The terms of a partial return or delayed settlement describe how the stake is handled; they do not turn the wager into a reliable way to win.

The New Zealand Rules Behind Online Roulette

Online roulette in New Zealand now sits within a changing legal framework. The Online Casino Gambling Act 2026 created the country’s first domestic licensing system for online casino gambling and took effect on 1 May 2026. Its introduction did not mean that every casino serving New Zealand customers immediately held a local licence. The legislation instead set out a transition before the full regime is expected to be operational in 2027.

The Gambling Act 2003 does not permit online casinos to be based in New Zealand. The new framework changes how online casino services are to be licensed, but it does not turn domestic casino websites into a permitted business. The distinction is between where a provider is based and whether its online service falls within New Zealand’s licensing arrangements.

The Department of Internal Affairs is New Zealand’s primary gambling regulator and oversees the online casino licensing framework. The Gambling Commission has a separate, decision-making role: it hears and determines casino licensing applications, issues online casino licences and approves changes to licence conditions. In practical terms, the Department provides regulatory oversight, while the Commission makes licensing decisions and handles applications. Those responsibilities are related, but they are not interchangeable.

Baccarat Strategy for New Zealanders: Rules and Limits

Learn how card counting differs from staking systems, why a Martingale rule changes wager size…

The transition period

The Act established a transition running from 1 May 2026 to 1 December 2026. Online casinos that were already providing services to people in New Zealand before the Act began may continue operating during that period under the transitional provisions. That permission is time-limited; it should not be confused with a licence issued under the new regime.

Providers that have not applied for a licence must stop conducting online casino gambling in New Zealand by 1 December 2026. A provider’s ability to operate during the transition therefore depends on its status under the transitional provisions and, at the end of the period, whether it has applied for a licence. The date marks a change in the applicable conditions, not a guarantee that every continuing service will ultimately receive approval.

The full licensed regime is not expected to be fully operational until 2027. Unlicensed operation is expected to become an offence from around mid-2027. Providers operating without a licence may face enforcement action and fines of up to NZ$5 million. Those consequences concern providers, not customers: a person playing at an offshore casino without a New Zealand licence does not face a fine or criminal prosecution for that reason.

What the framework means for players

Before the 2026 Act, online casino gambling supplied from outside New Zealand was not licensed or supervised by a New Zealand regulator. A customer wagering through an offshore website in that period had no domestic government recourse. The new framework establishes a route for domestic licensing, but the transition means the status of a service depends on when it began operating and what steps its provider has taken.

An offshore site may remain accessible to New Zealand residents; access by itself does not establish that the site holds a New Zealand licence. Nor does a casino’s use of roulette imagery, an app interface or a bonus offer demonstrate regulatory approval. These are separate matters from licensing status. The relevant question is whether the provider is authorised under the framework when the licensing regime applies to it.

The transition also places a restriction on promotion: online casino gambling must not be advertised to people in New Zealand during this period. This is a provider-side requirement, not a reason to treat promotional material as evidence of a licence. An offer can be visible without proving that the operator has local authorisation.

Licensed operators must implement stricter age and identity verification systems. Verification is therefore part of the regulatory obligations attached to licensed operation, rather than an optional feature that establishes the quality of a roulette game. The rules governing a casino’s permission to operate and the rules governing a particular table are different questions.

For a New Zealand player comparing online roulette services, regulatory status is a matter of timing and authorisation, not a ranking of games or a promise about outcomes. The new system gives the Gambling Commission authority to determine applications and issue licences, with the Department of Internal Affairs overseeing the framework. During the transition, some established providers may continue under transitional permission, while providers that have not applied must leave the market by 1 December 2026. A licence, once issued, is a regulatory status; it does not make a roulette bet predictable or remove the possibility of losing.

Blackjack for Real Money in New Zealand

Learn how online hands, table rules, apps and funding terms shape real-money blackjack; each round…

The House-Edge Calculation: European, American and French Roulette

Roulette’s house edge is the portion of stakes that the game is expected to retain over time under its stated rules. It is a comparison of long-run cost, not a forecast for an individual session. A wheel can produce a win or loss that differs sharply from that expectation; the percentage describes the structure of the game, not the next result.

European Roulette has 37 pockets: numbered pockets and a single zero. Its stated house edge is 2.70%. American Roulette has 38 pockets, adding a double zero alongside the single zero, and its edge is 5.26%. The extra pocket changes the calculation while the standard payout schedule remains in place. That makes the American layout less favourable on the stated edge measure.

A specialist review reports payout percentages of 97.30% for European Roulette and 94.70% for American Roulette. These figures express the same comparison from the player-return side: a higher payout percentage corresponds to a lower house edge. The reported European and American figures therefore align with the difference in their stated edges. They are not promises that a particular stake, session or account will return those proportions.

French Roulette uses the same single-zero layout as European Roulette, but its rules can change the cost for eligible play. When La Partage or En Prison applies, the house edge falls to 1.35%. The lower figure is conditional on those rules being in force; it should not be treated as the edge for every table labelled French Roulette or for every type of wager. Without the relevant rule, the single-zero layout alone does not establish that reduced edge.

For someone comparing the best odds at roulette, the useful order is to check both the wheel and the rules. A single-zero wheel has fewer pockets than an American wheel, but the French advantage in the figures above depends on the special rule as well as the layout. A table label is not a calculation. The stated edge is meaningful only when it matches the actual wheel and rule set.

The basic comparison can be set out plainly:

The figures do not mean that French Roulette pays a special amount on every winning bet. They describe the overall mathematical effect of the eligible rules on the house edge. That distinction matters when comparing a displayed payout with the expected cost of playing: the payout schedule and the rule attached to the zero both contribute to the calculation.

The best roulette payout, considered alone, is not a complete measure of value. A headline return percentage can obscure which rules produce it, while the house edge brings the wheel structure and relevant conditions into one comparison. European Roulette’s 2.70% edge and the review-reported 97.30% payout are two sides of that same stated assessment; American Roulette’s corresponding figures are less favourable. French Roulette’s 1.35% edge is the lowest of these stated figures only when its specified rules apply.

A roulette simulator can display outcomes, but simulated results do not replace the table’s rules or change its mathematical edge. For a comparison that holds up beyond a short run, the relevant checks are the number of pockets, the zero layout, the stated house edge and whether French rules apply. Those terms determine the calculation; a short sequence of spins does not.

What Betting Systems Can—and Cannot—Change

A roulette betting system is a method for deciding how much to stake and when to alter that amount. It organises the money placed on the table; it does not alter the wheel, the rules of a game or the casino’s mathematical advantage. The distinction is important because a disciplined sequence can look like a solution while leaving the underlying cost of play untouched.

Blackjack Strategies for New Zealanders

Learn how hit, stand, double and split decisions depend on hand totals and table rules,…

A flat-stake approach keeps the amount wagered the same from spin to spin. A progression changes the stake according to a chosen rule, often in response to wins or losses. Martingale is a familiar example: the stake is increased after a loss in an attempt to recover earlier losses when a later bet wins. That arithmetic can be described neatly. The result cannot be assured. A progression does not make a losing run impossible, and larger stakes mean more money is exposed before any recovery occurs.

Other systems, including D’Alembert-style progressions or Fibonacci-based staking, use different patterns for increasing and reducing bets. Their rules vary, but the financial function is the same: they redistribute when money is put at risk. They do not change the chance attached to a roulette outcome. Calling one the best roulette betting system therefore confuses a personal budgeting preference with a mathematical advantage.

The same applies to plans built around a target profit or a stop-loss. A target can define when a player intends to leave after a favourable session; a loss limit can cap the amount intended for play. Neither changes the game’s probabilities, and neither guarantees that the session will reach its target before reaching its limit. A limit is a boundary on intended exposure, not a device for turning the next result in a preferred direction.

A useful comparison is between changing a stake and changing a rule. Switching from one staking sequence to another only changes the size and timing of bets. Playing a variant with different rules can change the house edge, because the game itself has changed. That is a change to the terms of the calculation, not a benefit created by a system.

The best online roulette winning system does not exist if “winning” means a method that reliably overcomes the game’s advantage. A system may make stakes easier to track or help keep play within a planned budget. Those are organisational benefits, not evidence that the plan improves the underlying return. A larger stake after losses can make the potential loss grow faster; it cannot make the wager safer simply because the increase follows a formula.

Any staking plan should therefore be judged by the amount it can put at risk, not by the recovery story used to justify it. The affordable loss is the relevant ceiling; money beyond that amount has no place in the plan. When a system depends on continued play or steadily rising stakes, its apparent order can conceal a growing exposure. The calculation remains plain: stake management can control the shape of the risk, but it cannot remove the risk or guarantee a profit.

Roulette Apps: Check the Terms Before the Spin

An app can provide access to roulette, but its interface does not determine the amount ultimately available to withdraw. Where play is tied to a casino bonus, the relevant calculation is set by the offer’s conditions: eligibility, wagering requirements, expiry, game contribution and withdrawal restrictions. Those terms can change the result even when the roulette stake and spin are unchanged.

Blackjack Rules for New Zealanders: How Play Works

Learn how player choices shape a blackjack hand and how the dealer determines the result.…

Bonus money must be wagered 50X within 30 days. Winnings from free spins must be wagered 50 times within 7 days. These are separate requirements, and the applicable one depends on what the promotion supplies. A displayed bonus balance is therefore not the same as a withdrawable balance: the wagering condition must be met within the stated period, subject to the offer’s other terms.

Game contribution matters because the promotion’s rules determine how roulette play counts toward wagering. If roulette contributes only part of a stake, the full amount staked may not count toward the requirement. The offer’s wording should establish which games qualify and how their contribution is calculated. Without that detail, a progress figure cannot be translated into a reliable remaining requirement.

Expiry is another part of the arithmetic. A bonus that has not met its wagering condition by the deadline may no longer be available under the offer. The relevant amount is not simply the initial promotional balance; it is what remains after the wagering condition, eligibility rules, game contribution and expiry have all been applied.

Withdrawal has a direct consequence: a withdrawal request voids all active or pending bonuses. That can remove promotional funds and end the associated bonus terms before they are completed. A balance that includes an active bonus should therefore not be treated as though every displayed amount will survive a withdrawal request.

This applies whether roulette is opened through an app on an iPhone, iPad or Android device, or through another route to the same account. Device compatibility says nothing about bonus eligibility or withdrawal treatment. The offer’s terms, rather than the app’s presentation, govern those conditions.

For anyone comparing roulette apps, the useful distinction is between access to a game and access to funds under a promotion. A free-play mode and a real-money bonus are not interchangeable: the latter can carry wagering conditions and a withdrawal consequence. The final amount available depends on those written conditions, not on the label attached to the app or the size of a promotional balance.

Why No Roulette Strategy Can Predict the Next Spin

A roulette strategy can organise a decision, but it cannot identify the result of the next spin. Each spin is independent: earlier outcomes do not alter the probability of outcomes on a later spin. A run of one colour, a long gap between results or a sequence that appears to form a pattern is not a signal that the wheel must now behave differently.

This distinction matters because a record of past results can look persuasive without having predictive value. A sequence may seem unusual, balanced or overdue, but those descriptions concern what has already happened. They do not change the next spin’s probabilities. Treating a pattern as evidence of what comes next adds a story to the record, not information about the future.

The same limit applies to a best roulette strategy. If “best” means a method that predicts the next result, no strategy can meet that test. A method may determine which outcome to select or how much to stake, but those choices do not make the wheel’s next outcome more knowable. A favourable-looking history is not a forecast.

A best roulette spread cannot solve the prediction problem either. Dividing attention across several outcomes may change what happens to a particular set of bets if they win or lose, but it does not make the underlying spin predictable. More selections do not turn previous results into a guide to future ones. The number of choices on the table is not a measure of how much control a player has over the wheel.

That is also why a results display should be read as a record, not a recommendation. It can show past outcomes; it cannot establish that the next one is more likely to match or reverse them. Decisions based on perceived momentum, overdue results or a supposedly revealing sequence rest on an assumption that the sequence has predictive force. Independence removes that assumption.

Keno Results for New Zealanders in 2026

Understand New Zealand Keno results, match a ticket line to a completed draw, and see…

The practical value of a strategy is therefore limited to the decision it actually controls. It may give a player a consistent way to choose and place bets, but it cannot improve the probability of a future outcome by using previous spins. When a method is presented as able to forecast the wheel, the claim exceeds what the results can support.

For a money decision, the relevant distinction is between choosing a bet and knowing its outcome. The first is within the player’s control; the second is not. Roulette outcomes are random, and every real-money bet can lose. Any amount put at risk should remain within what can be afforded to lose, rather than being justified by a pattern or prediction that the next spin will confirm.

A Practical Check on Roulette Plays and Tips

A useful roulette tip is one that changes the terms of a bet, not one that claims to forecast the wheel. For even-money wagers, French Roulette has the strongest rule-based case: La Partage and En Prison reduce the disadvantage attached to zero. The benefit is limited to eligible even-money bets; it does not extend to every wager on the layout.

Before a spin, the practical check is the table’s stated rules. A French label alone is not enough to establish which zero rule applies. Where La Partage is offered, an eligible even-money bet has a lower effective house edge because zero does not cost the entire stake. En Prison also changes the treatment of that bet after zero. These terms make French Roulette the better choice for even-money play when compared on the relevant rules, but they do not turn the wager into a safe outcome.

The distinction matters because “best plays” can refer to two different things. A player can choose a bet type and a table rule; neither choice makes a particular result more likely. The useful tip is to match the intended wager to a table whose rules explicitly protect that wager. If the table does not state that La Partage or En Prison applies, that protection should not be assumed.

A short pre-play checklist keeps the decision factual:

• Confirm that the game is French Roulette and read the displayed table rules.
• Check whether La Partage, En Prison, or both apply to the intended even-money bet.
• Treat any other wager as outside those protections unless the rules expressly say otherwise.
• Set a stake that can be lost without affecting essential spending.

This is a choice about the cost of play, not a prediction. Roulette outcomes remain random, and every real-money bet can lose. French Roulette offers the best odds for even-money bets because of its special zero rules, but that advantage applies only within their stated scope. No tip can extend it to other bet types or guarantee a return. The calculation is simple: favourable terms reduce the cost attached to a specific risk; they do not remove the risk.

Roulette Numbers Are Not Due

A number that has not appeared for a while is no closer to appearing on the next spin. Roulette does not keep a record that changes the wheel’s probabilities: each spin is independent, and earlier results do not affect the likelihood of a later outcome.

That makes a “due” number a description of past results, not a forecast. A run of spins without a particular number may feel like evidence that it is waiting to arrive. It is not. The next result is not adjusted to compensate for what happened before.

The same applies to numbers that have appeared repeatedly. A recent run does not make another appearance more likely, nor does it make the number less likely. Counting past results can describe a sequence; it cannot turn that sequence into information about the next spin.

Live Dealer Casinos in New Zealand: How to Compare

Learn how real-time tables differ from digital games, and assess rules, shared-table play, bet costs,…

There is no best number in roulette in the sense of one with a higher chance because of its history. Choosing a favourite, selecting numbers that have been absent, or following a list of frequent results changes the selection, not the probability attached to the next outcome. A number can still win, but its past record does not provide a reliable reason to expect that result.

Number-selection methods also need to be kept separate from the terms of a bet. A pattern or personal preference may determine where a stake is placed, but it does not alter how the wheel produces an outcome. The distinction is straightforward: choice belongs to the player; probability does not respond to that choice.

A record of results may be useful for checking what has already happened. It is not a map of what must happen next. “Best numbers” can therefore only mean numbers someone prefers to select, not numbers made more likely by a gap, a streak, or a pattern in earlier spins.

Roulette Variants Change the Cost of Play

A roulette variant changes more than its name: wheel size and special rules affect the expected cost of betting. Mini Roulette illustrates why a smaller wheel is not automatically a cheaper choice. Its 13 pockets come with a stated house edge of 7.89%, higher than the standard single-zero European game.

Mini Roulette’s stated edge falls to 3.85% when La Partage applies. That lower figure is conditional on the rule being available; the variant name alone does not establish which terms govern a table. The rule must be confirmed before play, because the difference changes the cost calculation.

Craps for New Zealanders: Rules, Bets and Odds

Learn how the come-out roll, point, pass-line and don’t-pass bets work, including why some rules…

This is a useful distinction when deciding which roulette game is best for a given budget. A compact wheel may look simpler, but the relevant comparison is the edge under the actual table rules. With Mini Roulette, the listed edge without La Partage is materially higher than the edge when the rule applies. Choosing by appearance or wheel size would ignore the variable that changes the expected cost.

The best strategy for a roulette game cannot alter that underlying calculation. A staking pattern changes how much is placed on each spin, not the variant’s stated house edge. The practical choice is therefore about selecting rules with a lower cost, then setting a stake that can be lost without affecting essential spending. Every real-money bet can lose, and roulette outcomes remain random.

Variant labels are not a substitute for checking rules. For Mini Roulette, the key distinction is whether La Partage is included and applies to the bet in question. Without it, the 7.89% edge remains the relevant figure; with it, the stated edge is 3.85%. That is a terms difference, not a winning system.

Tactics Do Not Turn Roulette into a Guaranteed Win

A roulette tactic can shape how a session is approached, but it cannot make a real-money result certain. A bet still depends on where the wheel stops, and every real-money bet can lose. Calling a plan a tactic does not change that exposure.

Some choices affect the terms of play: the roulette variant, its rules and the type of bet placed. Those details can change how a stake is settled or the game’s cost over time. They do not turn a possible loss into a guaranteed return. A favourable rule is a condition attached to eligible bets, not a promise that any particular spin will pay.

The distinction is between controlling a decision and controlling an outcome. A player can decide whether to place a bet and how much to stake. No sequence of choices can dictate the wheel’s result. Nor does a previous win establish that another win will follow. Spins are independent, so earlier outcomes do not alter the probability of outcomes on later spins.

That limit applies regardless of how disciplined a plan appears. A tactic may set boundaries for participation, but it cannot remove uncertainty from roulette or convert chance into a dependable source of income. The word “best” has no force over the wheel.

The practical test is therefore not whether a method sounds persuasive, but whether its terms are understood and its possible loss is affordable. If the stake cannot be lost without affecting essential spending, the financial condition is already unsuitable. No tactic repairs that shortfall.

Roulette remains a game of chance. The result of a spin is uncertain, and a stake placed on it may be lost. Any account of tactics that omits that final amount at risk is incomplete; any account that promises a guaranteed win is not describing roulette accurately.

Responsible Gambling

Written by the editors at Casino Slots Adviser NZ.