Last reviewed and updated: February 2026. Written and fact-checked by our iGaming research desk. See our ongoing iGaming coverage for the methodology behind our game and operator testing.
Roulette looks simple, and that is exactly the trap. The wheel spins, the ball drops, and the game feels like it rewards intuition. It does not. It rewards the table you sat at.
So this guide is built terms-first. We start with the wheel and the layout, then the bets and their real probabilities, then variants, demo play, live dealer tables, offers and bonus weighting, banking, and the legal position in Australia. Along the way we flag what we could verify and what we could not, because roulette content is unusually full of recycled maths errors.
Who should skip this page? Anyone hoping to find a system that beats the house edge. It does not exist, and we are not going to pretend otherwise. Anyone who wants to understand the cost of a spin before staking a dollar is in the right place.
One more thing worth saying early. Roulette is entertainment with a measurable price tag. Knowing that price is what separates a controlled session from an expensive one.
Taken together, the peer-reviewed literature, regulator standards and operator data show that online roulette is a game whose mechanics and probabilities are mathematically well-understood. Which is precisely why the variant you pick inside an online casino lobby matters far more than any betting pattern you bring with you.
The single most consequential decision is the number of green pockets. A single-zero wheel returns a theoretical 97.30% to players; a double-zero wheel returns 94.74%. That gap translates into a practical difference in bankroll depletion, and it compounds over time.
«A $10 bet on American roulette produces roughly $260 in extra expected losses across 10,000 spins compared with European roulette.» Betting Data Lab, analysis of 100,000+ roulette spins (2024).
The second insight is that fairness in digital roulette is a certification question, not a matter of trust. Regulated random number generators must pass statistical batteries, including chi-square and runs testing, applied to at least 10,000 game outcomes.
«RNGs must be cryptographically strong and resistant to attack using contemporary computing resources.» GLI-19: Interactive Gaming Systems Standard, June 2024. https://gaminglabs.com/gli-standards/
The third insight is the Australian harm context. Australians lose more per capita on gambling than any other nation, equating to about $1,521 per adult.
«15% of Australian adults experience, or are at risk of, gambling-related harm.» AIFS data cited in the Australian Government Response to the House of Representatives inquiry into online gambling (2024 to 2025). https://aifs.gov.au
Fourth, the behavioural profile of roulette specifically deserves attention. Tracking data suggests that roulette in an online context may elicit relatively intense chasing behaviour.
«Roulette sessions showed shorter return intervals after losses (b = −0.13, p < 0.001) and wins (b = −0.08, p < 0.001) than slots sessions.» Behavioural tracking study of between-session chasing on an online gambling platform.
What this means in practice: pick a single-zero table, verify the certification of the game and the operator, decide your session budget before the first spin, and treat any promise of a "winning system" as a red flag.
Roulette is a game of chance in which you bet on which numbered pocket a ball will settle into on a spinning wheel. You place chips on the layout, the dealer spins wheel and ball in opposite directions, and the winning number is simply the pocket where the ball comes to rest. Bets can cover one number, a small cluster, or broad outcomes such as red, black, odd, even, high or low.
That is the whole game. Everything else is detail, and the detail is where the money sits.
A European roulette wheel has 37 pockets: numbers 1 to 36 plus a single green 0. An American wheel has 38 pockets: 1 to 36, 0 and 00. Numbers alternate red and black around the rim in a fixed, non-sequential order designed to distribute high, low, odd and even values evenly. Zero, and double zero where it exists, are green and belong to neither colour group.
That green pocket is the entire source of the house edge. It sits outside red/black, odd/even and high/low, so those bets lose when it lands.
The roulette table layout mirrors the numbers in a three-column grid of twelve rows, with inside bets positions on and between the numbers, and outside bets positions bordering the grid for dozens, columns, colour, parity and high/low. On French and many European tables you will also find a racetrack, an oval representation of the wheel order used for announced bets and neighbour bets.
Only the dealer handles chips once bets are marked. Online, the same sequence is automated by the interface with a visible countdown timer, usually 15 to 25 seconds on a live table.
| Zone | Where it sits | What it covers |
|---|---|---|
| Inside, straight up | On a single number | 1 number |
| Inside, split | On the line between two numbers | 2 numbers |
| Inside, street | At the end of a row | 3 numbers |
| Inside, corner | On the intersection of four numbers | 4 numbers |
| Inside, six line | Between two rows at the edge | 6 numbers |
| Outside, dozen | Boxes marked 1st 12 / 2nd 12 / 3rd 12 | 12 numbers |
| Outside, column | Boxes at the foot of each column (2:1) | 12 numbers |
| Outside, even money | Red/Black, Odd/Even, 1-18/19-36 | 18 numbers |
| Racetrack | Oval beside the layout | Announced and neighbour bets |
The betting menu is easy to learn once you see the split between inside and outside sections. Inside bets are tighter, more volatile and pay more. Outside bets are broader, pay less and hit far more often. Neither group beats the house edge. They simply choose your variance.
The reason even-money bets never quite break even is that green pocket sitting outside every one of those groups. Small pocket, permanent cost.
«Even-money bets on a single-zero wheel run at 18:19 odds; adding a second zero makes them 18:20, a small change with a cumulative effect.» RouletteVision, house edge explainer (2024 to 2026).
The five-number bet is the one wager that is worse than the rest of its own table, and informed players give it a wide berth.
«The 0-00-1-2-3 bet pays 6:1 with a 13.16% win probability, giving a 7.89% house edge, the worst in standard roulette.» Wizard of Odds, American roulette probability tables. https://wizardofodds.com/games/roulette/
| Bet | Numbers covered | Payout | Win probability (European) | Win probability (American) | House edge |
|---|---|---|---|---|---|
| Straight up | 1 | 35:1 | 2.70% | 2.63% | 2.70% / 5.26% |
| Split | 2 | 17:1 | 5.41% | 5.26% | 2.70% / 5.26% |
| Street | 3 | 11:1 | 8.11% | 7.89% | 2.70% / 5.26% |
| Corner | 4 | 8:1 | 10.81% | 10.53% | 2.70% / 5.26% |
| Six line | 6 | 5:1 | 16.22% | 15.79% | 2.70% / 5.26% |
| Five-number | 5 | 6:1 | n/a | 13.16% | 7.89% |
| Dozen / Column | 12 | 2:1 | 32.43% | 31.58% | 2.70% / 5.26% |
| Red/Black, Odd/Even, High/Low | 18 | 1:1 | 48.65% | 47.37% | 2.70% / 5.26% (1.35% with La Partage) |
Example 1, straight up, A$5 on 17 (European). If 17 lands, you receive A$5 × 35 = A$175 in winnings, plus your A$5 stake returned, for A$180 back. Expected value: (1/37 × 175) − (36/37 × 5) = −A$0.135, or 2.70% of your stake.
Example 2, A$10 on red (European). Win 18 times out of 37 for A$10 each; lose 19 times out of 37. Expected value: (18/37 × 10) − (19/37 × 10) = −A$0.27.
Example 3, the same A$10 on red at a French table with La Partage. Zero returns half your stake, so the long-run cost halves to about −A$0.135 per spin. Same bet, same colour, half the cost.
Example 4, A$1 across a six-chip Tiers du Cylindre. Total stake A$6 covering 12 numbers as six splits. A hit pays 17:1 on the winning split (A$17) plus that A$1 back, minus the five losing chips, so net A$12. Probability of a hit: 12/37 = 32.43%.
Variants change the maths by adding or removing green pockets, or by adding rules that soften the impact of zero. That is why the version you choose matters far more than most beginners realise. Different roulette games, genuinely different odds.
| Parameter | European | American | French | Lightning / multiplier |
|---|---|---|---|---|
| Pockets | 37 (1-36 + 0) | 38 (1-36 + 0 + 00) | 37 (1-36 + 0) | 37 (1-36 + 0) |
| House edge | 2.70% | 5.26% | 1.35% on even-money bets | ~2.90% base (provider-dependent) |
| Special rules | None | Surrender on some tables | La Partage / En Prison | Random multipliers up to x500 |
| Straight-up payout | 35:1 | 35:1 | 35:1 | 29:1, up to 499:1 with multiplier |
| Announced bets | Often, via racetrack | Rarely | Standard | Varies |
| Verdict for players | Recommended | Avoid where possible | Optimal choice | High volatility, entertainment-led |
European roulette is the cleanest mainstream version: one zero, 37 pockets, a flat 2.70% edge across every standard bet. French roulette uses the identical wheel but adds two zero-rules that only apply to even-money wagers (Red/Black, Odd/Even, High/Low):
Over a long session the difference is not academic. It is a real saving for players who lean on even-money bets.
«Across €2,000 in total wagers, the gap between a 2.70% and a 1.35% edge is roughly €54 versus €105 in expected losses.» AncientGames.org, roulette variant comparison (2024 to 2026).
Practical rule: if a table shows both La Partage and En Prison as dealer's choice, La Partage is generally the simpler and more predictable option, because En Prison locks your capital for an extra spin without paying winnings when it survives.
American roulette adds a second green pocket. Two losing pockets out of 38 gives 2/38 = 5.26%, nearly double the European figure, and the extra pocket makes every straight-up, dozen and even-money bet slightly harder to land. The wheel order also differs from the European sequence, with 0 and 00 placed opposite each other.
Some American tables apply Surrender, the double-zero analogue of La Partage. When 0 or 00 lands, you lose only half of an even-money bet, cutting the edge on those wagers from 5.26% to 2.63%. Always confirm from the game's rules screen, or ask the dealer at a live table, before assuming it applies.
A handful of operators also list triple-zero roulette, which adds a third green pocket and therefore pushes the house edge above the American double-zero format. Updated: we deliberately do not quote a precise percentage here, because our verified regulatory and mathematical sources do not state one for that format. Treat any figure you see elsewhere as unverified and read the game's own information panel.
Lightning-style games keep the single-zero core but strip the straight-up payout back, typically to 29:1, and use the difference to fund random multipliers of x50 to x500 applied to one to five "struck" numbers each round. The result is a base edge in the region of 2.90% with dramatically higher variance. Bigger potential payouts, not better maths.
Multiplier formats generally apply to inside bets only. Even-money bets pay standard 1:1 and receive no multiplier, which surprises a lot of first-time players.
Demo mode runs the same game logic and the same outcome engine as the paid version, but settles in virtual credits, so there is no cash result. It is the correct place to learn the layout, test the speed of the betting window, rehearse announced bets on the racetrack, and watch how a progression behaves across 200 spins without funding it.
Free play roulette is typically accessible via online casino websites or dedicated game portals without requiring a deposit. Note: access rules differ by operator and jurisdiction, and some sites gate demo play behind registration or geo-checks. That claim reflects common market practice rather than a verified regulatory requirement, so check the specific site's terms.
There is one documented psychological caveat worth naming. Because demo balances reset freely and losses cost nothing, players may overestimate their skill or underappreciate the impact of house edge and variance when they switch to real stakes. Note: we flag this as a widely reported practitioner observation. Robust quantitative evidence specific to demo-to-real transitions in roulette was not available in our verified source set, and we will update this section when it is.
In real-money RNG roulette the outcome is produced by a certified random number generator. Stake size cannot influence the number generated. Each spin is a statistically independent event, and no pattern of prior results alters the next one. Certification programmes require continuous monitoring rather than one-off approval, so a malfunctioning RNG is detected and pulled from service quickly.
«Chi-square evaluation of the most recent 10,000 outcomes is required, with automatic logging of results and entry into a tilt state after two consecutive failures.» Nevada Technical Standard 1, RNG requirements (2024 to 2027). https://gaming.nv.gov
Real-money access adds identity and age verification (18+ in Australian regulated gambling), deposit processing, withdrawal limits and KYC checks before payouts are released.
Disclaimer: roulette real money play is for adults aged 18 and over, and only where that format is lawful for you. Check the rules in your own state or territory, read the casino's terms, and set deposit and session limits before you fund an account. This is general information, not legal or financial advice.
| Factor | Demo roulette | Real-money roulette |
|---|---|---|
| Cost of play | None; virtual credits | Your own funds; losses are real |
| Session length | Unlimited; balance resets | Bounded by bankroll and limits |
| Winnings | Cannot be withdrawn | Withdrawable after KYC |
| Game selection | Often a limited subset of RNG titles | Full RNG catalogue |
| Live dealer tables | Not available | Available |
| Bonuses and wagering | Not applicable | Applicable, with game weighting |
| Best used for | Learning rules, layout, pacing | Playing for real outcomes, within limits |
| Risk of harm | Financially nil; can distort expectations | Real financial and behavioural risk |
Live roulette places a human croupier and a physical wheel in front of you through a real-time video feed from a studio or a land-based casino floor. You still bet on a familiar on-screen layout, but the dealer calls "no more bets", spins a real wheel and reads the real result. Cameras and optical character recognition capture the outcome, and the game server settles it. Which is why the "is the wheel rigged" question largely dissolves: you can watch it turn.
| Element | Live dealer roulette | RNG roulette |
|---|---|---|
| Outcome source | Physical wheel and ball | Certified software RNG |
| Result capture | Cameras plus OCR or sensors | Generated internally |
| Pace | Fixed by the dealer and betting timer | You control each spin |
| Social layer | Dealer and player chat, tipping | Single-player interface |
| Fairness control | Visible process plus audited dealing procedures | Certified RNG plus software audit trail |
| Typical minimum stake | Higher (A$0.20 to A$1+) | Lower (from A$0.10) |
Be sceptical of low-budget imitations. In testing, a few "live" tables turned out to be pre-recorded loops of a dealer and wheel. A genuine table shows current dealer names, a live round ID, real-time chat responses and other players' bet volumes updating spin to spin.
| Parameter | Minimum | Recommended |
|---|---|---|
| Internet speed | From 4 Mbit/s (SD stream) | From 12 Mbit/s (Full HD / 4K) |
| Latency (ping) | Under 100 ms | Under 30 ms |
| Device | Android 8.0 / iOS 12 / PC with Intel Core i3 | Android 12+ / iOS 15+ / PC with Intel i5 or Apple M1 |
| Browser | Chrome, Safari or Firefox with HTML5 and WebGL support | Current release versions |
| Connection type | Stable Wi-Fi | Wired Ethernet or 5G with no data throttling |
A webcam is not required, because the feed runs one way only. If your machine handles video calls smoothly, it will handle a live table. Nothing spoils a session faster than a laggy connection that closes the betting window before your chip lands.
A roulette casino is not judged by the size of its welcome banner. It is judged by what happens after you win. Terms first, promotion second, always in that order.
Most disappointment with casino bonus terms starts with an assumption. Players assume roulette counts. It often does not.
Editorial note, Claire Bennett, senior casino content editor: «If support cannot answer a plain question about roulette weighting in writing, that answer is itself the review.»
Important: no roulette strategy removes the randomness of a spin or the house edge, and none guarantees a win. Set a time limit and a money limit before you play. This information is general in nature and is not a substitute for professional advice. Gambling involves a risk of financial loss. If gambling has stopped feeling controllable, contact Gambling Help Online (gamblinghelponline.org.au) or the national gambling helpline on 1800 858 858.
Three things determine your long-run result: the number of green pockets, the payout ratios, and how much you wager in total. The order of bets, the timing of spins and the size of each individual stake affect variance only. Arranging chips differently does not change long-run expectation.
«Realised house edge measured 2.74% for European and 5.89% for American roulette across a 25,000-spin sample, close to theoretical values.» Betting Data Lab (2024).
Mathematically, every roulette bet has negative expected value under fixed payout rules. No stopping rule, sequence or chip progression can turn a series of negative-expectation bets into a positive-expectation one. A system can only redistribute when losses arrive.
| System | Mechanic | What actually happens | Hard ceiling |
|---|---|---|---|
| Martingale | Double the stake after every loss | Frequent small wins, then one long losing run that erases them | Table maximum and finite bankroll; required stake grows exponentially |
| Fibonacci | Progress stakes along 1-1-2-3-5-8 after losses | Slower escalation, longer recovery ladders | Still needs a win to clear the ladder; drawdowns deepen |
| D'Alembert | +1 unit after a loss, −1 after a win | Gentlest progression, feels controlled | Cannot offset the edge; extended losing streaks still compound |
| Flat betting | Same stake each spin | Most predictable variance, easiest to budget | Expected loss equals edge × total turnover |
The maths in one line: a Martingale sequence starting at A$5 requires A$640 on the eighth consecutive loss and A$1,280 on the ninth. Table maximums and your balance both arrive long before probability "owes" you a win.
Online play removes the friction of travel, closing time and cash withdrawal, which compresses the interval between one loss and the next bet. Research comparing internet and land-based gamblers suggests harmful outcomes are broadly comparable, whether roulette is played online or on a casino floor.
«Internet gamblers were younger, took part in more gambling activities and bet on sport more often than non-internet gamblers.» Survey of 6,682 Australian gamblers on the impact of internet gambling.
Early warning signs worth taking seriously: increasing stakes to recover a loss; playing longer than the time you allotted; depositing a second or third time in one session; hiding sessions from a partner; borrowing to fund play; or feeling relief only while a wheel is spinning.
Casino bonuses are written to protect the operator against low-volatility, low-edge play, and roulette is exactly that. Hence game weighting, which determines how much of each dollar wagered counts towards a wagering requirement.
| Game type | Typical wagering contribution | Why |
|---|---|---|
| Pokies and slots | 100% | Higher edge, high volatility |
| Roulette (RNG) | 0 to 10% | Low edge, even-money bets easy to hedge |
| Live roulette | 0 to 10%, often excluded | Same, plus low-risk table coverage |
| Blackjack | 5 to 20% | Low edge with correct play |
| Video poker and baccarat | 0 to 10% | Low edge |
Worked example. A A$100 bonus with 40× wagering means A$4,000 of qualifying turnover. If roulette contributes 10%, every A$1 you stake counts as A$0.10, so you would need A$40,000 of roulette turnover to clear it. That is why bonus hunting on roulette rarely makes sense.
Terms that void bonuses most often:
Reader's rule of thumb: if you intend to play roulette, decline the casino bonus, or ask support in writing whether roulette contributes before you opt in.
| Method | Deposit speed | Withdrawal speed | Notes |
|---|---|---|---|
| PayID or bank transfer | Minutes | 1 to 3 business days | Widely used in Australia; name on account must match |
| Visa or Mastercard (debit) | Instant | Often not supported for payouts | Credit cards for online gambling are restricted in Australia |
| Neosurf or prepaid voucher | Instant | Not available | Deposit-only; useful for hard budget caps |
| e-wallets | Instant | Hours to 48 hours | Availability varies by operator |
| Cryptocurrency | 10 to 60 minutes | 1 to 24 hours | Network fees and exchange-rate risk apply |
KYC in plain terms: expect to supply photo ID, proof of address dated within three months, and proof of payment ownership before your first withdrawal is released. Submitting documents before you win removes the most common cause of payout delays. Watch for pending periods, minimum and maximum withdrawal limits, and any fee applied to a second withdrawal in the same week.
Disclaimer: This information is general in nature and is not legal advice. Gambling legislation varies between jurisdictions and changes over time.
The Interactive Gambling Act 2001 (IGA) prohibits the provision of certain interactive gambling services to people physically present in Australia, most relevantly online casino games such as roulette, pokies and blackjack. The prohibition targets the supply of the service rather than the individual player, and licensed Australian online wagering is largely limited to sport and race betting plus lotteries. The Australian Communications and Media Authority (ACMA) enforces the Act, issues formal warnings, and asks internet service providers to block non-compliant sites, most of which offer casino-style games like roulette.
«In Q1 2023 the ACMA investigated 219 complaints, identified 20 IGA breaches and referred 56 sites for blocking, most offering casino games.» ACMA, Interactive Gambling Act enforcement report, January to March 2023. https://www.acma.gov.au
What this means for you in practice:
Confirm the operator has a mobile-optimised site or an app distributed through the App Store or Google Play. Never sideload an unknown APK.
Open the site in Chrome or Safari and add it to your home screen, or tap Download, then Install, then Open for a native app.
Register, complete age and identity verification, and set your deposit and session limits before funding.
Fund the account with a method that also supports withdrawals.
Filter the lobby by Roulette, then by European or French, and check the house edge and zero-rule on the info panel.
In landscape mode, use the racetrack for announced bets. Portrait mode usually hides it behind a toggle.
Switch to Wi-Fi or a strong 5G signal before joining a live table, and pause background downloads.
Set a reality-check reminder so a session cannot quietly become two hours long.
| Factor | Mobile | Desktop |
|---|---|---|
| Game range | Slightly narrower; some racetrack views simplified | Full catalogue, richest interfaces |
| Bet placement precision | Tap-and-zoom needed on inside bets | Precise chip placement with a mouse |
| Live stream quality | Adaptive, usually SD to HD | Up to Full HD or 4K with multi-camera views |
| Multi-tabling and statistics | Limited | Full statistics and history panels |
| Convenience | Play anywhere | Fixed location |
| Single account | Yes, same balance across devices | Yes |
So much roulette content recycles the same errors. Here are three we see repeatedly, and the correct version.
| Common claim online | Why it is wrong | Correct statement |
|---|---|---|
| "Straight-up odds are 36:1 in European and 37:1 in American roulette" | Confuses payout with odds against winning | Payout is 35:1. Odds against winning are 36 to 1 (European) and 37 to 1 (American). Win probability: 2.70% or 2.63% |
| "Play European roulette, where the house edge is higher than American, 2.7% versus 5.6%" | Internally contradictory, and 5.6% is wrong | The European edge is lower (2.70%) than the American edge (5.26%), which is exactly why European is the better choice |
| "On a European wheel you have 37 to 1 odds on a single number" | Misstates odds against | 1 winning pocket against 36 losing pockets = 36 to 1; the probability is 1/37 |
The rule that prevents all three errors: always state three separate figures, namely probability (1/37 = 2.70%), odds against (36 to 1) and payout (35 to 1). The gap between odds against and payout is the house edge.
Gambling is for adults 18+ only, and it is your responsibility to comply with the law in your own jurisdiction.
