Roulette wheels are distinguished mainly by the number of pockets and by rules applied when the ball lands on zero. The visual layout, wheel sequence and game name can vary, but the mathematical comparison begins with 37, 38 or 39 possible outcomes. Each extra zero reduces the return on standard bets because payouts remain based on a 36-number game.
Single-zero roulette
A European-style wheel has numbers 1 through 36 plus one zero, producing 37 pockets. A straight-up number pays 35 to 1. The fair net price would be 36 to 1, so the standard house edge is 1 divided by 37, or about 2.70 percent, on ordinary bets.
The number sequence around the wheel is not numerical. That arrangement matters for neighbour bets and physical-wheel analysis, but it does not change the probability that one fair pocket wins.
Double-zero roulette
An American wheel adds 00 for 38 pockets. Standard outside and inside payouts remain unchanged, raising the house edge to 2 divided by 38, or 5.26 percent. The five-number basket covering 0, 00, 1, 2 and 3 is worse because it usually pays 6 to 1; its house edge is about 7.89 percent.
GambleRoad's comparison of American, European and French roulette provides a broader rule overview. The central wheel-type decision is whether the second zero is present.
Triple-zero roulette
Some games add a third green pocket, often marked 000 or with another symbol. With 39 pockets and the same 35-to-1 straight-up payout, standard bets carry a house edge of 3 divided by 39, or 7.69 percent. The additional pocket is easy to miss when the product name emphasizes a theme or feature.
| Wheel type | Total pockets | Zero pockets | Standard-bet house edge | Return on $100 turnover |
|---|---|---|---|---|
| Single zero | 37 | 1 | 2.70% | About $97.30 |
| Double zero | 38 | 2 | 5.26% | About $94.74 |
| Triple zero | 39 | 3 | 7.69% | About $92.31 |
| Single zero with La Partage | 37 | 1 | 1.35% on even-money bets | About $98.65 on those bets |
The final column is theoretical long-run return, not an expected session cashout.
French roulette is a rule package
French roulette normally uses a single-zero wheel plus French-labelled betting areas and may include La Partage or En Prison. Under La Partage, half of an even-money stake is returned when zero wins, reducing the edge on red/black, odd/even and high/low bets to about 1.35 percent. En Prison can produce a similar long-run effect under its own carryover procedure.
These rules do not improve straight-up or other inside bets. Verify which even-money wagers qualify and how a second zero is handled. GambleRoad's guide to La Partage explains the settlement calculation.
The wheel and betting layout are related but separate
The felt shows numerical groups used for bets; the physical wheel places numbers in a nonconsecutive sequence. A racetrack display lets players choose neighbours around the wheel and named sectors such as Voisins du Zéro. Those bets are combinations of ordinary chips, not a new probability system.
Check the total stake created by a sector button. One click may place several units across multiple numbers, increasing turnover much faster than a single straight-up wager.
Live, automated and RNG wheels
Live roulette uses a physical wheel and video stream. Automated stadium products can use a physical wheel connected to terminals. RNG roulette produces results in software. Regulated games should disclose rules, payouts and likelihood of winning, and random games must use acceptable random outcome generation.
OLG's current online single-zero roulette page lists a 97.29 percent RTP, consistent with a 2.70 percent edge after rounding. The page is available at OLG Roulette. That figure applies to that identified game, not every product carrying the word roulette.
Wheel bias and history displays do not change standard odds
A physical defect can theoretically create bias, but a short sequence of recent results is not evidence. History boards describe previous spins; they do not make missing numbers due. RNG games are specifically designed so prior outcomes do not alter the next result.
Mechanical claims require large, carefully collected samples tied to one wheel and stable conditions. The ordinary player should assume the posted mathematical edge unless credible evidence shows otherwise.
Speed and limits affect total cost
A low minimum does not guarantee low exposure. Rapid automated games, repeat-bet controls and sector buttons can create high hourly turnover. The UK Gambling Commission's 2026 casino equipment requirements state that automated real roulette terminals should not exceed fifty games per hour and must display bet cost and prize information. The current requirements are available in the casino equipment standard.
Expected hourly loss is approximately average total stake per spin multiplied by spins per hour and house edge.
French call bets require another layer of stake accounting. Voisins du Zéro, Tiers du Cylindre and Orphelins place different numbers of chips across splits, corners and straight-up numbers. The sector name does not carry one uniform payout. Each component settles according to its ordinary bet type, so the total return depends on exactly where the ball lands.
When comparing live tables, note minimums by bet class. A table may advertise a low overall minimum but require a higher total for racetrack sectors or restrict the maximum on inside bets. Dealer speed, streaming quality and bet-closing time affect usability but not the underlying pocket probabilities. A smoother interface can still be more expensive if it encourages more spins per hour.
For RNG roulette, inspect whether special features add multipliers or altered payouts. A branded game may use a single-zero wheel yet fund multipliers by reducing ordinary straight-up payouts or changing another rule. Compare the complete paytable with standard roulette rather than classifying the game by pocket count alone.
Zero-rule value should be calculated from the exact procedure. Under La Partage, half the even-money stake is returned immediately. Under En Prison, the stake may remain for another spin, and later treatment can vary. The two rules can produce similar theoretical value under common implementations, but they create different cash-flow and dispute questions. Read the table's rule rather than assuming the labels are equivalent.
Electronic displays may round theoretical percentages. Recalculate from pocket count and payout when the stated value is unclear. For standard straight-up bets, the calculation is exact and does not require a simulation.
Mobile interfaces can hide the full table or zero pockets behind a compact view. Open the rules before placing a repeat bet, especially when moving from a familiar single-zero table to a branded or triple-zero version.
The pocket count should be visible in the help screen or wheel image before play begins.
Roulette wheel selection checklist
- Count the zero pockets rather than relying on the game name.
- Avoid the American five-number basket when comparing standard wagers.
- Confirm whether La Partage or En Prison applies to even-money bets.
- Check the full stake created by neighbour and sector buttons.
- Distinguish live physical, automated physical and RNG versions.
- Set a turnover limit based on stake and expected spins per hour.
- Do not treat recent numbers as a forecast of the next spin.
Roulette wheel types are straightforward to compare. More zero pockets mean a higher standard edge, while player-favourable French rules can reduce the cost of specific even-money bets.