On this page (9 sections)
Play money only — no simulator beats a negative edge
This wheel runs on play money that lives only in your browser tab. Nothing here predicts a real spin, nothing carries over to a real account, and no bet type, wheel view or staking pattern turns roulette's fixed −2.70% house edge into a positive number. That edge is built into the payout table itself, not into anything you do around it.
Play the simulator
This roulette simulator models a European, single-zero wheel: 37 pockets numbered 0 to 36, standard red/black colouring, and the four classic bet types — red or black, odd or even, a dozen, or a single straight number — all settled in play money with a live spin animation.
Pick a bet type and spin.
The 37 pockets and where the edge comes from
A European roulette wheel carries 37 pockets: the numbers 1 through 36, split evenly between red and black, plus a single green zero. That one extra pocket, which belongs to neither colour, dozen nor any outside bet, is the entire source of the house edge. Every payout table on the wheel is built as if there were only 36 numbers to divide up, so the 37th pocket is where the operator's margin sits.
A straight bet on a single number wins once every 37 spins on average and pays 35 units for every 1 wagered, plus the original stake back. A bet on a dozen — 12 consecutive numbers — wins 12 times out of 37 on average and pays 2 units for every 1 wagered. A bet on red, black, odd or even covers 18 numbers and pays even money, 1 unit for every 1 wagered, winning 18 times out of 37 on average.
Worked example: red or black
Win chance is 18/37 ≈ 48.65%. A winning bet returns your stake plus an equal amount, so the payout multiplier on a win is 2x your stake. Expected value per unit staked is (18/37) × 2 − 1 = 36/37 − 1 = −1/37 ≈ −2.70%. That figure is public, well-documented casino maths, not anything specific to this simulator or to any single operator.
Why every bet type carries the same 2.70% edge
Switch this simulator from red to a dozen to a straight number and the win-chance readout swings from 48.65% down to 32.43% down to 2.70% — three completely different experiences of variance. Run the same expected-value arithmetic on all three, though, and each one lands on exactly −1/37. That is not a coincidence: every European roulette payout is set specifically so the 37th pocket represents the same fixed slice of every wager, no matter which numbers you cover.
This is the same underlying idea our Dice simulator demonstrates with a single formula — roulette just spreads the same principle across eight different named bets instead of one slider.
Batch mode: 500 spins against the true odds
A handful of manual spins tells you almost nothing about whether a bet type is behaving correctly — you either hit or you didn't. Click “Run 500 spins” and the simulator plays your selected bet 500 times instantly, using the same bet size each round, then reports the observed hit rate next to the theoretical figure for that bet: 48.65% for red or black, 32.43% for a dozen, 2.70% for a straight number.
Run the same bet type three times in a row
The observed hit rate after 500 spins will usually land within a couple of percentage points of the theoretical figure, but it will rarely match it exactly — sometimes it runs a little hot, sometimes a little cold. That spread is ordinary sampling variance around a fixed probability, and it shrinks the more spins you run, never the other way round.
How to use this without fooling yourself
Four ways to get something real out of a play-money wheel
- Spin manually on red or black a dozen times and track how streaky it feels.
Where people trip up: Even a fair 48.65% bet produces runs of four or five losses in a row often enough that it feels meaningful in the moment — that feeling is exactly what the batch mode exists to correct.
- Switch to a straight number and spin manually the same number of times.
Where people trip up: Watching a 2.70% bet lose repeatedly for a while is the fastest way to feel the difference between a rare, large payout and a frequent, small one.
- Run the 500-spin batch on three different bet types and compare the final bankrolls.
Where people trip up: All three should land in a broadly similar range relative to their starting point, despite radically different hit rates along the way.
- Try doubling your bet after every loss for a short manual run, on purpose.
Where people trip up: It changes how fast your balance moves and how big a single swing can be — it does not change the 2.70% figure driving every spin, which is the point worth seeing once.
Roulette simulator vs. real-money roulette
| This simulator | Real-money roulette | |
|---|---|---|
| Funds at risk | None — play money only | Real deposited funds |
| Wheel and payouts | Standard 37-pocket European layout, shown in full above | Same public maths; specific table rules and minimums vary by operator and are not published here as a fixed figure |
| Fairness verification | Not applicable — no real wager exists to verify | RNG tables use provably fair or licensed RNG certification; live tables use a filmed physical wheel |
| Purpose | Teach the maths behind bet types and the fixed edge | Real wagering with real risk and real rakeback accrual |
The code to try it for real:
Honest pros and cons of practising here
What works about this offer
- Zero financial risk while you learn how the 37th pocket creates the entire house edge
- All eight bet types run the identical fixed-edge maths, shown in full on this page
- Batch mode compresses 500 spins into seconds, which real play never allows
- Useful groundwork before the inverse-CDF maths in Crash or the compounding maths in Mines
What to weigh up first
- Doesn't model any specific operator's live-dealer table rules or limits
- A hot batch run here proves the theoretical odds, not that a real session will match it
- Cannot replicate the pacing and social pull of a real live-dealer table
- Provably fair verification does not apply the way it does for RNG-based Originals
Licence, geo-restrictions and playing safely
Licensing. Access is blocked in a number of countries and US states, and the list is maintained by the operator rather than published as a stable document. The current restricted list and licence details live on the operator’s own terms page — read them before you register.
Your side of it. Crypto casinos settle fast, which is exactly what makes them easy to overplay. Before your first deposit at Rainbet, set a deposit cap and a session timer in the account settings — not in your head.
- Deposit limits and self-exclusion tools live in account settings, not support chat
- A losing session is not a signal to raise the stake — it is a signal to stop
- Bonus turnover targets are not a reason to keep playing past your limit
- Free, confidential help: BeGambleAware and Gambling Therapy
Frequently asked questions
How many pockets does this roulette simulator use?
Thirty-seven — a single zero plus the numbers 1 through 36, the European wheel layout. American double-zero wheels add a second green pocket and a noticeably worse house edge; this simulator does not model that 38-pocket version at all.
Why is the house edge exactly 1/37 on every bet here?
Because every payout on a European wheel is set so the operator keeps one thirty-seventh of every unit wagered, on average, regardless of what you bet on. A straight number wins 1/37 of spins and pays 35 to 1; red or black wins 18/37 of spins and pays 1 to 1; a dozen wins 12/37 of spins and pays 2 to 1. Work out the expected value for any of the three and you get the same −1/37, about −2.70%, every time.
Does betting on a dozen instead of a single number change my odds of profit?
It changes how often you win and by how much, not your long-run expected return. A dozen wins far more often than a straight number for a far smaller payout, and a straight number wins rarely for a much larger one. Both settle on the same −2.70% edge over enough spins, which the batch mode on this page demonstrates directly.
What does the batch mode actually prove?
It spins the wheel 500 times against your selected bet instantly and reports the observed hit rate next to the theoretical probability for that bet — 18/37 for red or black, 12/37 for a dozen, 1/37 for a straight number. Run it a few times and the observed rate will land close to, but rarely exactly on, the theoretical figure — that gap is sampling noise, not a flaw in the wheel.
Can betting patterns like doubling after a loss beat this simulator?
No. Every spin here is independent of the last one, exactly like a real wheel with no memory of previous results. Doubling your bet after a loss changes how quickly your bankroll can swing in either direction; it does not change the −2.70% expectation baked into every single spin, win or lose.
Is the color and number layout on this simulator the real European wheel order?
The pocket numbers, colors and payouts match the standard European roulette layout exactly — that part is public, verifiable maths. The physical wheel order (which number sits next to which) is not simulated because it has zero effect on the outcome probabilities; each spin here draws a uniformly random pocket from the 37, which is mathematically identical to a fair physical wheel.
Is this the same wheel and edge a real crypto casino uses?
The maths — 37 pockets, standard payouts, −2.70% edge — is public and universal to European roulette wherever it is offered. Specific operators may run a live-dealer table, a different table minimum or maximum, or a slightly different rule set on outside bets, and those specifics are not published here as fixed figures.