Free to play, no stakes, no account
Play Roulette Online Free: Spin a Real European Wheel in Your Browser
You can play roulette online free on this page: no account, no stake, no money in any direction. It is a real single-zero wheel rendered live rather than a video of one: 37 pockets, brass frets, and a ball that has to find one of them.
How to play the wheel: place a chip, spin, read the result
The odds the board quotes are the ones a real single-zero table pays. Nothing here involves money.
Place one practice chip
A wheel decides where the ball lands. The layout decides what that means for you, and the two are separate objects: most people only ever meet the second one already covered in other people's chips. This is the single-zero European layout on its own, thirty-six numbers in three rows of twelve with the zero at the head and the outside boxes along the near edge. Put a chip anywhere on it, by clicking the felt or by using the controls, and the board states what that chip covers in pockets and as a percentage before anything spins.
Where the chip lands is the bet, which is the part a payout chart never shows. Squarely inside a number is a straight up. On the line between two numbers is a split, on a corner where four meet is a corner, and on the outer boundary of the grid is a street of three or, on the join between two streets, a six-line of six. Click a line or a corner on the felt and the board names what you have just backed. Every one of those shapes is also in the controls, so you can place any of them from the keyboard.
The figure worth watching is the last one, because it does not move. A straight-up number covers one pocket in thirty-seven and pays 35 to 1. Red covers eighteen and pays even money. The first dozen covers twelve and pays 2 to 1. Cover more of the wheel and you win more often for less; cover less of it and you win rarely for more. In every case the house keeps 2.70%, because every price printed on the layout is calculated as though the zero were not there. Spreading chips across several boxes does not dilute that, and no arrangement of them ever has.
The chips have no denomination and the board keeps no balance. It counts spins and how often your chips covered the pocket, which is the only score that teaches anything.
Spin the wheel
Drag across the wheel to send it, the way a croupier does, or press the button.
No spins yet
Entertainment only. Nothing on this page involves money.
Explore a pocket’s neighbours
Choose any number to reveal the two physical pockets on either side. Spin results select themselves.
Anticlockwise: 26, 3 · Selected: 0 · Clockwise: 32, 15
Uniformly, using your browser's cryptographic random number generator, before the wheel starts moving, so the result owes nothing to how hard you flicked. Every pocket has exactly a 1 in 37 chance on every spin. We do not approximate the ball's physics, because any approximation biases the distribution, and a page arguing about probability has no business shipping a wheel that favours one arc.
Read what the result means
No result yet
Choose one practice bet. This panel will show its coverage, chance and table payout.
No spins yet
Because every price on the layout is short by exactly one pocket. A bet covering k pockets at odds of x to 1 returns (x + 1) times k thirty-sevenths of what it risks, and on this layout (x + 1) times k comes to 36 for every bet on the table: 36 times 1, 2 times 18, 3 times 12. One pocket in thirty-seven is 2.70%, and that is the whole of the house's advantage. It is charged the same way whatever you cover, which is why the choice between these bets is a choice about how the same cost is delivered rather than about how large it is.
Complete bet placement and payout reference
Use this reference when you need every accepted chip position. The practice lab above teaches one active bet at a time.
| Bet | Where the chip goes | Pockets covered | Pays |
|---|---|---|---|
| Straight up | Squarely inside one number. | 1 of 37 | 35 to 1 |
| Split | On the line between two touching numbers. | 2 of 37 | 17 to 1 |
| Street | On the outer edge of a row of three. | 3 of 37 | 11 to 1 |
| Corner | On the point where four numbers meet. | 4 of 37 | 8 to 1 |
| Six line | On the outer edge where two rows meet. | 6 of 37 | 5 to 1 |
| Column | In the 2:1 box at the end of a line of twelve. | 12 of 37 | 2 to 1 |
| Dozen | In the 1st, 2nd or 3rd twelve box. | 12 of 37 | 2 to 1 |
| Red or black | In the diamond or the colour box. | 18 of 37 | 1 to 1 |
| Odd or even | In the odd or even box. | 18 of 37 | 1 to 1 |
| 1–18 or 19–36 | In the low or high box. | 18 of 37 | 1 to 1 |
Why the European wheel is arranged the way it is
The European wheel's pocket order looks arbitrary and is precisely the opposite. Reading clockwise from zero, the sequence alternates red and black the whole way round, keeps high and low numbers evenly spread, and puts consecutive numbers far apart from each other. Nothing about that changes the odds on a single number (every pocket is 1 in 37 regardless), but it is what makes betting on a whole section of the wheel a coherent thing to do.
Those section bets are the reason the sequence is worth knowing. Voisins du zéro covers the seventeen pockets around zero, Tiers du cylindre the twelve opposite it, and Orphelins the eight left over. They are placed on the racetrack rather than the number grid, and they only make sense as arcs of the wheel rather than as groups on the felt.
Complete European wheel order
These are the 37 pockets clockwise from zero. The order is used for neighbour and sector bets; it is not a ranking or a results history.
- 0 green zero
- 32 red
- 15 black
- 19 red
- 4 black
- 21 red
- 2 black
- 25 red
- 17 black
- 34 red
- 6 black
- 27 red
- 13 black
- 36 red
- 11 black
- 30 red
- 8 black
- 23 red
- 10 black
- 5 red
- 24 black
- 16 red
- 33 black
- 1 red
- 20 black
- 14 red
- 31 black
- 9 red
- 22 black
- 18 red
- 29 black
- 7 red
- 28 black
- 12 red
- 35 black
- 3 red
- 26 black
What a free roulette wheel simulator is good for
A free roulette wheel simulator earns its keep in three ways, honestly stated. Learning the layout and the sequence without a stake on the table. Getting used to the pace, which is slower than most people expect and is the main reason a first live table feels rushed. And seeing for yourself that short runs look nothing like their long-run averages. Twenty spins produce streaks that feel meaningful and are not.
What it cannot do is test a betting system. That needs thousands of rounds and an average taken across them, which is precisely what the strategy simulator does. Nor does it tell you anything about a particular casino. For that, the live roulette comparison ranks them on catalogue depth and wheel variants, and the homepage ranks the licensed casinos themselves.
It is worth saying plainly what this wheel is not: it is not a demonstration that you can beat one. A free roulette wheel simulator is a good place to learn the game and a terrible place to form beliefs about winning, because nothing is at stake and no session ends because the money did. Use it to play roulette online free until the layout is second nature, then read the odds guide before you put a pound anywhere.
Lesson complete
What the wheel and layout taught you
The wheel chooses one winning pocket. Your chip position decides whether that pocket is covered. Broader bets cover more pockets and pay less, while every standard European bet keeps the same 2.70% house edge because of zero.
Continue with the odds and payouts guide, or test long-run strategy claims in the betting-system simulator.
Advanced: randomness and prediction
Can previous roulette results predict the next spin?
The practical lesson ends above. This section tests a different claim: whether wheel bias or a hot-and-cold board can make the next pocket predictable.
The wheel was beaten twice, and neither method survives
A physical roulette wheel is a machine, and machines wear unevenly. If one pocket sits a fraction shallower than its neighbours, or the wheel head leans by a hair, some numbers come up slightly more often than 1 in 37. The edge belongs to whoever notices, and twice in the recorded history of the game somebody did.
Joseph Jagger, a Yorkshire engineer who had spent his working life around textile machinery and knew that no machine is built perfectly, went to the Beaux-Arts Casino in Monte Carlo in 1873 and paid six clerks to write down every number its wheels produced. Five behaved. One did not, and he bet the numbers it favoured until the casino worked out what he was doing and began moving the wheels between tables overnight. A century later Gonzalo García-Pelayo did the same arithmetic with a computer, logging thousands of spins at Casino Gran Madrid from 1991. When the casino sued to recover the money, the Spanish courts found he had done nothing unlawful, the Supreme Court eventually putting it as ingenuity and computer technique and nothing more: he had recorded public information and drawn a conclusion from it.
Neither route is open now. Wheels are levelled and rotated on a schedule, pockets are machined to tolerances Jagger's contemporaries could not have held, and casinos run their own bias analysis on far more spins than any player can log by eye. A wheel biased enough to beat is a wheel that gets pulled from the floor.
Against an RNG table the question does not even arise, and it is worth being clear about why. There is no wheel. The animation you are watching is drawn after the number has been chosen, so there is no pocket to wear, no rotor to lean and no physical history for anyone to measure. The wheel on this page works the same way, which is precisely why watching it settle several hundred times tells you about the distribution and nothing whatsoever about the next spin.
Which makes the statistics panel on most live tables an odd thing to include. It shows the last few dozen results, flags hot and cold numbers, and counts how long it has been since red. All of it is accurate and none of it is predictive, because the wheel keeps no record of what it did before. The board is there because players like seeing it, and a number that has not landed in two hundred spins is exactly as likely as the rest to land next.
What the hot and cold board actually shows
The panel invites a strategy, so here are the three it invites, played against one shared run of spins: back whichever number has come up most often in the trailing window, back whichever has come up least on the theory that it is due, or ignore the board entirely and back 17 every time. Each rule stakes one unit on a single number at 35 to 1, and each sees exactly the same spins as the other two.
Hit rate is the measure to watch, because a hot number board makes a claim about hit rates. It says some pockets are landing more often than others and implies that this will carry on. If that were true, the rule that chases them would land more of its bets than the rule that ignores them.
| Spins | Chase the hot number | Back the cold number | Always 17 | Error on the realised return |
|---|---|---|---|---|
| 5,000 | 2.69% | 2.83% | 2.75% | plus or minus 8.3% |
| 100,000 | 2.69% | 2.65% | 2.59% | plus or minus 1.8% |
| 1,000,000 | 2.67% | 2.69% | 2.69% | plus or minus 0.6% |
The realised return is the noisier measure, and the width of that noise is the reason this idea survives contact with experience. Across 5,000 spins a single-number bet carries an error of roughly 8%, which is wide enough for any of the three rules to finish a run apparently in profit. Nothing has beaten the house when that happens. The run is simply too short to show the house edge at all. At 1,000,000 spins the error falls to about 0.6% and the three rules arrive at the same place.
The chart counts how often each pocket landed, plotted as the gap above or below the count you would expect. The shaded band is where a single number's count normally falls. The line above it is how high the highest of all thirty-seven counts normally reaches, which is a different and much larger figure: pick the biggest of thirty-seven numbers and you will beat any one number's typical range about five times in six. The tallest bar on the board sits under that line on almost every run, and that is the entire point. A hot number is what a fair wheel looks like, not a departure from one.
- Choose how many spins to run and how much recent history the board reads.
- Run one shared comparison of hot, cold and fixed-number choices.
- Compare every hit rate with 1 in 37 and read why the changing leaders are not predictions.
Change the board length and run it again. The hot number changes with it, and so does the cold one. That is the tell worth remembering: a genuine property of a wheel would not move about when you alter the size of the window you read it through. Joseph Jagger was not reading a panel of the last few dozen results. He paid six clerks to log every spin for weeks, and the wheel he found was defective in a way that modern maintenance would catch within a day.
Playing within limits
This page costs nothing and risks nothing, which makes it the right place to say the unglamorous part. If you move on to real money, decide the amount before you start, treat it as the price of the evening rather than a stake to win back, and set a deposit limit while you are calm.
GAMSTOP excludes you from every Gambling Commission licensed operator at once. GamCare runs the National Gambling Helpline on 0808 8020 133, any hour, free of charge, and BeGambleAware publishes self-assessment tools and advice for family. Our responsible gambling page gathers all of it together.
Frequently asked questions
Is this a real roulette wheel?
It is real geometry rather than a video or an image: a 0.81 metre single-zero wheel with 37 pockets, frets, a turret and a ball, rendered live in your browser. It is the same wheel the entrance animation on the homepage shows, mounted so you can spin it. Nothing on this page is a static render or a pre-recorded clip; the pocket is drawn afresh and the animation solved anew on every spin, so no two runs land the same way.
Do I need an account or a deposit?
Neither. There is no sign-up, no stake, no balance and no payout on this page. Nothing you do here involves money in any direction, and no result is recorded anywhere outside your own browser tab. Closing the tab clears everything instantly, and reopening the page starts a fresh sequence with no memory of what came before.
How is the winning number decided?
A pocket is drawn uniformly at random using your browser's cryptographic random number generator, and the animation is then solved backwards so the ball lands there. That is the honest way round: a physics approximation would introduce a small bias, and a page about probability should not ship a wheel that quietly favours one arc.
Why not simulate the physics of the ball instead?
Because any approximation of friction, the drop point and the fret bounces skews the distribution, and the skew would be invisible to you and misleading. Drawing the pocket uniformly guarantees every number has exactly a 1 in 37 chance, which is what a real, properly maintained wheel is engineered to achieve.
Can I use this roulette wheel to test a betting system?
For a handful of spins, by hand, yes. For anything conclusive, no: systems only reveal themselves over thousands of rounds, which is what our betting system simulator is for. It runs twenty thousand sessions and charts the average result. Hand-tracked spins also carry your own bias in when you decide to stop, which the simulator’s fixed session length removes entirely.
Why is the number order not 1, 2, 3?
The sequence is engineered so that red and black alternate all the way round, high and low numbers are distributed evenly, and consecutive numbers sit far apart. That arrangement is what makes section bets on the racetrack meaningful, and it is why the wheel order is worth knowing even though it changes no probability.
Does the wheel here have a zero or a double zero?
A single zero, which makes it a European wheel: 37 pockets and a 2.70% house edge. The American wheel adds a double zero for 38 pockets and 5.26%. The wheel comparison works through what that extra pocket costs. This page always uses the single-zero wheel, since that is the version worth practising on, while the double-zero version stays inside the comparison guide’s house-edge figures.
Are hot and cold numbers real?
No. Each spin is independent, so a number that has just appeared three times is exactly as likely next spin as one that has not appeared at all. Spin this wheel twenty times and watch clusters form anyway. That is what randomness looks like, and mistaking it for a pattern is the oldest error at the table.
Why does the wheel drift before a spin, then settle after one?
The drift is what a real wheel does. A 32-inch wheel spun properly takes something close to twenty seconds to come to rest, so a croupier launches the next ball while it is still moving, and a wheel sitting dead still between coups looks nothing like a table. Once a number is called this one comes to rest anyway, which a real wheel would not: a wheel still turning behind an announced result reads as though the coup is unfinished, and the result is the thing you came to the page to see.
Does it work on a phone?
Yes. Drag across the wheel to send it, or use the spin button. If your device or browser has no WebGL, or you have reduced-motion enabled, the page shows the pocket sequence as a table instead, which carries the same information without the animation. Touch and mouse use the same drag gesture, so nothing about the controls changes between a laptop trackpad and a phone screen.
Can I win anything here?
No, and that is deliberate rather than a limitation. This is a free wheel for learning the layout, the sequence and the pace. If you go on to play for real money, do it at a Gambling Commission licensed casino, with a deposit limit set before you start. Adding a payout here would need a licence this page does not hold and does not want; a free wheel exists so you can practise, and a payout would turn it into something else entirely.