The idea sounds reasonable. A dealer spins the wheel and launches the ball the same way hundreds of times an hour. So the ball should land roughly the same distance from the previous number each time — the dealer's signature. Log the last few dozen spins, find the distance that keeps coming back, and bet on that section of the wheel. A whole category of apps and "dealer signature trackers" is built on this.
We tested it the simplest way possible: we gave a tracker spins with no signature at all, and watched what it found.
How a tracker works
For every spin, the tracker measures how many pockets the ball travelled on the wheel from the last winning number — say, 12 pockets clockwise. After 30 spins it looks for the 9-pocket stretch of distances that came up most often. That is the "signature". On the next spin it bets on the 9 numbers that sit at that distance from the last number, $1 on each.
Nine numbers out of 37 should be hit 24.3% of the time by chance. If the dealer really has a signature, the tracker's section should be hit more often than that.
Test 1: Random spins still show a "signature"
We generated one million spins from a perfectly fair European wheel — every pocket exactly 1 in 37, every spin independent. No dealer, no habits, nothing to find. Then we let the tracker look at every 30-spin window:
| In the last 30 spins… | Random wheel |
|---|---|
| Share you would expect in any 9-pocket section | 24.3% |
| Share in the tracker's best section (average) | 39.0% |
| Windows where the best section had 40% or more | 50.2% |
| Best section in the top 10% of windows | 46.7% or more |
On a wheel with no signature at all, the tracker's favourite section held almost 40% of the spins — more than one and a half times what chance "should" give. Half the time it was 40% or higher. On a screen with a heat map, that looks exactly like a dealer who keeps hitting the same part of the wheel.
Test 2: Betting on it
Then we let the tracker bet on its signature section before every single one of the million spins:
| Result | |
|---|---|
| Spins bet | 999,972 |
| Signature section hit | 24.35% |
| Expected by pure chance | 24.32% |
| Return per $1 bet | −2.6% |
The section that held 39% of the last 30 spins was hit 24.3% of the time on the next one — exactly the chance of any 9 numbers. The tracker lost the same 2.7% house edge as someone picking 9 numbers at random. The difference in the last decimal is simulation noise.
But what if a real dealer has a signature?
Our test shows what a tracker sees when there is nothing there. Could a real dealer still have one? Experienced casino game-protection experts have looked hard at this and call it a myth: the ball bounces off the deflectors (the diamonds) and the frets between pockets in ways no dealer can control, and wheel speed changes from spin to spin. Casinos also rotate dealers and vary the wheel. Online, many roulette games have no dealer at all — the result comes from a certified random number generator.
And even if a small signature existed, you would have to prove it. To break even, the tracker's 9 numbers must be hit 25% of the time instead of 24.3% (nine $1 straight-ups pay $36 when one hits). Telling 25% apart from 24.3% with any confidence takes around 16,000 spins with the same dealer, the same wheel and the same habits. A tracker that decides after 30 spins is reading noise.
Why it feels like it works
Trackers are convincing because they are right about the past. The heat map really did light up that section — over the last 30 spins. And because the tracker's section covers 9 numbers, it hits about one spin in four, so a few hits in a row happen all the time and feel like confirmation. The misses are quickly forgotten when the next "hot" section appears. It is the same effect we found when we tested hot numbers and clusters: randomness is lumpy, but the lumps don't tell you where the ball goes next.
Our verdict
A dealer signature tracker will show you a signature on any wheel, including a perfectly random one — on average its favourite section holds 39% of recent spins when chance alone gives 24%. Betting on that section wins exactly as often as betting on random numbers, and loses the same house edge. Use a tracker for fun if you like, but don't pay for one or raise your bets because of one. If you want to know what a betting system really does, test it on 10,000 players in the System Builder or compare the 100 Club — no heat map needed.
Method: 1,000,000 simulated spins on a fair European wheel (wheel order 0-32-15-19-…-26). Tracker window 30 spins, 9-pocket sector (the best distance ±4 pockets), $1 on each of the 9 numbers per spin. Further reading: Steve Forte on the myth of dealer signature (Las Vegas Advisor).