Roulette tracker and bias wheel checker: records, not predictions

Is a roulette tracker legit?

A tracker that records results is perfectly legitimate: casinos show one above every live wheel. A tracker that claims its hot and cold lists tell you where to bet next is not, because a fair wheel has no memory and every number has the same chance on every spin. Believing otherwise is the gambler's fallacy, and the tracker above makes no such claim. It gives you descriptive statistics on numbers that have already landed, nothing more.

Do hot numbers in roulette keep hitting? We tested it

No. We simulated 100,000 sessions of 100 spins each on a fair European wheel, the sort of stretch a tracker app or a casino's stats panel shows. In 91.9% of sessions at least one number came up six times or more, more than double the 2.7 you'd expect, and in 53.9% a number hit seven or more. On average 2.4 numbers never appeared at all, so almost every session hands you a convincing "hot" number and a few "cold" ones.

Then we followed each session's hottest number for the next 100 spins. It landed 2.69 times on average, against 2.70 predicted by a fair wheel. The streak predicted nothing. The code and data are on our research page.

How do you test a roulette wheel for bias?

Compare how often one number hit with how often it should hit on a fair wheel, and measure the gap in standard deviations. That figure is a z-score: the bigger it is, the less likely the deviation came from chance alone. The checker above calculates it from your spins and hits, which is the same statistic wheel-clocking has always relied on.

How many spins do you need to detect bias?

Tens of thousands of spins on the same physical wheel. The checker flags anything under 10,000 spins as too small to say much, because a perfectly fair wheel routinely throws up numbers 20% or 30% over expectation across a few hundred spins. Joseph Jagger's clerks tracked thousands of spins at Monte Carlo in 1873 before he bet heavily, and even then he was exploiting a real manufacturing fault, not a hunch backed by a handful of results.

Testing one number is a trap

Checking only the number that already looks hot makes ordinary noise look like a discovery. A wheel has 37 or 38 numbers, so testing each at a 95% confidence threshold means roughly one in twenty will look "significant" on a completely fair wheel every time you run it. Proper bias analysis checks all the numbers together.

Our own data shows it. Across a million spins of a simulated wheel we know to be fair, number 6 came up 27,420 times, 2.42 standard deviations above expectation. Tested alone, you'd call it biased with 98% confidence. Tested with the other 36 numbers in a chi-square, the wheel scores 28.3 on 36 degrees of freedom: textbook fair.

Why live studios make bias hunting harder

RNG roulette has no physical wheel to bias at all; results come from a certified random number generator. Live-dealer roulette does spin a real wheel, but studios use precision-engineered wheels, maintain and rotate them, switch which wheel feeds a stream, and run their own statistical monitoring, often with the live dealer AI tracking covered in the glossary. They're set up to catch wear or a tilt long before an outside player could collect enough spins to notice it.

Red flags in paid tracker apps

Free stats panels are harmless. Be wary when a tracker app or channel charges for "signals", claims a win rate, shows only screenshots of wins, or links to a specific casino. A claimed hit rate of 48% to 49% on even-money bets is exactly what random guessing produces on a single-zero wheel. Anything that asks you to install software from outside the official app stores to "connect" to a casino's stream is a security risk in its own right. Game shows attract the same pitch, and the Crazy Time statistics page shows why "it's due" fails there too.

Using the tracker and checker together

Treat the tracker as a record and the checker as a reality check.

• Log spins to keep track of your session, not to guess the next number.
• Only run the bias test once you have thousands of spins from one physical wheel, and remember the one-number trap.
• Check the real cost of each bet with the roulette odds and payouts calculator, which applies whatever has landed recently.
• Compare your numbers with how far each pocket strays over a million fair spins in our roulette simulation.