Walk into any casino — physical or digital — and you'll see it: a display board showing the last 10, 15, or 20 roulette results. The numbers glow in red and black. Players study them intently, looking for patterns, streaks, and clusters. Number 17 has hit three times in the last 20 spins? It must be hot. Number 4 hasn't appeared in 40 spins? It's due.
This is one of the most persistent beliefs in gambling. It's also one of the easiest to test. So we tested it.
What Are Hot and Cold Numbers?
A hot number is one that has appeared more frequently than expected in recent spins. A cold number is one that has appeared less frequently than expected — or not at all — over a given window.
On a European wheel with 37 pockets, each number has a 1-in-37 chance (2.70%) of landing on any given spin. Over 100 spins, you'd expect each number to appear approximately 2.7 times. But randomness doesn't distribute evenly in small samples. Some numbers will appear 5, 6, or 7 times. Others might not appear at all. This is variance — the natural, expected deviation from theoretical probability.
The Simulation: 100,000 Spins
We simulated 100,000 spins on a European roulette wheel using a cryptographically seeded random number generator. At every 100-spin checkpoint, we identified the three hottest numbers (most frequent) and three coldest numbers (least frequent) from the preceding 100 spins. Then we tracked how those numbers performed over the next 100 spins.
Results — 1,000 Checkpoint Windows
Hot number average frequency (next 100 spins): 2.71 appearances
Cold number average frequency (next 100 spins): 2.69 appearances
Expected frequency (any number): 2.70 appearances
Hot number advantage vs expected: +0.37%
Cold number deficit vs expected: −0.37%
The results are unambiguous. Hot numbers in the past performed almost exactly at the expected rate in the future. Cold numbers did the same. The deviation between them — 0.37% — is within the standard error margin and statistically insignificant. Past frequency had no predictive power.
Why Your Brain Sees Patterns That Aren't There
Humans are pattern-recognition machines. It's one of our most powerful cognitive abilities — and one of our most dangerous biases when applied to random sequences. Three specific biases drive the hot/cold number fallacy:
The Clustering Illusion
Random sequences naturally produce clusters — runs of the same outcome, or repeated appearances of the same number in a short window. Your brain interprets these clusters as meaningful patterns because that's what it evolved to do. In nature, clusters often are meaningful (predator tracks, fruit-bearing trees). In roulette, they're noise.
The Gambler's Fallacy
The belief that a number is "due" because it hasn't appeared recently is the textbook gambler's fallacy. Each spin is independent. The wheel has no memory. The probability of any number landing is 1/37 on every spin, regardless of what happened on the previous 10, 100, or 1,000 spins.
Confirmation Bias
When you bet on a hot number and it hits, you remember. When it doesn't, you forget — or rationalize ("the streak ended"). This selective memory creates the subjective experience that hot numbers "work" even when the data shows they don't.
The Window Problem
Even if you believe that frequency data has predictive value, there's a practical problem: how many spins should you look back? The casino display shows 20 spins. But why 20? Why not 50? Or 200? Or 5?
We tested multiple lookback windows against our 100,000-spin dataset:
| Lookback Window | Hot Number Hit Rate (Next 37 Spins) | Cold Number Hit Rate (Next 37 Spins) | Expected |
|---|---|---|---|
| 20 spins | 2.74% | 2.66% | 2.70% |
| 50 spins | 2.71% | 2.68% | 2.70% |
| 100 spins | 2.71% | 2.69% | 2.70% |
| 200 spins | 2.70% | 2.70% | 2.70% |
No lookback window produced a statistically significant edge for either hot or cold numbers. The shorter the window, the noisier the signal — but even the noisiest windows converge to the expected 2.70% hit rate over enough trials. There is no magic window.
Why Do Casinos Show the Display Board?
If hot and cold numbers don't predict outcomes, why do casinos prominently display recent results? Because the display board encourages betting. Players who believe in hot numbers bet more — they place additional straight-up bets on "trending" numbers. Players who believe in cold numbers also bet more — they add bets on "due" numbers. The casino wins either way because both strategies increase total action, and every additional bet carries the same 2.70% house edge.
The display board is not an analytical tool. It is a marketing tool. The casino is not giving you an advantage by showing it. They are giving you a reason to bet more.
When Frequency Analysis Does Matter
There is one scenario where number frequency data has legitimate predictive value: biased wheels. A physical roulette wheel is a mechanical device subject to manufacturing tolerances, wear patterns, and environmental factors. If a wheel is imperfectly balanced, certain numbers may genuinely land more often than others over very large sample sizes (thousands of spins).
Wheel bias is real and has been exploited historically. But detecting a genuine bias requires thousands of recorded spins, statistical testing for significance, and access to a specific physical wheel over an extended period. It has no relationship to the "hot number hit three times in 20 spins" observation that drives most players' decisions. Modern casinos also regularly inspect and rotate wheels specifically to prevent bias exploitation.
Track Number Frequency in Real Time
Our simulator records every spin and calculates frequency distributions, z-scores, and deviation analysis automatically. Run 1,000 spins and see exactly how random variance produces "hot" and "cold" streaks.
Open the SimulatorThe Bottom Line
Hot numbers are not hot. Cold numbers are not cold. They are random outcomes that your brain retroactively assigns meaning to. The data from 100,000 spins is clear: past number frequency has no predictive power for future outcomes on a fair wheel.
This doesn't mean you can't use frequency data — it means you need to use it correctly. Understanding how variance creates the appearance of patterns in random data is the foundation of informed play. The goal is to see the game as it is, not as your brain wants it to be.
The scoreboard shows you the past. It tells you nothing about the future. The only number that matters on every spin is the same one it's always been: 2.70%.