Every chart here is a record of spins you have already entered. None of them forecast anything — this page explains exactly what each one counts, and what it does and doesn't tell you.
Ground rules
An American wheel: 38 pockets — 1 through 36, plus 0 and 00. That shapes every percentage on the site.
| Bet | Pockets that win | Chance per spin |
|---|---|---|
| A single number | 1 of 38 | 2.63% |
| Red, Black, Odd, Even, High, Low | 18 of 38 | 47.37% |
| A dozen or a column | 12 of 38 | 31.58% |
| 0 or 00 specifically | 1 of 38 each | 2.63% each |
Notice that red is not 50%. It's 47.37%, because 0 and 00 belong to neither side. That 2.63% gap is the house edge — on this wheel it works out to 5.26% of every even-money bet — and it's the reason an even-money side is slightly behind over a long session rather than dead level.
Chart 1
One dot per spin, at the height of the winning number, coloured red or black. The gold line running through them is the running average of every number so far:
Green spins (0 and 00) are left out of this chart entirely. Including them would drag the average toward zero and make it look like the wheel was running low, when really nothing about 1–36 had changed.
That's the arithmetic middle of 1 to 36 — (1 + 36) ÷ 2 = 18.5,
rounded to the label at 18. It is a reference mark, not a prediction. Over a long
session the running average drifts toward it simply because averages of many
equally-likely numbers converge on their midpoint.
The red band above 21 and the blue band below 15 are eyeball guides. If the gold line is sitting in the red band, this session's numbers have averaged high so far. That's a description of the spins you entered — it says nothing about where the next one lands.
Chart 2
A straight tally: how many times each pocket has hit this session. Hot lists the five most frequent, cold the five least frequent among numbers that have actually appeared. Numbers that haven't come up at all are excluded, because otherwise every unseen number would tie for coldest and the list would be meaningless.
With 38 pockets, you need a lot of spins before "hot" means anything but noise. After 38 spins the average number has hit once — so a number showing three times looks dramatic but is completely ordinary. Expect big spread early and less later.
Charts 3, 4 and 5
Odd/Even, High/Low and Red/Black all work identically — only the rule for which side a number belongs to changes.
A running total. Each spin adds or subtracts one:
So a line at +6 means the top side has landed six more times than the
bottom side this session. Above zero is the top side's territory (shaded), below
zero is the bottom side's. Green spins leave the line flat for one step, which is
correct: 0 and 00 are neither odd nor even, neither high nor low, neither red nor
black.
Each moving average is the plain average of the unit-count line over its last N spins:
A line stays blank until enough spins exist to fill its window, so MA 50 only appears from spin 50 onward. Because the unit count moves by exactly one step per spin, these averages behave like a lagged, smoothed copy of the line rather than a noise filter.
When MA 10 sits above MA 50, the recent stretch has leaned toward the top side more than the older part of the session did. That is a fact about the spins behind you. It is not a signal, and a crossover does not mean a change is starting or ending.
The same data as a share instead of a total:
Green spins are excluded from both the top and bottom of that fraction, which is why it's labelled "% of applicable spins." The 50% line is a dead-even split, not the expected value — remember the true long-run figure for each side is 47.37%, with the rest going to 0 and 00. Moving averages are hidden in this view, since averaging a running percentage tells you very little.
The panel in the top-left corner of each chart shows every line's current value. Hover the chart and it switches to the values at the spin you're pointing at. Click any entry to hide or show that line — useful for isolating one average.
Session totals for that chart. The side that's ahead shows in bright green, the side behind in bright red. The labels never swap places — only the numbers and their colours change — so Odd stays on the left whichever way the session is running.
All Spins fits the whole session into the visible width, so the chart never scrolls. Last 50 shows only the most recent 50 spins, and the unit count restarts from zero at the left edge of that window — it is a fresh count of those 50 spins, not a zoom into the session line. Moving averages follow the same rule, so MA 50 needs the full window before it appears.
Chart 6
Three bars: how many times 0 has hit, how many times 00 has hit, and how many of each you'd expect on average.
Actual and expected almost never match, and the gap is usually large in percentage terms early on. At 100 spins you'd expect 2.6 of each, so seeing zero of one and five of the other is entirely normal variance. The bars converge on the expected value slowly — think thousands of spins, not hundreds.
Charts 7 and 8
Two or more consecutive spins landing in the same column (or dozen). A single isolated hit is not a streak and isn't counted. A 0 or 00 breaks any streak in progress, because it belongs to no column and no dozen — exactly as it would settle a real column or dozen bet as a loss.
The bar chart in each card shows how many times a streak of each length occurred. The summary row shows total occurrences, share of all spins, and the longest run so far.
Streaks look surprising and are not. Each column hits 12/38 = 31.58% of spins, so back-to-back is about 10% and three in a row about 3% — meaning in a 200-spin session, several two-streaks and a few three-streaks are the expected outcome rather than an anomaly. The distribution should taper steeply: many short streaks, few long ones.
Common questions
No. Every spin of a fair wheel is independent of every spin before it. There is no pattern in the charts that changes what happens next, because the wheel has no way of knowing what it did last time. The charts are a session record — useful for seeing what happened, interesting to watch, but not predictive in any way.
No — this is the gambler's fallacy, and it's the single most expensive idea in casino gambling. The wheel doesn't balance itself out. Long-run percentages settle down because later results dilute early imbalances, not because the wheel corrects them. After 12 blacks, red is still 47.37% on the next spin.
On an American (0 and 00) wheel it's 5.26% on essentially every bet. In plain terms: over time the game is expected to keep about $5.26 of every $100 wagered. No betting pattern, progression or chart-reading changes that number — it comes from the payout structure, which pays as if 0 and 00 didn't exist.
A European wheel with a single zero has a 2.70% edge. This tracker's percentages assume the American 38-pocket layout.
No. Auto Spin generates random numbers in your browser to fill the session with sample data — it's for exploring the charts without typing in a hundred spins by hand. Real sessions come from entering each winning number yourself.
No. Everything lives in the page while it's open. Reloading or closing the tab clears it, and nothing is uploaded. See the About page for detail.
Because that view counts only those 50 spins. It starts from zero at the left edge and tallies forward, so it answers "how have the last 50 spins gone" rather than "where is the session total right now." Switch to All Spins for the session-long figure.
Because 0 and 00 aren't odd or even, high or low, red or black. Scoring them as either side would misrepresent what happened. They still occupy a spin on the x-axis, so you'll see the line hold flat for one step, and they still appear in the session table and hot/cold counts.