A live roulette session recorder. You enter each winning number as it comes up, and it keeps the running record and the charts for you — no download, no sign-in, nothing stored on a server.
Roulette players have been writing spins on scorecards for as long as the game has existed. This is that scorecard, with the arithmetic done automatically. Enter a number, and it updates:
Every figure is derived from the spins you entered — nothing is estimated, simulated or filled in. The FAQ shows the exact formula behind each chart.
There's no betting system built in, no "next number" suggestion, and no strategy advice. The house edge on an American wheel is 5.26% on essentially every bet, and no pattern of play alters it. This is a record-keeping and curiosity tool: it's genuinely interesting to watch a run of spins take shape, and it's a good way to see for yourself how ordinary streaks and hot numbers really are.
Spins you enter live in the page while it's open. There's no account, no database, and no upload. Reload the page or close the tab and the session is gone — which also means there's nothing to delete and nothing that could be breached.
A small counter in your browser's local storage tracks how many times you've opened the page, shown in the corner of the tracker. It's labelled as a this-browser-only count precisely because it isn't a real visitor total. Clearing your browser data resets it.
The charts are drawn with Chart.js, loaded from a public CDN, so your browser does request that file. Share buttons only open the relevant site when you click one — they don't load trackers in the background.
Percentages assume 38 pockets: 1–36 plus 0 and 00. If you're playing a European single-zero wheel, the structure of the charts still works, but the reference figures (2.63% per number, 47.37% per even-money side, 5.26% edge) will read slightly pessimistic — a European wheel is 2.70%.
Most of what looks striking in the first hundred spins is noise. Hot numbers, lopsided colour counts and long streaks are all expected outcomes of a fair wheel over short runs. Treat early readings as scenery, not signal.
The tracker can't know if you mistyped a spin. If a session looks strange, check the results table at the bottom of the page first — you can remove any row.