Competitive PUBG Mobile analytics
A live zone simulator, the probability chart behind circle reads, the arithmetic of a 1v4, and a playbook for the 2026 tournament pool — built on a published candidate algorithm and honest about what hasn't been measured.
A disc of radius 2r′ − r around the current centre cannot be caught outside the next zone. It is geometry, not a forecast, and it needs no match data — only the radius schedule.
See it on the chartCircle-fitting error, carried through the key statistic, grows from ±0.012 at phase 1 to ±0.79 at phase 8, on a scale of 0 to 1. Phases 1–4 hold the usable signal.
Read the precision analysisOf eighteen accuracy-and-armour cases, four fit in an extended M416 magazine — none of them against Level 3 armour, which is what the survivors wear when clutches happen.
Run the numbersWatch the zone close, drag a team into its path, and see whether they get out. Every match feeds a live test battery.
Where the next centre can land versus where a team can stand and stay inside. Different questions; positioning needs the second.
The ammunition a 1v4 costs, the armour grid, and double-gun loadouts ranked by margin.
Anti-rush defence, the 1v4 clutch method, and Erangel, Miramar and Rondo map by map.
The method, the patent evidence, the power analysis, and what is still unmeasured.
Every tool on this site as a desktop program that works offline. Windows 10 and 11, 64-bit.
The same site packaged as an app that works offline. Android 7 or newer; installed from the APK file, not Google Play.
The website, the Python research toolkit, the workbook and both apps, with every release.
Neither app carries a paid code-signing certificate. Windows SmartScreen may say it doesn't recognise the installer: choose More info → Run anyway. Android asks you to allow installs from your browser the first time.
The zone models come from a published Tencent patent describing a centre-selection routine. A patent describes an embodiment, not shipped code, and no tournament matches have been annotated yet.
PUBG Mobile publishes no weapon statistics. Every figure here is a private-lobby estimate, the sources disagree, and each row says how much to trust it.
Everything computes from inputs you can replace with your own timings and tests. Quote the method, not the numbers — then measure and re-run.
Download the combat workbook