CTRLCHK is a browser-based diagnostic workspace for people who are tired of testers that light up buttons and call it science. It keeps observed measurements separate from product guidance, so you can always inspect the underlying signal.
What we measure
Resting stick drift (offset, jitter, P95, RMS, peak), a recommended software deadzone, 8-sector range coverage, rest-corrected circularity error, snapback overshoot, required-button coverage, analog trigger peaks, browser-observed update rate, optional dual-rumble, and button chatter or stuck holds. Focused tools cover the deadzone test, button test, trigger test, and a calibration check that cannot write firmware. Measurement version 3.0 is printed on every report. If we change a threshold, the version bump is how you know two scores are not comparable.
What we refuse to measure
USB 1000 Hz polling, DualSense adaptive triggers, Elite paddles, gyro, and true haptic spectra. The Gamepad API does not provide them. Competitors sometimes invent a number anyway. We would rather leave a blank and say why.
Scoring philosophy
Health classifications are versioned guidance, not universal hardware specifications. Confidence reflects mapping support, sample sufficiency, and stage completion. Capability-limited stages — rumble on a browser without vibrationActuator, analog triggers on an empty mapping — never reduce the score. An invalid hands-off centre withholds the overall score entirely, because a drift number from a moving stick is worse than no number.
No controller input needs to leave your device. The Gamepad API is polled locally and reports are generated in your browser. See privacy for the short version of that promise.