Input and sampling
The tester reads the browser’s Gamepad API during animation frames. It retains bounded observation buffers and limits interface updates. Browser scheduling, connection, mapping and input translation can affect what is available. This is not a raw USB report capture.
Resting-center check
During an untouched-stick capture, each sample’s radial distance is calculated from its X and Y values. The result averages those distances and also retains the largest observed distance. Averaging distances prevents opposite offsets from cancelling each other. It does not identify why an offset exists.
Range and circularity
The current range calculation divides the stick sweep into 32 angular sectors and retains the greatest observed radius in each. Coverage counts sectors that reached a radius of at least 0.7. Circularity is withheld until all sectors reach that coverage threshold.
When complete, the calculation averages the absolute difference between each sector’s recorded maximum radius and a unit-circle reference. This is this tool’s method, not an industry certification or a universal wear threshold. Different tools and firmware may produce different results.
Buttons, triggers and response
The button capture records which buttons were observed pressed across the session. Trigger capture retains maximum reported values for standard-mapped trigger positions. An unpressed or unsupported control is not automatically a failure.
The response view uses intervals between observed input-state changes, with bounded buffers and an idle timeout. It is not the controller’s true USB polling rate. The reaction exercise includes the participant’s response, browser and display timing; it cannot isolate hardware latency.
Demo, export and hardware status
Demo inputs are generated in software and reports are explicitly labelled as simulated. JSON exports describe observations; save a real-device report only after a real device is selected.
Physical hardware, vibration and calibration outcomes still require model-specific testing. The calibration area adapts DualShock Tools under MIT and retains a visible verification warning. Do not interpret the presence of an operation as proof that it has been tested on your hardware.
Reproducible comparisons
Record model, firmware if known, OS, browser version, connection and any mapping utilities. Keep the same actions and conditions for repeated checks. Share the method and limitations alongside results, and never compare synthetic observations with hardware measurements as though they were equivalent.