Choose the appropriate protocol
A range test involves deliberate motion, so its center statistics should not be read as drift. Select the range protocol, start the session and move one stick slowly through its intended travel before testing the other. Return each control to rest between movements so the trace remains easy to interpret.
For buttons, choose the button protocol and press each intended control slowly enough for browser snapshots to observe it. The report stores normalized button values and the browser-reported pressed state as a bit mask. A counted transition means the sampled state changed from unpressed to pressed; a very brief press between two samples can be missed.
Use numbered axes and mapping information
Controllers do not all expose the same layout. A standard mapping helps the browser present common controls consistently, while an unknown mapping needs interpretation from the numbered axis and button list. The lab does not infer a manufacturer model from a device name or pretend that every pair of axes is physically a conventional stick.
The plotted axis pairs show the first two pairs, with numeric statistics available for the supported remaining axes. Range coverage counts twelve angular sectors reached beyond radius one half. This gives a straightforward observation of sampled movement. It is not a precise physical-circle calibration or a circularity pass/fail certificate.
Keep the original values and limits visible
The lab leaves raw values unchanged. A value reaching the normalized limit means that the browser reported that limit; it does not measure the angle of the physical stick. Firmware and driver processing may already have shaped the values before the browser sees them. The chosen review radius is only an annotation.
Stop when the intended protocol is complete and download JSON or CSV. The record is bounded by time and snapshot count, and it stops if the page is hidden or the device disconnects. Do not combine unrelated rest, range and button routines into one comparison and then interpret their aggregate statistics as the same experiment. Separate sessions make the evidence easier to explain and reproduce.