Start with untouched sticks
Move and release both sticks naturally, then place the controller on a stable surface. Take a five-second resting sample without touching the sticks. Repeat it. A value captured while moving the controller is not a resting baseline. Save the connection and any remapping layer with the report.
Read the formula
DeskProof uses sqrt(x²+y²) for each pair of browser axes. A 0.03 horizontal and 0.04 vertical offset has magnitude 0.05, displayed as 5%. A diagonal at (1,1) is about 141.4% under this definition. That percentage is a distance from browser centre, not wear, damage or a universal failure probability.
Keep the input layers separate
A game may apply its own deadzone after the browser would have received a value. A driver or remapping layer may already apply a deadzone before the browser. Thus a non-zero browser reading need not move a game character, and a zero reading cannot rule out a sensor problem. Record the actual in-game symptom separately.
Compare before calibration
Use the same surface, sample duration and connection for a baseline pair. Compare the maximum and mean of each stick without turning the difference into a health score. If calibration is appropriate, follow the specific manufacturer procedure and keep before/after reports. This website never writes calibration to the controller.
Keep both reports in Device Lab →