What double-clicking by itself actually is
A mouse button is a mechanical switch, and mechanical switches bounce. When the contacts close they do not close once, cleanly. They make and break contact several times over a few milliseconds as the metal settles. Every mouse contains debounce logic, in firmware, that ignores further transitions for a short window after the first, which is what turns a messy physical event into one clean click.
A worn switch bounces for longer than that window. The contacts oxidise, the spring weakens, and the settling time grows past the debounce period until a second transition arrives after the firmware has stopped ignoring them. The mouse then reports a second click that your finger never made. This is why the failure appears gradually, why it worsens over months, and why it happens to the left button first: it takes the most actuations.
The measurable signature is the interval. A human double click has a gap between the two clicks in the range of roughly 80 to 500 milliseconds, bounded by how fast a finger can move. Switch chatter produces a second click within a few milliseconds of the first, usually under twenty, which no finger can do. That threshold is what separates a diagnosis from a guess, and it is what this tool measures: not whether a double click happened, but how long after the first click the second one arrived.
The same timing view answers a different question people often have, which is whether the mouse is dropping clicks rather than adding them. A dropped click is a press with no release, or a release with no press, and it shows up in the raw event stream as an unmatched pair rather than as a short interval.
Polling rate is a separate axis and worth not confusing with this. A 125 Hz mouse reports its position every 8 milliseconds and a 1000 Hz mouse every 1 millisecond, which affects how smooth the cursor feels and has nothing to do with whether the button chatters. A high polling rate on a worn switch is a fast report of a wrong click.