Quick answers before you test
Mouse acceleration test
Record slow and fast swipes over the same desk distance. If the fast swipes report more counts, something in the path accelerates the pointer: the OS curve, a driver profile or firmware.
Angle snapping and prediction
Draw one long stroke. Human hands always wobble a little; if nearly every event has zero sideways movement and the line is unnaturally straight, snapping or prediction is smoothing your aim.
Sensor skipping at speed
A fast swipe should be one continuous run of counts in one direction. Backwards jumps or isolated spikes mean the sensor lost tracking, usually on a glossy surface or with a dirty lens.
Measure observed DPI as well →What the result actually measures
The lab locks the pointer to a capture field and records raw movement counts per event. The acceleration stage compares median travel of slow versus fast swipes over the same distance. The stroke stage measures the share of events without minor-axis movement and the residual against a fitted line. The rest stage counts movement while the mouse is untouched. The skip stage detects sign reversals and spikes during a fast swipe. The surface stage stores two sets of results and compares them.
Measurement boundary: Counts arrive through the operating system and browser. OS pointer acceleration, driver smoothing and browser scheduling are part of what is measured, so a finding describes the path on this machine. Confirm firmware settings in the vendor software and repeat on a known-good surface before judging the sensor.
How to interpret the output
Speed ratio
Fast counts divided by slow counts over the same distance. 0.90–1.10 is 1:1 tracking within hand tolerance; above 1.10 indicates acceleration, below 0.90 negative acceleration or skipping.
Zero minor-axis share
The percentage of movement events with no sideways component. Natural strokes stay well below 85%; a near-100% share with a tiny residual points to angle snapping.
Skip score
Reversals and spikes per 100 events during the fast swipe. Clean sensors stay near zero; a score above 1.5 or more than two reversals deserves a surface change and a lens check.
Troubleshooting unexpected results
Pointer lock is refused
Click inside the capture field first, keep the page in the top-level tab and allow the browser prompt. Some kiosk or embedded browsers block pointer lock entirely.
Ratio changes every run
Swipe distance must be identical for slow and fast passes. Use physical marks, keep the same grip and record three of each so the median absorbs one bad swipe.
Everything reads suspicious on a laptop
Touchpads and OS acceleration curves produce exactly these patterns. Plug in the mouse you want to test and disable "Enhance pointer precision" or the macOS acceleration curve for the run.
Frequently asked questions
Can a browser detect mouse acceleration?
Yes, indirectly: pointer lock exposes raw movement counts, so identical physical distances at different speeds should produce identical counts. A consistent difference reveals acceleration somewhere in the path.
Is angle snapping bad?
It helps drawing straight lines and hurts fine aim corrections. Competitive players usually disable it; this lab shows whether it is active so the choice is yours.
Does this replace MouseTester?
MouseTester on Windows plots raw counts against time with more detail. This lab brings the core acceleration, snapping, rest and skip checks to any browser with pointer lock and adds surface comparison and a device passport.
Can a chatbot tell me whether my sensor skips?
No. Skipping only shows up in counts recorded from your mouse on your surface, which is what the capture field does.
