Mouse Drift and Tracking Test

Two questions, one page: does the cursor move while the mouse sits still, and does the sensor keep up while you move it? Nothing is installed and nothing leaves your device.

Sensor drift

Let go of the mouse, then start — 3 seconds without touching it.

Furthest drift
—
Total wander
—
Samples
—

A healthy mouse sends no movement whatsoever while it sits still. If this run came back clean but the cursor still wanders in daily use, try the longer run — some sensors only twitch every so often.

Trace pad and drag test

Move the cursor here to draw. Hold the left button and drag for the heavier line.

Samples
0
Largest jump
0px
Dropped drags
0

Draw slowly, then quickly. Corners and straight cuts across a curve are the sensor skipping. Hold the left button while you draw: if the heavy line thins out before you let go, the button released on its own.

Why a Cursor Drifts and a Line Breaks Up

Both tests above watch the same thing from opposite ends: what the sensor reports when it should report nothing, and what it reports when it should be reporting everything.

Drift: movement out of nowhere

An optical sensor is a tiny camera photographing your desk thousands of times a second and working out how far the picture shifted between shots. When it can find nothing to compare — or compares the wrong things — it guesses, and the guess arrives as cursor movement you never made. On a mouse standing still, that is drift.

Jitter: a line that comes apart

Move quickly and the sensor has less overlap between one photo and the next. Past its limit it loses the thread, so the line jumps instead of curving. The largest jump on the pad puts a number on it — though a big gap can also be a slow report rate rather than a lost sensor.

Dropped drags come from the switch, not the sensor

A drag that lets go halfway across the screen is almost never a tracking problem. It is the left switch bouncing open for a millisecond under a worn spring, which the computer reads as a release. The pad counts every release that happened while the cursor was still travelling.

Clean it before you replace it

Turn the mouse over and look into the lens under a lamp. A single hair or a smear of grease is enough. Wipe the lens with a dry cotton bud, clear the ring of grime around it, and give the mouse a plain matte mat rather than bare desk, glass or a busy pattern.

When the mouse really is finished

A clean lens on a plain mat, a fresh battery or a different USB port, and the drift is still there — that is a sensor at the end of its life. Dropped drags on a mouse a few years old are the same story for the switch. Both are cheaper to replace than to live with.

Mouse DPI test
A cursor that feels twitchy may simply be set too sensitive. Measure what your mouse actually reports per inch.
Measure DPI
Polling rate test
Gaps in the trace are often a slow report rate rather than a failing sensor. See how often your mouse speaks up.
Check the report rate
Gamepad stick drift
The same fault on a different device: a controller stick that keeps moving after your thumb lets go.
Test stick drift

Frequently asked questions

Something is making the sensor believe the mouse moved. Dust or a hair on the lens, a glossy or glass desk, and mouse pads with a fine repeating pattern account for most cases; a worn sensor accounts for the rest. Run the test above with your hand off the mouse — if the cursor drifts more than a pixel or two over a few still seconds, the sensor is reporting movement that never happened.
Work through it in order. Turn the mouse over and clean the lens with a dry cotton bud. Move to a plain matte mouse pad — not glass, not a polished desk, not a busy pattern. On a wireless mouse, change the battery and move the receiver to a port near the mouse, away from other transmitters. Then run the test again. If it still drifts after all that, the sensor is at fault.
There are two harmless reasons before a broken one. The sensor may be past the speed it can track, which is what its maximum-speed rating means, and that shows up only in fast strokes. Or the mouse simply does not report often enough — at 125 Hz the browser hears from it every 8 ms, so a quick flick leaves visible gaps. Draw the same shape slowly: if the line is smooth then, the gaps are speed, not a fault.
It is when you are holding the left button to drag something and it lets go on its own halfway. The switch inside the button bounces open for a fraction of a millisecond, the computer sees a release, and the file lands wherever the cursor happened to be. The trace pad counts them: hold the button, draw, and any release recorded while you were still moving is shown as a dropped drag.
More than most people expect. Optical sensors need texture to lock onto, so glass, high-gloss desks and clear plastic are the worst surfaces there are. Mirror-finish and very finely patterned pads confuse them in a different way: the sensor matches the pattern to the wrong repeat and invents movement. A plain matte pad in a mid grey fixes both problems for the price of a coffee.
Plug the mouse into another computer, or another mouse into this one — that answers it in a minute. Short of that: drift that follows the mouse from surface to surface is the mouse, while a cursor that jumps only when the machine is busy is the computer struggling to keep up. Pointer-precision settings, mouse utilities and remote-desktop software can also move the cursor themselves, so turn them off before condemning the hardware.
It can behave worse, though usually not by drifting. A weak battery or a receiver crowded in among other USB devices makes the mouse skip and stutter, which shows up on the trace pad as gaps rather than as movement while it sits still. Change the battery, plug the receiver into a port near the mouse, and keep it clear of Wi-Fi routers and USB 3 drives.
Yes to both. Everything runs inside your browser, nothing is uploaded and there is nothing to install. Leave the page and the measurement is gone with it.