Mouse tester

How to test a mouse

1. Point into the test box and press each button once: left, right, the wheel as a button, and the back and forward side buttons. Each lights in the drawing while held and its count goes up by one.
2. Tap Start 20 s run and single-click each button at your normal pace, about 20 times each. Read the verdict per button when the run ends.
3. Scroll the wheel slowly one way for a few seconds, then the other. Each notch adds one to Wheel up or Wheel down; tilt it sideways if it tilts.
4. Tap Start polling test and move the mouse in fast circles for 3 seconds. Read the median and peak reports per second.
5. Check the event log for anything odd: a press you did not make, or a release missing.

How the double-click test works

Every press inside the box is stored with the browser's timestamp, to a fraction of a millisecond. During the run, each press is compared with the previous press of the same button; when the two start less than the threshold apart (80 ms unless you change it), the pair is flagged with its gap. Changing the threshold re-checks the presses already recorded.

Worked example: you click the left button 25 times. Presses 14 and 15 start 23 ms apart and presses 21 and 22 start 31 ms apart. Both are under 80 ms, so the table shows Left: 25 presses, flagged 23 ms, 31 ms, double-click fault: 2 double presses. One flagged pair in a run reads one suspicious double press, because a fast deliberate double click can now and then land near the threshold; a fault shows up again on a second run.

Fault-prone switches fail on the main buttons most, because those get pressed millions of times. If a switch double clicks only sometimes, press it lightly and slowly: worn switches chatter most on light presses.

How the wheel test works

Each wheel event with a vertical movement counts as one tick up or down, and one with sideways movement counts as a tilt left or right. Ticks less than 0.3 s apart belong to one scroll. Inside one scroll, a stretch of one or two ticks the other way with at least two ticks of the main direction before and after it is flagged as reversed: 10 ticks down, 1 up, 5 down is one reversed tick. Changing direction on purpose has no same-way ticks after it, so it is not flagged.

A notched wheel sends one event per notch in most browsers. A touchpad, a free-spinning wheel or a smooth-scrolling driver sends many small events, so its counts run higher than the notches you feel; the reversed check still works on those.

How the polling rate is measured

The mouse reports its movement to the computer a fixed number of times per second, its polling rate (125 Hz is the old USB default, 1000 Hz is common on gaming mice). The page counts reports that reach it. Where the browser offers the pointerrawupdate event it listens to that, which is sent as soon and as often as the browser can produce it; otherwise it unpacks the batched reports inside each pointermove with getCoalescedEvents(); failing both it uses plain pointermove timestamps. The source used is printed under the result.

The median is 1000 divided by the middle gap between reports, so stray slow moments do not drag it down. The peak is the fastest stretch of 16 reports in a row. Both raw events need a secure (HTTPS) page and are not offered in every browser, so the same mouse can read 1000 Hz in one browser and 60 to 125 in another.

What it cannot do

• It cannot measure DPI or CPI. That needs the mouse moved over a known distance, and a web page only sees screen pixels after the system's pointer speed and acceleration are applied.
• It cannot change mouse settings: polling rate, DPI, button mapping and lighting live in the system settings or the maker's own software.
• Side buttons beyond back and forward, and keys the maker's software turns into keyboard shortcuts, are invisible to a browser.
• On a phone or tablet it tests a connected mouse only. Touches are not counted as clicks.
• Some browsers do not let a page stop the back and forward buttons in every case; if one goes back, return to this page and test it again.
• A bounce shorter than the time the system takes to report a press can be merged by the mouse or the system into one press and will not show here.

Privacy

Clicks, scrolls and positions never leave your device and nothing is saved. The page sends an anonymous page-view count and one of two fixed words, mouse-start and mouse-dblclick-test; never a reading.

Questions

How do I do a double click test on my mouse?

Tap Start 20 s run, then single-click each button at a normal pace inside the test box, 20 or more times each. Every two presses of one button that start less than 80 ms apart are listed with their gap, and each button gets a verdict: no fault seen, one suspicious double press (run it again), or double-click fault. A worn switch usually shows gaps of 5 to 40 ms.

How do I test the right click without the menu opening?

Right-click inside the test box. The page stops the context menu there, lights the right button and adds one to its count. Outside the box right-click works as normal. The back and forward side buttons are caught the same way so the page does not go back while you test them.

Can I check my mouse polling rate online?

Yes, roughly. Tap Start polling test and move the mouse in quick circles for 3 seconds. The page counts how many position reports reach it per second and shows the median and the peak. A 1000 Hz mouse should read close to 1000 in Chrome or Edge; browsers may cap or batch events, so a lower figure in a browser does not prove the mouse is slow.

Is this mouse test online free, and does it need a download?

It is free, with no account and nothing to install. Everything runs in the browser tab; no click, position or timing leaves your device. It works with any mouse the computer sees, wired, wireless or Bluetooth.

Why does my scroll wheel jump the wrong way?

The wheel's encoder is usually worn or dusty, so one notch now and then is read as the opposite direction. Scroll steadily one way over the test box: a single tick against the direction inside a steady scroll is flagged as reversed. A few in a long scroll mean the encoder is going; cleaning it can help.