What this mouse test checks
This mouse test answers four practical questions about the mouse on your desk. Does every button send a signal? Does any button send two signals when you pressed once? Does the scroll wheel report the right direction? And roughly how often does the mouse report its position to the computer?
A drawing of a five-button mouse sits in the test area. Each part turns orange while you hold it and green once the browser has registered it. Below it you get press counts with a status for each button, a double-click and release detector with a timing table, a log of every press, release and wheel step, and, in Chrome and Edge, a polling rate estimate.
How to use the mouse test
- Move the pointer into the test area. The corner label changes to "Listening". Only events over the box are counted.
- Press every button. Left, right, middle (click the wheel), and the back and forward side buttons if you have them. Each tile shows its press count.
- Scroll the wheel both ways. The arrows beside the wheel light up, and the scroll tiles count steps and show the newest direction and delta.
- Check for double clicks. Click each button slowly, about once a second, until its tile reads OK (20 clean presses). Watch "Flagged presses" and the red lines in the event log.
- Check holds. Hold each main button for about 3 seconds, then hold it while you move the pointer around inside the box. Presses outside the box are not counted. Watch "Quick re-presses".
- Adjust the limit if needed. The slider sets how close two presses of the same button can be before they are flagged: 80 ms by default, 20 to 300 ms in range.
- Estimate the polling rate in the box at the bottom, if your browser shows it.
- Save the result. "Copy report" and "Download report" give a plain-text summary, and the download adds the event log. "Reset test" clears everything.
The toggle beside the buttons stops the page scrolling while the pointer is over the box.
How each reading is measured
The browser identifies each button with a number. These are the values the tool reads.
| Button | event.button | Bit in event.buttons |
|---|---|---|
| Left (main) | 0 | 1 |
| Middle (wheel click) | 1 | 4 |
| Right (secondary) | 2 | 2 |
| Back | 3 | 8 |
| Forward | 4 | 16 |
Double-click detection measures from one mouse-down to the next mouse-down of the same button. Different buttons never count against each other, so pressing left and right together is not a fault.
Worked example: you press the left button at 12.000 s, 12.050 s and 12.250 s on the session clock. The gaps are 50 ms and 200 ms. With the default 80 ms limit, the second press is flagged, the closest pair reads 50 ms, and the verdict becomes a warning. Raise the limit to 250 ms and both later presses are flagged; lower it to 40 ms and neither is. Flags are recomputed from stored timings, so moving the slider never loses data. Each log line also shows the time since the previous release, which helps separate a bouncing switch from a real double tap.
Each button tile also carries a status. It reads "Not tested yet" until the first press, a count such as "7 of 20 presses" until the button has 20, "OK" once it has 20 or more with nothing flagged, and "Possible fault" with the number of problem presses as soon as one is flagged or counted as a quick re-press. Worked example: you press Left 20 times about a second apart and Right 7 times, and never touch Back. The table reads "OK – 20 clean presses" for Left, "Keep going (7 of 20)" for Right and "Not tested" for Back. "Buttons OK" shows 1 of 2, because untested buttons are left out, and the verdict asks for more Right presses. OK describes this session only; faults can be intermittent.
A notched wheel typically sends one wheel event per notch, and that is what "Scroll up" and "Scroll down" count. Trackpads and free-spinning wheels send many small events per gesture, so their counts run higher. A direction flip is a wheel event in the opposite direction less than 150 ms after the previous one. Reversing a wheel that fast is uncommon in normal use, so repeated flips while you scroll steadily in one direction can point to a dirty or worn wheel sensor. A deliberate quick flick up and down can also trigger one, so judge by repeats, not a single flip. The delta is shown in the unit the browser reports (pixels, lines or pages). Chrome and Edge usually report pixels, while Firefox can report lines, so one notch may read differently across browsers.
"Browser double-clicks" counts the browser's own double-click event, which follows the operating system's double-click rule. Windows, for example, defaults to 500 ms between the two clicks, far longer than the gap this detector flags.
Diagnosing a double-click fault
A double-click fault means one physical press produces two presses. Most mice put a small mechanical switch under each main button, and its metal contacts close when you press. As the contacts wear, or a film builds up on them, they can part and touch again a few milliseconds apart while you are still pressing. Optical switches have no metal contacts, so this explanation does not apply to them. Spec sheets quote anywhere from a few million to tens of millions of clicks, but that does not stop an individual switch from failing early.
| What you see | Likely cause | Next step |
|---|---|---|
| Flags on one button, gaps under about 40 ms | Switch contacts bouncing | Try the fixes below, then the warranty |
| Flags on every button | Software, macro or driver | Close vendor apps, test on another PC |
| Flags only when you click fast | Normal fast clicking | Lower the limit to 40 ms and repeat |
| A button lets go while held, or clicks twice on release | Worn switch or loose contact | Hold it for 3 seconds, click slowly, read "Quick re-presses" |
| A button never turns green | No signal reaches the browser | Check cable, receiver and remapping |
To separate hardware from software, run the same 30 slow clicks on a second USB port and, if you can, on another computer. A fault that follows the mouse is in the mouse. A fault that stays behind is in the computer.
The test cannot read intent: if you tap twice in under 80 ms on purpose, it flags you. Slow, deliberate clicks make the result meaningful.
Testing a button that lets go or bounces on release
A worn switch can fail in two ways that the press-to-press limit misses. It can open for a few milliseconds while you hold it, which drops a drag or a text selection halfway. Or it can bounce as you let go, so the browser sees the release and then a second press. If the press lasted longer than the limit, the gap between the two presses looks normal.
The detector therefore also counts a quick re-press: a press that comes less than 30 ms after the same button's release and was not already flagged. The 30 ms figure is this tool's design choice, not a measured human limit, and very fast clicking techniques can trip it, so click slowly.
Worked example: Left is pressed at 0.000 s, released at 1.500 s, pressed again at 1.510 s and released at 3.000 s. The gap between the presses is 1,510 ms, so the 80 ms limit stays clear. The second press came 10 ms after a release, so "Quick re-presses" reads 1, the log line turns red and the verdict warns. If Windows ClickLock is on, turn it off before a hold test.
Fixes to try before replacing a mouse
- Rule out software. Quit vendor apps, autoclickers and macro tools, then retest. Some gaming mice have a debounce or click-delay setting, and raising it slightly can mask a mild fault at a small cost in responsiveness.
- Swap the connection. Try another port without a hub, another cable if it detaches, or the wireless receiver on an extension lead.
- Retest on a second computer. It settles the software question in two minutes.
- Read the warranty. Terms vary by maker and country, and a switch that fails early may be covered. The copied report lists the flagged presses and their gaps.
- Decide on repair or retirement. Replacing a switch takes soldering skill and opening the mouse may void the warranty, so it rarely pays off on a cheap model.
When a browser test is not enough
This page sees what reaches the browser after the operating system and the mouse driver. That answers the question most people have, whether a button reaches their apps, with nothing to install. A few checks need to look below the browser.
| Situation | What to use |
|---|---|
| A button never registers here | Check that the primary button is not swapped ("Primary mouse button" in Windows 11 mouse settings) and that vendor software has not remapped it. On Linux, run sudo evtest, pick the mouse from the list and press the button. It prints the kernel's button codes, such as BTN_LEFT, BTN_RIGHT, BTN_MIDDLE, BTN_SIDE and BTN_EXTRA, so you can see whether the hardware sends the press at all. |
| Double-click feels wrong in every app, but nothing is flagged here | The Double-click speed slider on the Buttons tab of Windows Mouse Properties. Windows 11: Settings > Bluetooth & devices > Mouse > Additional mouse settings. Windows 10: Settings > Devices > Mouse > Additional mouse options. |
| Holds and drags act strangely | Clear "Turn on ClickLock" on the same Buttons tab. ClickLock lets you highlight or drag without holding the button down. |
| You want a debounce setting or a firmware update | The mouse maker's own app |
Run this page again afterwards to confirm the fix.
Polling rate explained
Polling rate is how many times per second the mouse reports to the computer. Each report carries the movement since the previous one, so a higher rate means smaller steps between cursor updates. The interval is 1,000 divided by the hertz.
| Rate | Time between reports |
|---|---|
| 125 Hz | 8 ms |
| 500 Hz | 2 ms |
| 1,000 Hz | 1 ms |
| 2,000 Hz | 0.5 ms |
| 4,000 Hz | 0.25 ms |
| 8,000 Hz | 0.125 ms |
The estimate uses the browser's pointerrawupdate event, which reports pointer changes as soon as the browser receives them. It exists in Chromium browsers such as Chrome and Edge, on secure (HTTPS) pages only. Where it is missing, the box is hidden rather than showing a guess. The tool finds runs of continuous movement lasting at least 0.3 seconds, divides the updates in each run by its length, and keeps the highest.
Worked example: a run lasts 0.8 seconds and contains 801 updates. That is 800 intervals, so 800 ÷ 0.8 = 1,000 updates per second, which matches a 1,000 Hz setting. The figure snaps to a common rate within 15 percent of one, runs above 8,800 Hz are ignored as timing artifacts, and the estimate is capped at 8,000 Hz.
Slow movement reads low, because a mouse reports only while it moves. A busy computer, power saving and a weak wireless link can also delay reports, so take the highest reading from several fast runs.
DPI, sensitivity and polling rate
| Term | What it controls | Changing it affects |
|---|---|---|
| DPI (or CPI) | Sensor counts per inch of movement | How far the cursor travels |
| Sensitivity (game or system) | A multiplier applied after the sensor | How far the cursor or view travels |
| Polling rate | How often reports are sent | Smoothness and latency |
A common shorthand multiplies DPI by in-game sensitivity. At 800 DPI and a sensitivity of 0.5, the product is 400, and in the same game 400 DPI at 1.0 gives a similar speed. Pointer acceleration, such as Windows' "Enhance pointer precision", varies speed with movement and breaks such comparisons.
Who uses a mouse test
- Used-mouse buyers run 30 slow clicks on every button and scroll both ways before paying.
- IT support staff test a reported sticky button on a spare computer to see whether the fault follows the mouse.
- Gamers confirm side buttons register after installing software and compare their polling rate with the setting they chose. Each button needs its own test, because a clean left button says nothing about the right one.
Tips and common mistakes
- Click slowly for the double-click check. Fast clicking belongs in the click speed test.
- Mind swapped buttons. If you switched the primary button in system settings, the browser normally reports the swap, so the physical right button may show as Left.
- Expect missing side buttons after remapping. Vendor software that turns them into shortcuts hides them from the page.
- Test the rest of the desk. The keyboard tester catches keys that double-type, and the dead pixel test checks the monitor.
Privacy and limitations
The page reacts only to events over the test area and keeps the counts in memory, so closing the tab erases them. Only the detector limit and the scroll toggle are saved, in your browser's local storage.
The browser sees what the operating system and driver pass on, not the raw switch signal, so a fault hidden by the mouse's own firmware will not show. Browsers may round timestamps, so gaps of a few milliseconds are approximate. Back and forward blocking applies while the pointer is over the box and can vary by browser and driver. Touch input is ignored. This page is not affiliated with any mouse maker or browser vendor.
Frequently asked questions
How do I test my mouse buttons online?
Move the pointer into the test area and press each button once: left, right, middle (the wheel click), and the back and forward side buttons if your mouse has them. A button turns orange while it is held and green once the browser has registered it, and its tile counts the presses. If a button never changes color, the browser never received a press from it, which points to the switch, the cable or the mouse software.
How do I know if my mouse is double-clicking by itself?
Click one button slowly, about once a second, for at least 20 presses, until its tile reads OK, while you watch the double-click detector. A healthy mouse shows no gaps below the limit. A worn switch sometimes registers a second press a few dozen milliseconds after the first, and the tool flags it in red with the exact gap. Faults can be intermittent, so repeat the test on each button and on a second USB port before you decide.
What limit should I use for the double-click detector?
The default is 80 ms, which equals 12.5 presses per second from a single button. Deliberate single clicks almost never come that close together, while bounce-type faults usually show up as gaps of a few milliseconds to a few tens of milliseconds. If a very fast clicker triggers flags on purpose, lower the limit toward 40 ms. To catch milder faults, raise it toward 120 ms and click slowly to avoid false alarms.
How do I fix a mouse that double-clicks?
Start with the cheap checks: try another USB port or computer, update or reset the mouse software, remove macros, and raise any debounce or click-delay setting it offers. Then look at the warranty, because switches can fail early. Some people report that cleaning the switch helps for a while, but a worn contact usually returns. Past that, the switch can be replaced by someone comfortable with a soldering iron, or the mouse retired.
Why don't my back and forward buttons show up?
The browser only reports them when the mouse driver sends them as standard buttons four and five. Some vendor software remaps side buttons to keyboard shortcuts or macros, so the page sees a key press or nothing at all. Try closing or disabling the vendor software and press again. On a mouse with only left, right and wheel buttons, the back and forward tiles will simply stay untested.
What is polling rate and how do I check it here?
Polling rate is how many times per second the mouse reports to the computer: 125 Hz means every 8 ms, and 1,000 Hz means every 1 ms. In Chrome or Edge, the estimate box counts position updates while you circle the pointer quickly and shows the highest sustained rate, capped at 8,000 Hz. It is an estimate, and slow movement or a busy computer can read low. Other browsers lack the needed event, so the box is hidden.
Do I need to download a program to test my mouse?
No. To check that each button registers, catch double clicks and test the wheel, this page is enough, and it runs in your browser with nothing to install. An installed tool helps when a button never shows up here: the Buttons tab in Windows' Mouse Properties holds the double-click speed and ClickLock settings, and on Linux evtest shows the raw button codes the kernel receives.
Is anything I do in the mouse test recorded or uploaded?
No. The page listens to mouse events in your browser and keeps the counts in memory until you reset or close the tab. Nothing is sent to a server. Only two settings are saved in your browser's local storage: the double-click limit and the scroll option. The report you can copy or download is created on your device and contains only what you see on the page.