Device tests · Input Devices
Mouse Polling Rate Test
Move a physical mouse continuously inside the test area for ten seconds to estimate the rate of mouse events delivered to this browser. The result includes median timing, average rate, sample count and consistency.
This is a browser-observed event rate, not direct USB polling measurement. A lower estimate cannot establish that a mouse configured for 1000 Hz or more is faulty.
- Local observations
- No hardware certification
Move a physical mouse continuously in circles inside the area during a run. Touch and pen input are excluded.
Mouse movement area
Only mouse events inside this area are sampled.
Browser-observed event rate
Input method: Not started. One event timestamp per delivered event; no coalesced-event double counting.
No completed estimate yet.
Your result
What your result means
The primary estimate is 1000 divided by the median valid event interval in milliseconds. Median timing is less sensitive than a single fastest event, which can reflect scheduling artifacts. The displayed average uses valid interval count divided by their accumulated time.
Intervals over 100 ms are excluded as movement pauses and reported separately. Accumulated sample time is therefore not the same as the ten-second wall-clock session. With very few intervals, the page marks the evidence insufficient instead of implying a stable hardware rate.
Mostly consistent means median absolute deviation is no more than 15% of the median with at least 30 valid intervals. Variable intervals mean the browser-delivered timing varied more. Neither classification describes manufacturer compliance or input-to-display latency.
Procedure
How the test works
- 01
Prepare a physical mouse
Use a supported connection and comfortable sensitivity. Touchscreens and pen pointers are excluded. Keep the tab visible and avoid running an unrelated heavy task at the same time.
- 02
Start the session
Press Start Polling Test. The page chooses pointerrawupdate where supported, otherwise pointermove. It does not combine both streams or add coalesced samples on top.
- 03
Move continuously
Make smooth circles inside the area for ten seconds. Slow movement or stops can reduce events. Stay inside the area so the page observes a consistent region.
- 04
Read the evidence
Compare the median estimate, average, median interval, valid samples and excluded pauses. A run with little movement is not useful for comparison.
- 05
Repeat comparable runs
Keep connection mode, power settings, browser and movement similar. Compare OS or manufacturer settings separately; do not snap the estimate to a marketed number.
- 06
Cancel if interrupted
Stop and Reset remain keyboard accessible. Hidden-tab sessions are cancelled because background scheduling would make the run unreliable.
Methodology
How Testpartout measures this
This implementation uses browser observations and local calculations, with no hardware certification.
What the implementation does
- One high-resolution event.timeStamp is collected for each delivered mouse event inside the active area and ten-second session.
- pointerrawupdate is used when the browser exposes it; pointermove is the fallback. No coalesced-event expansion is performed, avoiding duplicate counting.
- Positive consecutive intervals at most 100 ms form the calculation set. Duplicate/nonpositive times are discarded, and longer gaps are counted separately.
- The primary value comes from the median interval; average rate uses summed valid intervals. Consistency uses median absolute deviation and a minimum sample count.
- The active listener and timer are removed on Stop, completion, hidden-tab cancellation and route exit. Raw timestamps are not persisted.
Limits of a browser measurement
- Physical USB reports are not directly available to this browser tool. Operating-system and browser scheduling can combine, delay or omit delivered events.
- High-Hz devices may exceed what a particular browser exposes. A 4000 or 8000 Hz setting cannot be certified by a browser event stream.
- Pointer movement speed, power saving, wireless mode, system load and window focus affect comparability. This is not a latency test or DPI detector.
Privacy
What happens to your data
This section describes the behaviour of the code on this page, not an intention.
- Diagnostic events, coordinates, timings and observations stay in current page memory. They are not uploaded or saved as raw diagnostic data.
- No camera, microphone, account or server storage is used by this test.
- Only start and complete lifecycle events may be sent after analytics consent; no result values are included.
Troubleshooting
If the test does not behave as expected
A 1000 Hz mouse shows less
Check continuous movement and sufficient samples first. Browser coalescing or scheduling can lower the observed rate even when hardware reports at its configured setting. Compare another supported browser without treating either as a USB analyzer.
Large gaps
Pauses or leaving the area create gaps. They are excluded from active interval time, not silently counted as ordinary slow samples. Repeat if pauses dominate the run.
Raw events unavailable
pointermove is a valid fallback but may be synchronized or coalesced by the browser. The displayed method explains which path was sampled; a fallback is not a device fault.
Factors
Interpret observations in context
Wireless and power modes
A manufacturer may offer different reporting settings across Bluetooth, a dedicated receiver and wired mode. Compare only supported documented modes, with adequate charge and stable receiver placement.
System load
Heavy tasks or power-saving settings can alter scheduling. Repeating under quieter conditions improves comparability, but does not transform the browser into a calibrated measurement instrument.
Touchpad versus mouse
A touchpad can produce mouse-like pointer events and may be measured by this browser path. That does not establish its USB polling rate or make its rate directly comparable with a gaming mouse.
FAQ
Frequently asked questions
- Does this measure true USB polling rate?
- No. It estimates delivered mouse pointer-event rate. Browser and operating-system processing stand between hardware reports and the page.
- Why not show the fastest event as the result?
- One unusually short interval can be a scheduling artifact. Median timing gives a more robust primary estimate; average and sample evidence provide context.
- How long does a run last?
- Approximately ten seconds. Positive event intervals up to 100 ms are used; longer gaps are excluded as pauses and reported separately.
- Are coalesced events double-counted?
- No. This implementation uses one timestamp per event from one selected stream and does not expand getCoalescedEvents.
- Can it verify 8000 Hz?
- Not reliably. Browser event delivery may limit high-Hz readings, so this tool does not certify any advertised hardware polling rate.
- Does analytics receive mouse Hz?
- No. Timestamps, Hz, sample counts and consistency remain local. Only consent-controlled start and complete events may be sent.
Found an issue with this test? Report it.