Measure repeatability, not a universal score
Scroll event rate is shaped by hardware, firmware, settings, the operating system, the browser, and how you move the wheel. The most defensible use is comparing controlled runs from one setup.
How to use the Mouse Scroll Speed Test
Set a duration
Choose 5, 10, 15, or 30 seconds. Use the same duration for every scroll speed comparison.
Wait for the countdown
Press Start, move the pointer into the capture card, and begin when the countdown finishes.
Compare up to five runs
Keep your grip and wheel mode consistent, then review EPS, timing, peaks, and variation together.
What a scroll speed test can measure
This scroll speed test counts `wheel` events delivered to the page during a fixed interval. Events per second is total captured events divided by run duration. The chart also divides the run into fixed one-second buckets, revealing whether input arrived steadily or in bursts. Median interval represents typical spacing between adjacent events, while p95 interval highlights the slower end of the observed timing distribution without treating a single maximum as typical.
Average absolute delta adds context because two runs can produce similar event counts but different raw magnitudes. It still is not a physical speed measurement. A driver can turn one motion into line-mode steps, pixel-mode values, acceleration, or smoothed messages. The mouse scroll speed result therefore belongs to the exact browser and settings used for that run.
Keep each mouse scroll speed run comparable
Use the same test duration, browser zoom, device connection, wheel mode, hand position, and movement pattern. Begin only after the countdown so the button press does not alter your rhythm. Keep the tab visible; background throttling or a focus change can invalidate a timing comparison. Run the scroll rate test at least three times rather than choosing the fastest attempt, then look for a stable range and investigate only repeatable changes.
The optional history keeps five summaries. By default it disappears on refresh. Enabling Save on this device writes only those summaries to localStorage, so you can revisit them on that browser. The coefficient of variation describes spread relative to the mean EPS; it is not a percentile, ranking, quality grade, or comparison against other users.
Why systems produce different scroll speed data
A notched wheel, free-spin mode, touchpad gesture, trackball, or remapped input can all generate different event sequences. Vendor software may add smoothing or per-application profiles. The operating system can apply scroll speed or lines-per-step preferences, and a browser can expose line, pixel, or page units. Even the page receiving the input can influence default scrolling. These layers explain why a scroll speed test is not portable like a stopwatch measurement.
Raw delta comparisons across devices are especially risky. A value of 100 in pixel mode and a value of 3 in line mode do not share a physical scale. Even values with the same `deltaMode` can have different origins. For repeatability, compare the same device, connection, system, browser, and configuration; record any change before interpreting the next mouse scroll speed run.
Events per second is not polling rate
Hardware polling rate describes how often a device or connection reports state in its own input pipeline. Wheel events per second counts messages that a browser dispatches to this page. Operating-system processing, browser scheduling, coalescing, wheel mechanics, and user motion stand between those concepts. A high EPS result does not establish a high polling rate, and a low EPS result does not establish a low polling rate.
The number also does not equal physical wheel notches. One notch can lead to one or several browser messages, and smooth scrolling may continue after a mechanical movement. Use the scroll speed test to compare the responsiveness and regularity you observe in one environment, never to publish a hardware specification that the browser did not measure.
Practical uses for a wheel-event rate
For gaming or rapid list navigation, compare whether repeated runs have similar event timing and whether the application reacts as expected. For long documents, steadiness may matter more than the highest peak. In free-spin use, a tall early burst followed by decay can be normal. The chart helps distinguish those shapes without inventing a universal better score. Your comfortable, controllable pattern is more meaningful than a global leaderboard.
If a new driver or setting changes scroll behavior, take a before-and-after set under identical conditions. If only one program feels wrong but this browser result is repeatable, inspect that program’s scroll settings. If timing becomes irregular across applications, continue with the jitter test and the troubleshooting checklist while preserving the same controlled setup.
Mouse Scroll Speed Test questions
What is a good score in the scroll speed test?
There is no universal score. A useful result is a repeatable range that suits your task in the same device and software environment.
Why is p95 higher than the median interval?
The median represents typical spacing, while p95 is near the slower end. Pauses and uneven bursts naturally make p95 larger.
Should I save run history?
Only if local comparison is useful to you. Saving is optional and stores up to five summaries on this browser, not on a server.
Can I compare two different mice by raw delta?
Not reliably. Different units, drivers, smoothing, and wheel modes can make raw delta values incomparable.