Understand the signal before drawing a conclusion
A browser can reveal useful event patterns, but those patterns sit at the end of a chain involving the device, its firmware, drivers, operating-system settings, and the browser. This guide keeps that boundary visible throughout the test.
How to use the Free Online Scroll Wheel Test
Choose the test setup
Select 15, 30, or 60 seconds, or Manual. Use Free for intentional direction changes; use One direction for a controlled same-direction scroll wheel test that establishes a dominant sign.
Capture a representative run
Press Start and keep the pointer inside the capture area while scrolling. For broader report coverage, collect at least 40 vertical events across three strokes separated by brief pauses and over at least one second. Avoid intentional reversals in One direction mode.
Generate and interpret the report
Let the timer finish or select Finish & build report after recording input. When local report generation completes, the page moves to the report automatically. Review the diagnostic overview, charts, flagged details, and parameter guide, then repeat under the same conditions before drawing a conclusion.
What this scroll wheel test measures
The scroll wheel test counts standard browser `wheel` events only after you start a session and only while the pointer is inside the capture card. It records relative timing, `deltaX`, `deltaY`, `deltaZ`, `deltaMode`, and modifier-key states. From that local event stream it calculates event count, a recent one-second rate, median timing, accumulated absolute delta, observed extremes, and a short bounded trace. These are browser-received input values, not physical distance, wheel revolutions, or a measurement of the encoder itself.
In One direction mode, the mouse wheel test uses the first valid samples to determine the dominant sign for the action you chose. Later events with that sign are expected-direction events; events with the opposite sign are candidates worth reviewing. Free mode intentionally permits direction changes. That distinction prevents ordinary up-and-down exploration from being mislabeled as a fault and makes the scroll test useful for both inspection and a more controlled check.
How to run a reliable mouse wheel test
Place the mouse on a stable surface, keep the same grip, and decide whether you are testing short notched steps or a continuous free-spin motion. Close overlays that might intercept input, keep the browser tab visible, and avoid changing operating-system or vendor settings between runs. Start the scroll wheel test, move into a steady rhythm, and complete several strokes. For a one-direction run, do not intentionally reverse until the session ends because the calibration assumes one requested action.
A useful comparison controls the environment. Repeat the wheel input test in the same browser, at the same zoom, with the same device mode and similar hand movement. Then repeat in another browser or application only when you are trying to locate an environment-specific difference. A single surprising row may be caused by momentum, remapping, focus changes, or timing variation; a pattern that repeats under controlled conditions carries more troubleshooting value.
How to read each scroll test metric
Total events is the number of `wheel` messages captured during this session. Events in the last second describes a rolling browser event rate, not hardware polling rate. Median interval summarizes typical spacing between messages and is less affected by a single pause than an average. Latest, minimum, maximum, and cumulative absolute `deltaY` describe raw vertical signal values in the current `deltaMode`; they do not represent centimeters or a guaranteed number of physical notches.
A nonzero `deltaX` count means the browser received horizontal components. That could follow a tilt wheel, a trackpad gesture, Shift mapping, software remapping, or another compatible source; the scroll wheel test cannot identify the hardware source reliably. `deltaMode` says whether delta values are expressed as pixels, lines, or pages. If units change during a session, raw magnitudes should not be merged into one hardware comparison.
Reverse scrolling, skipping, jitter, and dead input
Reverse scrolling can mean an operating-system direction preference, an intentional direction change, inertial input, or a short unexpected opposite-sign sequence. The controlled mode calls these sequences candidates because the browser event alone cannot reveal their cause. Skipping describes missing or uneven movement as experienced by the user, while jitter usually describes instability such as small reversals, bursts, or irregular timing. Dead input is harder to prove: a browser cannot record an action for which it receives no event.
When the main scroll wheel test shows a repeatable candidate, use the dedicated direction and jitter tools. Direction testing separates an Up run from a Down run and recalibrates each action. Jitter testing groups input into strokes and uses robust timing and amplitude checks. If the browser records nothing, compare applications, check focus and settings, then inspect power or connection conditions before assuming a mechanical cause.
Mouse wheel and touchpad input are not interchangeable
Mouse wheels often produce step-like bursts, but firmware and software can accelerate, smooth, combine, or remap those steps. Touchpads commonly produce smaller, more frequent deltas with momentum after the fingers leave the surface. Trackballs, accessibility devices, and synthetic inputs can also create `wheel` events. The browser normally exposes the event, not a trustworthy device category, so this mouse wheel test reports the observable pattern without claiming to identify the source.
Choose Free mode for a touchpad-like gesture because deliberate reversals and inertia are normal. Choose One direction only when you can make a controlled action and understand that calibration belongs to that session. Raw delta values are most useful when comparing repeated runs from the same setup. Comparing a trackpad’s pixel-mode deltas with a notched mouse’s line-mode deltas would mix different pipelines and yield a misleading benchmark.
Browser, operating-system, and privacy limitations
The operating system may apply natural scrolling, speed settings, line counts, accessibility features, or application profiles before the event reaches this page. The browser may coalesce activity, expose different units, reserve Ctrl-plus-wheel for zoom, or allow a page to scroll without a one-to-one relationship between input and movement. Background tabs may be throttled. For these reasons, a scroll test session is evidence about this environment at this moment, not a manufacturer-grade specification.
All wheel samples, timestamps, summaries, charts, and exports are created in your browser. The site has no account or server endpoint that receives the tool's in-memory samples or exported files. Microsoft Clarity and Google Analytics measure site usage separately. JSON and CSV downloads contain the captured fields but no screen coordinates or identity fields. Theme and language preferences may be stored locally, and speed history is stored only when you explicitly enable that option. Resetting a session removes the active in-memory samples.
Free Online Scroll Wheel Test questions
Can this scroll wheel test confirm that my mouse is healthy?
No. It can show whether the browser received a consistent pattern during a controlled session. It cannot inspect physical components or provide a warranty-grade conclusion.
Why does deltaY change between browsers?
Drivers, operating-system settings, browser behavior, delta units, acceleration, and input mode can all change the values exposed to a page.
Does events per second equal mouse polling rate?
No. It counts browser wheel events during this run. Hardware polling rate is a different property and cannot be inferred from this number.
Why is normal up-and-down scrolling marked as mixed?
Mixed is expected if you reverse direction during a controlled one-direction run. Use Free mode when deliberate direction changes are part of the test.
Can I share a result without sending it to the site?
Yes. Copy the summary or create a local JSON or CSV file. Export happens in the browser and does not upload the session.