About the workbench
Browser-observed scroll diagnostics, explained
ScrollTest is a free mouse-wheel diagnostics and repeatability tool. It records the events and delta values delivered to your browser, then makes the limits of those observations as visible as the results.
Why it exists
A scroll-wheel problem can be difficult to describe. A page may jump in the wrong direction, input may arrive intermittently, or one browser may feel different from another. A trace that separates up, down, horizontal-only and true-zero events, alongside timing and delta units, gives you a repeatable starting point for troubleshooting.
The same trace cannot prove that a physical detent occurred or failed. Device firmware, drivers, operating-system settings, momentum, acceleration, browser scheduling, and application load all sit between an action and JavaScript. ScrollTest reports that uncertainty instead of hiding it behind an accuracy score.
Three useful intents
Diagnostic
Run controlled down and up phases at normal and faster paces. Inspect reversals, horizontal-only events, true-zero events, gaps, and bursts; an under-sampled phase is inconclusive, not clean.
Repeatability
Run the same motion three times with one device, browser, duration, input category, and declared setup. Compare the median event rate and run-to-run variation without mixing incompatible runs.
Challenge
Try a 5, 10, or 30 second browser-events-per-second challenge. Valid results show and retain callback-delay context; a delay over 250 ms invalidates the run.
What ScrollTest does not claim
- Browser events are not labelled as physical wheel notches.
- Line and page deltas are not relabelled as pixels.
- A short browser sample is not a definitive hardware diagnosis.
- There are no invented gamer bands, universal quality grades, or percentiles.
- Mouse and touchpad results are not treated as interchangeable.
- User-editable shared summaries are never presented as verified records.
Transparent by design
The methodology defines every field, unit, timing rule, comparison boundary, and diagnostic limitation. Test history stays local to your browser. A normal report records the declared setup and input-device connection context, protocol, browser and OS, report ID/date, metric version, duration, validity and callback delay, with a link to the methodology. Raw traces remain local unless you explicitly choose a raw export, and ordinary share links contain only a small self-reported summary.