Before / after check
Compare scroll test results
One change, two runs, one fair comparison. The guided lab below walks through fixed phases, saves the first run as Run A, then re-runs the identical phases as Run B after you change exactly one condition. Everything stays in your browser.
Scroll Signature Lab
Idle01What problem are you investigating?
02Input type
A guided before/after check. Compare two browser-observed runs locally. No account required. All data stays in this browser. Nothing is uploaded and nothing is put in the URL.
03Phase
Focus here, then scroll
waiting
04Event river
Demonstration — synthetic data, not measured from your hardware.Event river: order and relative size of received wheel events. One mark per event; height reflects size. Meaning is in the labels and counts, not in color.
Text table alternative
Change exactly one condition, then record Run B with the same settings.
A / B
05Saved runs (this browser only)
All data stays in this browser. Nothing is uploaded and nothing is put in the URL.
06Safe troubleshooting path — repeat the same test after each single change
- Focus the test area and repeat with a steady motion.
- Change only the USB port, then repeat the same test.
- Change only the browser, then repeat the same test.
- Disable extensions or try a private window, then repeat the same test.
- Check OS scroll speed / direction, make one change, then repeat the same test.
- With the mouse powered off, wipe around the wheel with a dry cloth, then repeat the same test. Do not open it or put liquid inside.
- Try another computer, then repeat the same test.
Differences describe browser-observed input only. A change does not prove a repair.
Comparability checklist
Two runs are comparable only when every item below matches. The tool warns you when they do not — heed the warning instead of reading a difference as an improvement.
- Same input type (standard wheel or touchpad / free-spin on both runs).
- Same phase set and phase durations (the lab enforces this when you use Record Run B).
- Same browser, or a deliberate single-condition browser change you can name.
- At least 8 events in each run; no stopped-early partial samples.
- No interruptions (focus loss, hidden tab) on either run.
- Exactly one changed condition between A and B — port, browser, extensions, OS setting, or cleaning step.
Reading the comparison
The table puts duration, event count, direction pattern, opposite-direction observations, unit types, and test conditions side by side. A difference describes browser-observed input only — it does not prove a repair, and it does not diagnose hardware. If the verdicts differ, repeat each setup once more before drawing conclusions.
Prefer the quick single-run tool? Run the mouse scroll test. Want the method behind the numbers? Read how the scroll test works.