Mouse wheel test
Put the pointer in the area below and roll the wheel. The page shows the scroll direction, the accumulated notches and the direction of each notch in real time, and marks it out when you keep scrolling one way yet a reversal shows up — the classic look of a wheel encoder starting to skip.
Direction of the last 80 notches
Scrolling down sinks to the bottom, scrolling up rises to the top. The red ones are reverse glitches — you kept scrolling one way, and one notch alone jumped the other way.
How to read the result
- Suspected skips is the only number here that points at a fault. Its definition is narrow: one isolated reversed event, with the same direction on both sides. Steady scrolling never produces that sequence, so when it appears it is worth suspecting. But it is not proof — a slipping finger, a trackpad mis-touch or the system doing smooth interpolation can all produce the same shape. So the right way to use it is: scroll steadily one way for twenty-odd notches, and if the red keeps appearing, that is the encoder.
- Direction reversals are not a fault indicator on their own, which is why they are not on the panel above. Every time you change direction it goes up by one, and a few trips up and down pile up a heap of them — using it to judge a fault would only scare you.
- Raw notch value tells you what deltaY your system reports. Check it at a glance: on Windows, Chrome is usually 100 (pixel mode) and Firefox usually reports 3 (line mode). If your number looks nothing like either, the system or driver has changed the scroll settings, and the notch conversion is then only a rough guide.
What the browser actually gives you
A wheel event carries deltaY plus a deltaMode, and the latter only says whether the unit is pixels, lines or pages — it does not say how tall a line is or how many pixels a notch is. So any attempt to turn deltaY into a "notch count" carries an assumption. The assumptions this page uses are written out in the table below, so you can check them against your own setup:
| deltaMode | Name | Meaning | Treated as one notch |
|---|---|---|---|
| 0 | DOM_DELTA_PIXEL | Counted in pixels | 100 |
| 1 | DOM_DELTA_LINE | Counted in lines | 3 |
| 2 | DOM_DELTA_PAGE | Counted in pages | 1 |
And precisely because the conversion is only an estimate, this page treats direction as the main indicator. Direction needs no conversion at all — just read the sign of deltaY. It is unaffected by system settings and consistent across browsers.
Common questions
No, it is an estimate. The browser only tells you the value and the unit of deltaY (pixels/lines/pages); it never says how big a "notch" is — the system setting for how many lines one scroll moves, the driver’s smooth scrolling, and whether the mouse is a free-spinning wheel all change that number. This page converts by the common conventions (pixel mode: 100 per notch, line mode: 3 lines per notch), so what is really trustworthy is direction and direction reversal — those need no conversion at all.
The most typical sign is a "reverse glitch": you keep scrolling one way, and in the middle one event suddenly jumps the other way, with normal scrolling on either side. This page counts that shape on its own — it is the "Suspected skips" figure below. Note that it is only a suspect: a slipping finger or a trackpad mis-touch can produce the same shape. So the way to read it is: scroll steadily in one direction for twenty-odd notches and see whether a glitch appears.
That is deliberate. If the page scrolled along, the test area would slide out from under the pointer and the more you scrolled the less it would measure. So this page intercepts the scroll events — it only counts, it does not scroll. To go back to normal browsing, move the pointer out of this area.
A free-spinning wheel (the kind that keeps turning and glides on under its own momentum) sends dozens or hundreds of small consecutive deltas in a single flick, so the strip refreshes very fast and the accumulated notch count comes out absurdly large — that is how it is meant to work, not a fault. To judge this kind of mouse, look at whether the direction is consistent, not at the notch count.
