- Click your blocker’s icon in the toolbar (uBlock Origin, AdBlock, Ghostery, Brave Shields, etc.)
- Turn it off for
clickspeedtest.info - Reload the page
Mouse wheel diagnostic
Scroll Test
Put the pointer in the box and scroll. This checks that every notch registers, in both directions, and looks for the two faults a plain scroll counter cannot see: phantom reversals and skipped detents from a worn encoder.
Scroll inside this box
Roll the wheel up and down. Tilt it left or right too, if your mouse can.
Total ticks
0
Peak ticks/s
—
Reversals
0
Step consistency
—
Scroll at a normal pace first, then try a fast flick. Encoder faults usually appear at speed, not at a slow careful roll — which is exactly why a wheel can feel fine in a document and misbehave in a game.
What this scroll test checks
A mouse wheel is a rotary encoder: a notched disc that interrupts a beam or closes contacts as it turns, producing one pulse per detent. When it wears, it does not usually stop working. It starts producing pulses that do not match what your hand did.
This test records every wheel event the browser reports and looks at four things:
- Direction coverage. Up, down, and horizontal tilt if your mouse has it. A direction that never registers is a dead channel on the encoder.
- Reversals. A tick in the opposite direction to the previous one, arriving within 120 ms. Your hand cannot change direction that fast mid-scroll, so this is the encoder misreading its own rotation.
- Step consistency. How uniform the reported scroll distances are. A healthy notched wheel produces one repeated value; a worn one produces scattered values as detents get partially read.
- Peak rate. Ticks per second at your fastest, which is where faults surface.
Reading your results
| What you see | What it means |
|---|---|
| Reversals: 0 | Healthy. The encoder is reading rotation direction correctly. |
| Reversals: 1–2 over a long test | Usually you genuinely changed direction quickly. Reset and scroll one direction only to confirm. |
| Reversals climbing during one-direction scrolling | Encoder wear or contamination. This is the classic “page jumps back up” fault. |
| Step consistency above 90% | Normal notched wheel with clean detents. |
| Step consistency below 60% on a notched wheel | Detents being partially read, or OS smooth-scrolling is interpolating. Check the wheel type line below the stats. |
| One direction counts far lower than the other | A weak channel. Often reversible with contact cleaner; often not. |
| Tilt counts stay at zero | Normal for most mice — only tilt-wheel models report horizontal deltas. |
Notched wheels, free-spin wheels, and why the numbers differ
Two mice can produce completely different readings here while both being perfectly healthy, because they are built differently.
| Wheel type | What the test sees |
|---|---|
| Notched (stepped) | One consistent step value repeated. High step consistency. Most office and gaming mice. |
| Free-spin / infinite scroll | Many small varied deltas in rapid succession. Low step consistency, very high peak rate. Not a fault. |
| Dual-mode (clutch) | Behaves as either, depending on the mode the wheel is engaged in. Test both. |
| Trackpad or touch | Continuous pixel deltas, no detents at all. Step consistency will be low and that is expected. |
Before concluding your wheel is faulty: operating-system smooth scrolling also converts clean detents into streams of small interpolated deltas, which looks identical to a free-spin wheel from inside a browser. If your step consistency is low but scrolling feels normal, smooth scrolling is the likelier explanation. Turn it off in your OS or browser settings and retest.
Why the same wheel scrolls different amounts in different browsers
Browsers do not report scrolling in a single unit. Each wheel event carries a delta mode saying what the number means, and the test displays which one your browser used.
| Delta mode | Unit | Typically seen on |
|---|---|---|
| 0 — pixel | Pixels to scroll | Chrome, Edge, Safari, and trackpads generally |
| 1 — line | Lines of text | Firefox on Windows and Linux with a notched wheel |
| 2 — page | Whole pages | Rare; some accessibility and legacy configurations |
This is why a wheel that moves three lines per notch in one browser moves a hundred pixels in another. It is not a hardware difference and not a fault. It is also why comparing raw scroll numbers between two people’s tests is meaningless unless both are in the same delta mode.
How a scroll wheel actually works
Understanding the mechanism makes the failure modes obvious, and it is the part almost every scroll-testing page skips.
Inside the wheel is a rotary encoder: a small drum with a slotted or toothed wheel and two sensing channels, usually labelled A and B. As the wheel turns, each channel opens and closes in turn. Because the two channels are offset from one another, the order in which they change tells the mouse which way the wheel is rotating — A leading B means one direction, B leading A means the other. This offset arrangement is called quadrature encoding, and it is the same principle used in everything from printer rollers to industrial motor controllers.
The physical detent — the click you feel — is separate. It comes from a sprung arm riding over the teeth of the drum. On a free-spin wheel that arm is disengaged, which is why the wheel coasts and why the delta stream looks so different.
| What fails | What you feel | What this test shows |
|---|---|---|
| Contamination on one channel | Occasional jump in the wrong direction | Reversals climbing during one-direction scrolling |
| Worn contacts on one channel | Scrolling one way works, the other stutters | Up and down counts badly unbalanced |
| Both channels degraded | Scrolling feels vague, sometimes does nothing | Low tick count relative to how far you turned the wheel |
| Detent spring worn or unseated | Notches feel mushy or absent | Step consistency drops while ticks still register |
| Wheel axle out of its cradle | Wheel wobbles, scroll intermittent | Erratic mix of missed ticks and reversals |
The reason a failing wheel misreads direction rather than simply stopping is the quadrature scheme itself. If one channel drops a transition, the mouse briefly sees the pattern that means “the other way”, and reports a tick in the wrong direction. That is precisely the event this test counts as a reversal, and it is why direction errors — rather than dead scrolling — are the classic symptom of a worn wheel.
Why encoder faults show up at speed
A wheel that behaves perfectly in a document can misbehave badly in a game or a long web page, and the reason is mechanical rather than mysterious.
- Contact time shortens. Turning faster gives each contact less time to settle. A slightly dirty contact that closes reliably at a slow roll may not at speed.
- Debounce windows get crossed. Mice apply a small debounce delay to wheel channels. Fast rotation can push genuine transitions inside that window, so they are discarded as noise — the skipped-notch symptom.
- Momentum on free-spin wheels. A coasting wheel can outrun the sampling rate entirely, producing bursts and gaps that look like faults but are not.
This is why the instruction above is to scroll normally first and then flick. The slow pass establishes your baseline; the fast pass is what surfaces the fault.
Fixing a wheel that jumps or skips
- Confirm it is the mouse, not the software. Test on this page in a second browser. If the reversals only appear in one application, the encoder is fine and something is intercepting scroll input.
- Clean the encoder. Dust and skin oil inside the encoder housing are the most common cause of intermittent reversals. Compressed air first; isopropyl alcohol through the encoder slots if you are willing to open the mouse.
- Check the wheel is seated. A wheel axle sitting slightly out of its cradle produces exactly the partial-detent pattern that shows as low step consistency.
- Try a different port or receiver. Wireless interference can drop individual events, which reads as skipped notches rather than reversals.
- Replace the encoder. They are a few units each and a common repair on otherwise good mice. This is the usual endpoint for a wheel that has been jumping for months.
What a browser cannot tell you about your wheel
Being straight about the limits, because several pages in this category are not.
- The physical detent count. A browser sees scroll events, not the encoder. Detents per rotation is a hardware specification you have to look up.
- Whether smooth scrolling is on. The interpolated deltas the OS produces are indistinguishable from a genuinely free-spinning wheel at this level.
- Polling rate of the wheel channel. Wheel events are delivered on the OS event schedule, not the raw USB report rate. The polling rate test estimates the device rate with the same caveat.
- Which physical encoder is at fault on a mouse with more than one wheel or a tilt mechanism.
FAQ
How do I test my mouse scroll wheel?
Put the pointer inside the box above and roll the wheel up and down at a normal pace, then flick it quickly. Watch the up and down counts rise evenly and the reversal count stay at zero. Reversals appearing while you scroll in a single direction are the sign of a worn encoder.
Why does my scroll wheel jump back up when I scroll down?
The rotary encoder is misreading its direction, usually because of dust or wear inside the encoder housing. This test counts those events as reversals. Cleaning fixes many cases; replacing the encoder fixes nearly all of them.
Why is my step consistency low even though scrolling feels fine?
Most likely operating-system smooth scrolling, which turns each detent into a stream of small interpolated movements, or a free-spin wheel that genuinely has no detents. Neither is a fault. Turn smooth scrolling off and retest if you want a clean reading.
Why do the tilt left and tilt right counts stay at zero?
Because most mice have no tilt wheel. Only models with horizontal scroll report the deltas this test reads for those counters, so zero is normal unless you know your mouse tilts.
Does this scroll test work on a laptop trackpad?
Yes, and two-finger scrolling registers. Expect low step consistency and a very high tick rate, because a trackpad produces continuous pixel movement rather than discrete notches. That is not a fault reading.
Can a scroll wheel be repaired?
Often, yes. Cleaning the encoder with compressed air or isopropyl alcohol resolves a large share of intermittent jumping. If that fails, the encoder itself is an inexpensive part and is a routine replacement on mice worth keeping.
How is this different from a mouse click test?
This checks the wheel’s rotary encoder. A mouse click test checks the button switches, which fail in a different way — one press registering as two rather than a direction being misread.
Sources
- MDN Web Docs — WheelEvent.deltaMode. The pixel, line and page units explained above, and why scroll amounts differ between browsers.
- MDN Web Docs — WheelEvent.deltaY. The vertical scroll value this test reads to count ticks and measure step consistency.
- MDN Web Docs — WheelEvent.deltaX. The horizontal value behind the tilt-wheel counters.
About · Contact · Privacy Policy · Affiliate Disclosure
Related tests and guides
- Mouse Click Test — check your buttons register cleanly
- Mouse Polling Rate Test — how often your mouse reports to the PC
- Double Click Test — the other classic switch fault
- CPS Test — clicks per second for left, right, both buttons or spacebar
- Click Speed Test — the main test, 1 to 60 seconds
- Reaction Time Test — measure your reflexes in milliseconds
- Keyboard Test — chatter, rollover and auto-repeat diagnosis
- Mouse Tester — full button, side-button and scroll diagnostic