Mouse Skip Test

Does your cursor jump or stutter? Press Start to go fullscreen, sweep your mouse all over the screen, and this test flags skips against your own motion baseline — press Esc to stop and see your score out of 100.

How the detection works

The test samples your cursor's position and timestamp on every pointer event, then computes the distance, time gap and speed of each movement segment. Instead of fixed thresholds, it builds an adaptive baseline from your recent motion (the 95th percentile of distance and time gaps), so a fast mover and a slow mover are each judged against their own style.

The two skip rules

A segment is flagged when it's a distance spike — a jump far larger than your recent norm inside a very short time window — or a time gap then jump, where reports stop briefly and the cursor reappears somewhere new. Skips per 1000 px of travel are reported so the rating is normalised by how much you actually moved.

How to run it

Press Start and the test goes fullscreen so the whole screen becomes the capture area — high-rate movement needs the room. After a short countdown, sweep your mouse all over the screen; clean motion draws in blue and flagged skips draw in red. Press Esc to stop, which exits fullscreen and brings you back to your score and full stats.

Getting a clean reading

Move at a relatively constant pace rather than alternating between crawling and flicking, and give the run a second or two to warm up before trusting the score. Because browser events aren't raw sensor data, treat this as a strong indicator, not a hardware verdict — pair it with the polling rate test for the full picture.

Detection method adapted from Alejandro Lopez Osornio's open-source mouse-skip-detector.

FAQ

What is mouse skipping?

A skip (or cursor jump) is when the pointer leaps across the screen instead of moving smoothly — a sign of dropped sensor reports, wireless interference, a dirty sensor, or a surface the mouse can't track. This test watches your motion and flags those jumps.

How does the detector decide something is a skip?

It learns your normal movement from your recent samples (a p95 baseline) and flags a segment when it's a large distance spike in a tiny time window, or a noticeable time gap followed by a position jump. The first ~1.2s of a run is treated as warm-up and isn't judged.

What does the score mean?

It's a 0–100 rating derived from how many skips were flagged and how often they occur per 1000 px travelled. 92+ is excellent, 80+ acceptable, 65+ needs work, below that is poor. Move at a steady rhythm for the most meaningful result.

Can a browser test prove my mouse is faulty?

Not definitively. Browser pointer events sit above the OS and USB stack, so this infers skip behaviour rather than reading raw sensor data. It's great for spotting a pattern; confirm hardware faults with the polling-rate and button tests too.