How it works
Click the centre target to start. One target appears at a time at a bounded distance from the last one. Each hit spawns the next target immediately. When the timer ends you see your hits, accuracy, time to acquire each target, and a replay of your cursor path.
The standard run is 30 seconds with a medium target. Shorter and longer runs and other target sizes are separate presets with separate personal bests, so a record always compares like with like.
Scoring rules, version 1
- A press inside the active target, boundary included, is a hit.
- A press inside the arena but outside the target is a miss. Misses lock out hits for 150 ms, so rapid clicking around a target cannot beat deliberate aim.
- Presses during the lockout are ignored. They are neither hits nor misses.
- Hits completed is the headline result. Ties break on accuracy, then on median acquisition time. There is no composite score.
- Escape ends a run early. Incomplete runs are shown but never saved as records. A run also ends if the tab loses focus.
- Target positions come from a seed. The same seed and preset produce the same layout on any screen. That reproduces the task, not your hardware or display.
- On touch screens targets are 2.6 times larger, because the arena is small and a finger is wide. Touch runs are recorded as touch and never compete with mouse runs for records.
Challenges and ghosts
After a run, Challenge a friend copies a link that reproduces your exact target layout. Whoever opens it plays the same targets with your pace shown as a ghost cursor. The link carries only when you hit each target, where you missed, and two samples per target to shape the ghost's path. It never carries your full cursor trace. The challenger's score is shown as an unverified comparison.
What the path replay shows
The map draws your cursor path over the arena. Segments within 120 ms of a miss are highlighted so overshoots and corrections stand out. Crosses mark where misses landed. The cursor dot replays the run at four times speed by default; switch to real time or scrub to a moment. With reduced motion enabled the map is static and the scrubber still works.
Movement is sampled from pointer events, using coalesced events where your browser provides them. Samples are kept at most every 8 ms and the path is a browser-observed record, not a polling-rate measurement.
Mouse accuracy versus game sensitivity
This is a cursor task. It measures how precisely you place the operating-system pointer, which is the real input in MOBAs, real-time strategy games, Age of Empires, Old School RuneScape, and most desktop work. It does not convert to a 3D shooter sensitivity and makes no claim about aim in any specific game.
Records and privacy
Personal bests and run history are stored on this browser only. Cursor paths never leave it. Anonymous usage counts such as runs finished and links opened are sent, as described on the about page, and nothing is sent at all if your browser asks not to be tracked. Touch runs are recorded separately from mouse runs. You can clear everything from the progress page.
The suite
Questions
- Is this an aim trainer?
- It is a 2D precision trainer for the pointer. If you want 3D flick and tracking practice with sensitivity matching, a desktop trainer fits better.
- Why no next-target preview?
- Showing the next target changes planning and anticipation. The standard mode hides it so results stay comparable. A lookahead variant may come later as its own mode.
- Does it work on a phone?
- The page works on phones and touch runs are recorded, but they are kept apart from mouse runs and a desktop mouse is recommended.
- Why 150 ms?
- Long enough that spam-clicking loses to accurate aim, short enough that an honest miss costs about one target. The number is part of rules version 1 and will not change silently.