Rules & quick reference
| Session | 5 valid rounds |
|---|---|
| Random wait | 1,500–4,500 ms |
| Early presses | Retry; excluded from the mean |
| Result | Mean and fastest valid response; lower is faster |
How to take the reaction test
Start the test and keep your pointer over the large panel. The waiting panel is red and the cue panel is green. Respond as soon as you see the green cue. A press during red produces a Too soon message and does not add a result. Select the panel again to retry that round. After each valid response, start the next round when ready; five valid responses finish the session.
What the millisecond result measures
The page schedules the visual cue with requestAnimationFrame and records a monotonic performance timestamp when the cue state changes. A pointerdown handler records your response before a mouse button is released. Space and Enter are also supported while the panel has focus. The difference includes browser and device effects. The timestamp is not a measurement of the instant that the panel physically emitted light, so decimal precision would overstate accuracy.
Use the mean together with the fastest round
The mean includes every valid round, including a distraction or a slow response. The fastest round describes one attempt and can be influenced by anticipation. A result that looks unusually fast is not automatically evidence of a faster nervous system. Start another session if you want a repeat measurement. This page intentionally does not convert a score into an age estimate, medical assessment or global percentile without a defined comparison dataset.
Compare conditions before comparing people
Use the same input device, browser zoom, display and test method when tracking progress. Touch and mouse input follow different paths; switching between them changes the experiment. Background tasks, a busy browser and a display power-saving mode can also change a result. Keep this tab visible: a hidden tab cancels the run rather than letting an interrupted session contribute a misleading score. Take breaks and avoid using a single result as a judgment of gaming ability.
Your result stays on this device
Completed game results are saved in this browser when local storage is available. Bests and recent history describe this device’s saved runs, not a public leaderboard. Changing browsers, clearing site data or using private browsing can remove them. Copy score writes only the visible result summary to your clipboard when you request it; sharing is your choice. Clear local game data removes scores, daily completion markers and the saved 2048 board for this origin. No game result is uploaded, and the games remain usable when persistent storage is unavailable. Use the rule table to check what the active mode measures before comparing another session. A changed duration, input method or grid difficulty can change the task even when the displayed score uses the same unit.
Frequently asked questions
Is this my exact biological reaction time?
No. The result includes browser scheduling, display presentation, input hardware and your response.
Why use pointerdown instead of click?
Pointerdown responds to the initial press. A click event usually follows release and can include time spent holding the button.
Do early clicks count?
No. An early press repeats the current round and is excluded from the five-round mean.
Where is my percentile?
There is no validated reference population for this implementation, so the page does not invent a percentile.
Built by Mihai · Reviewed 2026-10-24 · Read the related technical guide · Testing policy