How it works
Click the panel (or press Space or Enter) to start. After a random wait between 1.5 and 4.5 seconds the panel turns green. Press as soon as you see it. You do this five times, and we show the median of the five as your result.
If you press before the panel turns green, that press does not count: the trial is marked as a false start and repeats. Pressing Escape during a run cancels it without saving, and so does switching to another tab.
The start time is recorded in the browser frame callback that turns the panel green, which is the closest point the page can observe to the stimulus being drawn. The end time is the timestamp of your press event.
What the number includes
The result is a browser-observed visual response time. It includes your display's refresh interval and processing delay, your mouse, trackpad, touchscreen or keyboard, the operating system, and the browser. It is more than the time your nerves and muscles take, and we do not claim otherwise.
Other reaction sites often quote around 30 ms of system lag as a typical figure. We do not guess a correction for your setup. Instead, after each run we show a note with your measured display refresh rate and how long one frame lasts at that rate.
Why median
A best-of score rewards guessing: if you anticipate the colour change and happen to hit it, one lucky trial defines your result. The median of five is the middle value, so one unusually fast or slow trial barely moves it and the result is more stable from run to run. We also show your best, mean, the spread between your fastest and slowest trial, and flag any trial under 100 ms as likely anticipation.
Typical results
Medians published by large reaction-time sites are around 270 to 290 ms, and those figures include their own display and input lag. Your number here is not directly comparable to theirs, because the sites measure and time things differently.
Comparisons between people on different hardware are not fair either: a high refresh rate display and a low-latency mouse can shave off tens of milliseconds without any change in the person. Compare yourself against your own earlier runs on the same setup.
Mobile and touch
Touchscreens add their own latency, and browsers handle taps differently from mouse presses. Runs made with touch are recorded separately from mouse runs, and personal bests are only compared within the same input type. The page works on a phone, but a mouse or keyboard on a fast display will give the most repeatable numbers.
FAQ
Is my data uploaded?
No. Results are stored in this browser only, so you can see your personal best. You can clear them from the settings menu.
Why was my run cancelled?
If the tab is hidden during a trial, the timing is no longer reliable, so the run is cancelled and not saved.
Why is my score slower than on other sites?
Different sites time different things and use different colours, waits and averaging. We show the median of five trials, not your best, and we do not subtract any assumed system lag.
Can I use the keyboard?
Yes. Press Space or Enter on the focused panel. Keyboard presses are recorded with an unknown input type, so they are kept apart from mouse and touch personal bests.