Reaction Time

About the Reaction Time Test

Reaction time is the delay between a stimulus and your response, measured in milliseconds. Click when the red box turns green. Do that five times and you get your average, in milliseconds. That is the whole test — no sign-up, no download, and it runs entirely inside your browser.

How it works

• Click once to arm the test, then wait. The box turns green somewhere between 1.4 and 4 seconds later, and the delay is re-rolled every attempt, so you cannot learn the rhythm and jump early.
• Click the instant it turns green. Click before that and the attempt is discarded and restarted — it costs you nothing but time, which also means you cannot rescue a bad average by bailing out of a slow round.
• After five clean attempts we report the mean of all five. Space works if you prefer the keyboard.

The result is one number: your average reaction time in milliseconds.

What your millisecond number is actually made of

The clock starts the moment our code decides to turn the box green, and stops the moment your browser reports the press. Everything in between lands in your score: the browser painting the new colour, your monitor drawing it, light reaching your eye, your brain recognising it, your finger moving, and your mouse telling the computer about it. Only the middle of that chain is you.

The screen is the biggest slice you don't control. A 60 Hz display only redraws every 16.7 ms, so a green box that is ready in software can sit unshown for most of a frame — about 8 ms on average, up to 16.7 ms when the timing falls badly. A 144 Hz display cuts that worst case to 6.9 ms. Your mouse adds its own: polling at 125 Hz it can hold your click for up to 8 ms before reporting it, where a 1000 Hz mouse reports within roughly 1 ms. A touchscreen is slower still. We subtract none of it, because a web page cannot see your hardware.

That is why the same person scores differently on a phone and on a gaming PC, and why holding your number up against a stranger's mostly compares equipment. The score means something when you keep the equipment still — same screen, same mouse, same browser. Then a change in the number is a change in you.

What counts as a good reaction time

The most useful reference point is Woods and colleagues (2015), who measured simple visual reaction time in 1,469 people and found a mean of 231 ms — about 213 ms once they corrected for the lag of their own equipment. So a healthy adult reacting to a light lands somewhere near 210–230 ms under laboratory conditions. Browser tests, this one included, read slower than that, because the hardware lag described above is still sitting inside the number.

Ignore the 130–150 ms figures attached to famous players. Those come from single clicks on a livestream, not from measurement, and they are the best of many tries. When researchers put esports players through standardised tests they get roughly 220–290 ms — and the more careful studies find no meaningful gap between professionals and trained athletes who don't game at all. One 2023 study measured esports players at 229.5 ms and sports-science students at 229.5 ms, with a p-value of .999. Reaction speed is simply not where elite players are unusual.

The Elite and Pro labels we hand out are our own informal tiers for this game, not scientific thresholds — we picked them, and you should read them as a game mechanic. The live distribution of real scores, with the sample size it is drawn from, is on our stats page.

What actually moves the number

• Sleep, more than anything else. Being awake for 17 to 19 hours degrades performance about as much as a 0.05% blood alcohol level, in a study that compared the two directly.
• Age — but slowly, and later than you would guess. Simple reaction time holds up until roughly 50 and then drifts by about half a millisecond a year. It is the harder, choice-based reactions that begin slipping in your mid-twenties.
• Caffeine, a little. The pooled effect across studies is about g = 0.28: real, small, and not reliable in every trial.
• Your hardware, which is not you but is in your score anyway. If you want the number to fall without getting any faster, a higher-refresh monitor and a 1000 Hz mouse will do it.

Practising this test will make you better at this test. Whether that carries over to anything outside it is a far weaker claim than most reflex sites imply, and we would rather not make it.