What is a good reaction time?
On a reaction time test, most adults respond to a light in a little under a quarter of a second. Across 1,469 adults aged 18 to 65 measured by Woods and colleagues in 2015, the average was 231 ms, and once the lab’s own screen and mouse were taken out, about 213 ms of it was eye, brain and finger. A phone or laptop reads higher: browsers add a median 67 ms of delay across the ten device-and-browser combinations Pronk and colleagues measured in 2020, so the middle adult would average about 280 ms here.
This test is a simple visual reaction time task, laid out like motor-racing start lights. Five red lights come on one by one, hold for a random 1.0 to 1.8 seconds — the same five holds Woods used — then go out as a green bar lights, and you tap anywhere. There are three counted starts and no practice. A press before lights out, or within 110 ms of it, is a jump start and that start is run again; nothing slower than a second counts, the study’s own window. Your score is the mean of the three, ranked against how three-start averages spread across Woods’s adults, or one age group of them, with the population moved 49.5 ms for typical device delay.
It is not a medical or driving assessment. Three starts give a rough place, not an exact one: the same person’s average moves by about ±23 ms from one run to the next, so someone truly in the middle can land anywhere from faster than 14% to faster than 87% of people. Sleep, caffeine, attention, the hour of the day and your device all move the number; in Pronk’s measurements an iPhone added the least delay and a Mac in Safari the most.
What the test measures
One thing: how quickly you respond to a light you can’t predict.
The task
- Simple visual reaction time, start-light style: five red lights, a random hold of 1.0–1.8 s, then lights out and a green bar
- Three counted starts with no practice; tap anywhere on the screen, click, or press any key
- A jump start (before lights out or within 110 ms) or a lapse (over 1,000 ms) doesn’t count, and that start is rerun
- Timed from the frame that draws the green to the input event’s own timestamp, not by a timer
Your result
- Your average reaction time over three starts, with your median and your best
- Where you stand among Woods et al.’s 1,469 adults aged 18–65, or against one of seven age groups from 18–24 to 59–65
- A verdict after every start, and how much your three times varied
- Everyday terms: how far a car travels before you move, and how far a dropped ruler falls before you pinch it
- A shareable card and a link that challenges a friend to beat your time
Before you start
- About a minute; free, no sign-up
- A finger, a mouse or a keyboard all work; sound is optional, and the go signal is silent
- Stay on the page: leaving in the middle of a start voids that start
- Your times are worked out in your browser; a challenge link carries your average after the # in the address
Reaction time FAQ
Is this test free? Do I need to sign up to see my result?
Yes, it’s free, and there’s nothing to sign up for: no account, no email, no payment. Your result appears on this page the moment you finish, and you can share it or play again straight away.
What is the average human reaction time?
About 230 ms to a light in a lab. In Woods and colleagues’ 2015 study of 1,469 adults, the mean simple reaction time was 230.8 ms, and 17.8 ms of that was the study’s own display and mouse, leaving about 213 ms for the eye, brain and finger. On a phone or in a browser the same reaction reads tens of milliseconds slower because of device delay.
How accurate is an online reaction time test?
Accurate enough for a rough place, not to the millisecond. A browser never sees the real light or the real touch: Pronk et al. (2020) measured 57.6–132.9 ms of extra delay across ten phone and laptop browsers. This reaction time test corrects with their median, 67.3 ms, so your device may be tens of milliseconds kinder or harsher, and three starts wobble by about ±23 ms between runs.
Which device gives the fastest reaction time?
In Pronk et al.’s 2020 measurements, Safari on an iPhone added the least delay (57.6 ms) and Safari on a Mac the most (132.9 ms); Android and Windows browsers sat between about 62 and 70 ms. Those gaps come from hardware and software, not reflexes, so compare scores taken on the same device.
Does reaction time slow with age?
Only a little. In Woods et al.’s sample the average went from 217.9 ms at ages 18–24 to 239.1 ms at 59–65, roughly half a millisecond a year. One person’s single presses vary by about ±40 ms, more than that whole age gap, which is why one start says little and an average says more.
What counts as a jump start in a reaction test?
A press before the lights go out, or within 110 ms after. Nobody can see a light and move a finger that fast, so such a press is a guess rather than a reaction; Woods et al. used the same 110 ms cut-off, and about 3.2% of presses in their study were early. A jump start here doesn’t count and that start is run again.
Is a reaction time test a medical or driving test?
No. This test measures one simple reaction to a light on a screen and gives a rough place among adults. Sleep, caffeine, attention and the time of day all move the number, and a one-minute run can’t tell you anything about a medical condition or about whether you are fit to drive.
Where the numbers come from
Rankings compare your three-start average with a log-normal model of three-start averages built from Woods et al.’s means and within-person spread, shifted 49.5 ms for typical browser delay. The site does not publish its own players’ data.
- Woods, Wyma, Yund, Herron & Reed (2015), Frontiers in Human Neuroscience 9:131 — Factors influencing the latency of simple reaction time
- Pronk, Wiers, Molenkamp & Murre (2020), Behavior Research Methods 52:1371–1382 — Mental chronometry in the pocket? Timing accuracy of web applications on touchscreen and keyboard devices
How the score is worked out
Scoring window
Only presses 110–1,000 ms after lights out count, as in Woods et al. (2015). Faster is anticipation, slower is a lapse.
Three-start spread
A mean of three starts is noisier than Woods’s 120-trial means, so you are ranked against how three-start means spread across people (about ±35 ms), not 120-trial ones. A Monte Carlo check of 400,000 simulated people put the formula within 1.6 percentage points of the truth.
Device shift
Pronk et al. (2020) found browsers overstate reaction time by 57.6–132.9 ms, median 67.3. Woods’s figures already include 17.8 ms of their own hardware, so the population is moved by the difference, 49.5 ms.