Reaction Time Test
Wait for the signal, then tap as fast as you can — a pad that lights up, the five red lights of a Formula 1 start that all go out at once, or a beep. 5 rounds, under a minute, free and with no account, or 15 rounds when you want a steadier number. It is the test people mean by reflexes or reaction speed, and it reports the median of your rounds in milliseconds — plus a straight answer about how much of that number is you and how much is your equipment.
Take the test
Tap the pad the moment it lights up. Not before.
Five red lights, one by one. Go the instant they all go out.
Tap on the beep. The pad lights up at the same moment.
Your browser is not letting this page play sound. The pad still lights up on the signal, so the run works — check the tab is not muted, then test the sound again.
READY
0
ms, median
Your median: the slowest round dropped, the middle of the rest. Simple reaction time is how fast you can act on a signal you were already expecting, and published figures for adults usually land somewhere around 200–300 ms — but on a web page your display and input device are inside that number.
Every round was a false start. Wait for the pad to change before tapping, and the run will measure something.
Median — Best —
- Best
- —
- Average
- —
- Valid trials
- —
Your slowest round is dropped and the median of the rest is the headline. When a run ends it is sent to us — its times, the signal, the input device, the language of this page — so the line about where you stand can exist; nothing that identifies you goes with it. Tapping early is a false start (a jump start, in F1 terms) and the round is repeated. Space or Enter works too.
That was a single run. In the app it becomes a short daily session, with a Focus Index to show for it.
Point your camera at the code — the App Store opens.
What a reaction time test measures
The gap between a signal appearing and your finger moving is not one thing. Light has to reach your retina, a signal has to travel, something has to recognize it as the event worth acting on, and a command has to reach your hand.
Simple reaction time — this test — strips out every decision: there is one signal and one response, and you already know what to do. What is left is close to the floor of how fast you can act on the world at all.
That floor is why the measurement matters. Every more interesting task, from choosing between two answers to naming an ink color, is built on top of it — so knowing your baseline is what lets a slower number elsewhere mean something other than “I am slow today”.
Three ways to run it: tap target, F1 start lights or sound
Same stopwatch, same scoring, three signals. Pick whichever you will actually come back to — the trend on one device is the only number here that means anything.
Tap target is the textbook version: the pad lights up after a random 1.5–3.5-second wait and you tap it. One onset, one response. It is the form nearly every published figure was collected in, so it is the one to use if you want your number to sit next to the 200–300 ms you have read about.
F1 start lights borrows the Formula 1 start procedure: 5 red lights come on one every 1 s, hold for a random 0.2–3 seconds, then all go out together — and you go on lights-out. The countdown tells you roughly when, the hold makes sure you do not know exactly when, and reacting to an offset (lights disappearing) is its own small skill. A tap while any light is still on is a jump start: the round is discarded and repeated, the same way the tap mode treats an early tap.
Sound adds a short 1 kHz beep to the signal — about 120 milliseconds of it — after the same random 1.5–3.5-second wait as the tap mode. It is generated in your browser, so nothing is downloaded and there is nothing to switch on. Press Test sound before you start: a muted tab or a browser that refuses to play audio is invisible until the moment something is meant to make a noise, and finding out mid-run costs you a round. If audio cannot start at all, the test says so instead of leaving you waiting.
Two honest caveats about that mode, and they pull in opposite directions. The audio path has its own delay — a Bluetooth headset can add a great deal more than wired headphones or a laptop speaker — and that delay is inside your number exactly the way display latency is in the other two modes. And the pad also lights up at the moment of the beep, so the mode can be run without hearing it at all; if you want a number that is purely a reaction to sound, look away from the screen or close your eyes.
Where it comes from
- 1868
Donders times a thought
Franciscus Donders measures how long a mental step takes by timing two tasks and subtracting one from the other: react to any signal, versus react only to a particular one. The difference is the cost of deciding. It is the founding trick of mental chronometry, and the reason a bare reaction time is the baseline everything else is measured against.
- Since
The most measured number in psychology
Simple reaction time turns up everywhere: sports screening, sleep-deprivation studies, drug trials, ageing research. It is trivial to collect and it moves in response to almost anything that affects the nervous system, which is both its usefulness and its weakness.
- On the web
The hardware joins in
A browser test measures your nervous system plus your screen plus your input device plus whatever the browser is doing. That does not make it worthless — it makes it a measurement of you-on-this-device, comparable to itself and to nothing else.
How to read your result
Three numbers: the one the run reports, the one it deliberately does not, and how much of the run survived. The interesting part is the distance between the first two.
-
Your median
The headline. Your slowest round is dropped, and the middle of what is left is your number — so one lapse of attention cannot decide it, the way it decides an average, and neither can one lucky round, the way it decides a best.
-
Your best round
Kept for the pleasure of it, not as the claim. The fastest of a handful of rounds is also the luckiest of them, and it gets luckier the longer you play — which is exactly why it is no longer the number on top. A best far below your median means attention wandered on the other rounds, not that you are secretly faster.
-
Valid rounds
Out of 5, or out of 15 if you chose the long run. Tapping before the signal — before the pad lights up or the beep sounds, or while any of the F1 lights is still on — is a false start: the round is thrown away and repeated, because guessing the timing is not reacting.
See where everyone else lands: the reaction time statistics →
Reaction speed by age and by activity
The two questions people arrive with — how does mine compare for my age, and are gamers really faster — have honest answers. Neither of them is a number from us.
We publish no norms by age or activity, so everything below is a published laboratory finding, collected on equipment built for the job rather than on whatever phone you are holding. Read it as background, not as a scale to place yourself on.
Age. Simple visual reaction time is at its quickest in early adulthood and slows gradually from there, with the change easier to see past middle age; children are slower than adults and speed up into their late teens. The part usually left out of the tables is that the spread inside any one age group is wider than the gap between neighboring groups — which is why your own earlier runs tell you more than a decade label ever will.
Activity. Athletes and practiced players of fast games do tend to post quicker times than untrained comparison groups, but the differences reported are small — tens of milliseconds — and the direction of cause is not settled, since quick people gravitate toward quick activities as readily as the activities sharpen them. It is also precisely the size of difference that swapping one phone for another can manufacture on its own.
Everything else. Sleep loss, alcohol, illness, medication, time of day and plain boredom all move reaction time, often by more than age or training do. That is the real reason a single number cannot be read as a fact about you: on any given afternoon it is mostly a fact about that afternoon.
What this version cannot tell you
-
Your screen adds time
A 60 Hz display can sit on a finished frame for up to about 17 ms before showing it, and touchscreens add more. Tens of milliseconds of what you measure here belong to the hardware, not to you.
-
Mouse, keyboard and finger differ
The three input paths do not cost the same. Switching device mid-comparison is enough to invent an improvement that never happened.
-
This is simple, not choice, reaction
One signal, one response, nothing to decide. Choice reaction time — where you must pick the right response — is a different and slower measurement, and the two are not interchangeable.
-
It is not a diagnosis
Reaction time shifts with sleep, caffeine, mood, illness and age all at once, which is exactly why a single number cannot tell you which of them caused it.
Questions about reaction time
What is a good reaction time?
Published laboratory figures for adults on a simple visual reaction task usually land somewhere around 200 to 300 ms. On a web page, add whatever your display and input device contribute, which is often tens of milliseconds. That makes a browser number fine for tracking yourself and poor for comparing yourself to anyone else.
Is this a reflex test?
In the everyday sense, yes — it is the thing people mean when they say they want to test their reflexes, and every reaction test online is doing the same job. In the strict sense, no: a reflex is involuntary and runs through the spinal cord without waiting for you, which is why a tapped knee jerks before you have decided anything. What you are measuring here is a voluntary response — see the signal, decide it is the signal, move — and that route is the longer one, which is why reflexes proper are measured in tens of milliseconds and this test in hundreds.
What is the average human reaction time?
There is no single human average, and that is the first honest thing to say about the question: the figure moves with the signal, the response asked for, the equipment and the group measured. Laboratory studies of adults reacting to a simple visual signal do cluster in the region of 200 to 300 ms, and browser tests run higher because the display and the input device add delay of their own. We publish no average of our own: the line under your result places you among runs taken on this page, which is a fact about this site’s visitors and their equipment rather than about people in general. So treat any number quoted as the average, that range included, as one collected under conditions that were probably not yours.
Why is the wait before the signal random?
So you cannot learn the rhythm. With a fixed delay you would start moving before the signal and measure your timing rather than your reaction. In tap and sound modes the wait is a random 1.5–3.5 seconds; in F1 mode the lights come on one every 1 s but then hold for a random 0.2–3 seconds before going out. Tapping during any of those waits is a false start.
What are the F1 start lights, and why do I react to them going out?
It is the Formula 1 start procedure: 5 red lights come on one every 1 s above the grid, stay lit for a random 0.2–3 seconds, and then all go out at once — and the race starts on lights-out, not on any light coming on. Reacting to something disappearing is a slightly different task from reacting to something appearing, which is exactly why it is worth having as a second mode rather than a second page.
How fast do Formula 1 drivers react to the start lights?
Published figures for drivers reacting to the start lights going out sit around two tenths of a second, with the very quickest starts a little under that. Two cautions come with the number. The sport does not measure it the way this page does: a Formula 1 jump start is not a reaction quicker than some threshold, it is any forward movement in the grid slot before the lights go out at all, picked up by transponders under the tarmac — the penalty is for anticipating, not for being fast. And a driver's figure is measured through a car, on one heavily rehearsed response, to a signal whose arrival window they know far better than you know this page's. Same idea, different instruments; neither number belongs on the other's scale.
Is there a reaction time test with sound?
Yes — the third signal is a 1 kHz beep about 120 ms long, generated by your browser after the same random 1.5–3.5-second wait as the tap mode. Nothing is downloaded. Use the "Test sound" button first, because a muted tab or a browser blocking audio is invisible until the signal is due. Note two things about the number it gives you: your audio hardware adds its own delay, so headphones and speakers do not score the same, and the pad lights up at the same moment as the beep so that the mode is usable without sound at all. Look away from the screen if you want a purely auditory reaction.
What counts as a jump start?
Any tap before all 5 lights have gone out — the second before the first light comes on included. The round is discarded, “JUMP START” shows briefly, and the same round is run again — up to 3 times before the run ends on its own, so you cannot keep guessing the hold until you get lucky. The same rule applies to tapping the pad before it lights up in tap mode.
Can I use the spacebar?
Yes. While a reaction time run is live, Space or Enter counts as a tap — the keypress is read the moment the key goes down, not when it comes back up, so a keyboard is as honest an input as a tap on the screen. Mouse and touch are read on press for the same reason.
Why is my score the median and not my best round?
Because your best reaction time is the luckiest round you had, and a luckiest round is not a measurement of you: it improves simply by playing longer, and two honest runs on the same afternoon can differ by more than the gap between a quick person and an average one. An average has the opposite fault — one lapse of attention drags it down and it says more about that moment than about you. The median sits between those failures: your slowest round is dropped, the middle of the rest is reported, and both the best and the average are still on the card so you can see how far apart they are.
Should I take 5 rounds or 15?
Take 5 rounds if you want a reaction time now — that is why the short run is the one the card starts on. Take 15 if you want one that holds still: attempt-to-attempt spread on this task runs to a few tens of milliseconds, and averaging that away is what more rounds buy you. The measurement is identical either way; only the steadiness of the summary changes. If you compare two of your own runs, compare runs of the same length — a median of 5 and a median of 15 are not estimates of the same precision.
Can I train my reaction time?
Reaction time improves a little, and mostly early on: some of what looks like improvement is you learning the task and settling into a rhythm of expectation. Sleep, caffeine and simple alertness move the number more than practice does. Anyone promising you a dramatic, lasting drop is selling something.
What is the difference between simple and choice reaction time?
This page measures simple reaction time: one signal, one response, and you know in advance what to do, so nothing has to be decided. Choice reaction time gives you several possible signals and a different response for each, and the extra step of picking the right one adds time — reliably enough that Donders built the whole method of mental chronometry on subtracting one from the other in 1868. Choice times run longer than simple ones and the two are not interchangeable; if you have read a figure somewhere, check which of the two it was.
Why is my time here different from other reaction time sites?
Mostly because the hardware and the software around a browser reaction time measurement differ, not because one site is right and another wrong. A test that starts its clock when it asks the browser to paint and one that starts it when the frame is actually shown will disagree by a frame or more; display refresh rate, touch panel sampling and whatever else the browser is doing add the rest. That applies to every browser test including the well-known benchmark sites and including this one, which is why we ask you to compare your runs to each other on one device instead of to a stranger on another.
Do I need an account?
No, and there is nothing to sign up for. The reaction time test itself runs in your browser; when a run ends, the run is sent so that the line comparing it to everyone else’s can exist. Nothing that identifies you goes with it and nothing about you outlives the tab.
If it is the trend you want, that needs an app
A single number tells you about one afternoon. NEURA tracks reaction time with Reaction Tap and processing speed across several more games, rolls them into one Focus Index and shows you the line over weeks — which is the part no single page can do, whatever it measures.
