Reaction Time Tool

Reaction Time Statistics

The whole distribution, not just the average. Find where your number sits, and see plainly where these figures come from.

The distribution

Reaction times are not spread evenly. They pile up between about 200 and 300 milliseconds and then trail off slowly to the right, because there is a hard floor near 100 ms set by nerve conduction but no ceiling at all on being tired or distracted.

100 150 200 250 300 350 400 450 500 median ~250 ms milliseconds
Share of people at each reaction time, derived from the reference distribution this site uses. Faster is to the left.

Percentile table

Read this as: a score of X beats this percentage of people.

AverageBeatsHow it reads
120 ms99.9%Elite
150 ms99.5%Elite
170 ms97.5%Excellent
200 ms86%Fast
230 ms65%Above average
250 ms49%Above average
280 ms30%Average
300 ms20%Below average
340 ms8%Slow
380 ms4%Slow
450 ms1.5%Slow

Look up your own score

Type any time in milliseconds to see where it falls.

Your saved scores

Every personal best this site has stored in your browser. It is read from your own device and has never been sent anywhere. An empty row means you have not taken that test yet.

Loading your scores.

Where these numbers come from

The curve is a reference distribution for a simple visual reaction with a button or mouse press, assembled from published population figures: a median in the mid 250s, a floor near 100 ms set by nerve conduction and muscle latency rather than by effort, and a long right tail. It is the same table the tests on this site use to work out a percentile, and it lives in one file so the number you see on a result screen and the number you see here cannot drift apart.

Two caveats travel with every figure on this page. The first is equipment: a browser measurement includes your monitor, your input device and your operating system, which is commonly 10 to 40 ms before your nervous system is involved at all. The second is that a percentile is a comparison against a population, not a verdict about you. The average reaction time page goes through both properly, and reaction time by age covers how the curve shifts across a lifetime.

Why we do not publish visitor data

We could label this chart as our own data and almost nobody would check. We have not, because it would not mean what it appears to mean. A distribution built from site traffic is shaped by how many people mistapped on a phone, how many were testing the page rather than their reflexes, and how many retried until they got a number they liked.

Nothing you do here is uploaded, so there is no visitor database to publish even if we wanted one. Your scores stay in your browser, which is also why the table above can only show what this device has seen. It is the same reason there is no leaderboard: on a test this easy to fake, a public ranking would measure persistence rather than reflexes.

Ready to add a number to it? The reaction time test takes about twenty seconds.

All the tests

Reflex and speed

Memory and attention

Reading

Frequently Asked Questions

What percentile is my reaction time?
Use the lookup on this page. As a rough guide, 250 ms is around the middle, 200 ms beats about 86 percent of people, and 170 ms beats about 97 percent. Those figures come from a reference distribution assembled from published population data, not from this site's visitors.
What is the median reaction time?
Close to 250 ms on the reference distribution used here. Large public tests often report a median nearer 270 to 280 ms, and the gap is mostly equipment and self selection rather than a disagreement about human biology.
Why is your median faster than Human Benchmark's 273 ms?
Because they are different statistics. Theirs is the median of everybody who has taken their test, which includes every mistimed attempt, every phone and every distracted round. Ours is a reference distribution for a simple visual reaction, which is the figure the research literature reports. Neither is wrong, they just answer different questions.
Is this data from real users of this site?
No, and we would rather say so plainly. Publishing a distribution built from our own traffic would mean publishing a number shaped by how many people mistap on a phone. The curve here comes from published population figures and the page says so wherever a percentile appears.
How much of my score is my computer?
Between 10 and 40 milliseconds for most setups. A 60 Hz monitor alone can hold the change back by up to 16 ms, a wireless mouse adds a few more, and a loaded browser adds a little again. That overhead is why comparing against your own previous scores on the same machine is worth more than comparing against a stranger.