Short-term memory
Digit Span Test
Digits appear one per second. Type them back in order — the list grows until you miss twice.
Take the test →about 150 seconds · no sign-upReaction Time Test
Short answer
Most adults land between 190 and 280 milliseconds on this task, and around 230 is an ordinary result. Your browser and screen add tens of milliseconds of their own, so treat the number as “this device, today” and use your own past runs as the real comparison.
These tests are for curiosity and self-comparison. They are not medical instruments, they do not screen for or diagnose any condition, and no result here says anything about your health. If something about your memory or concentration worries you, talk to a doctor rather than to a website.
What this test measures
One signal, one response, nothing to decide. The clock starts when the screen turns green and stops when you click or press a key. Because there is no choice to make, the number lands close to the floor of how fast you can notice something and act.
It says nothing about your attention span, your memory or how clever you are. Those need different tasks — two of them are on this site. The value of this one is that it is narrow and repeatable.
How to take it
Press Start and keep your finger on the mouse button or the space bar. The wait before green is random, between roughly one and three and a half seconds, so there is nothing to anticipate.
Click too early and the attempt simply repeats. The same happens to anything under 100 milliseconds, because that is anticipation rather than reaction and counting it would flatter the score. Neither is a failure, and the counter shows how many you made.
- Mouse, trackpad or keyboard all work. The difference between them is smaller than the difference between two of your own runs.
- Five attempts take about forty seconds.
- Doing a few runs in a row usually shaves a little off the number, then it settles.
How to read your result
Your score is the middle value of five attempts, in milliseconds — lower is faster. The middle value is used instead of the average because one slow attempt, from a sneeze or a notification, would drag an average up and misrepresent the run.
The percentile answers one question: out of a large sample of adults, what share was slower than you. The 70th percentile means seven in ten were slower. It is a comparison, not a grade.
What the percentiles mean
The comparison comes from a study of 1,469 adults aged 18 to 65 doing this same task, where the typical result was 231 milliseconds. That study published an average and a spread rather than a full table, so the rows below are fitted to those two numbers, leaning slightly towards the slow side — real reaction times have a long slow tail and a hard floor near 150 milliseconds.
| Result | Percentile | Meaning |
|---|---|---|
| 175 ms | 99th | 99% of the sample was slower |
| 189 ms | 95th | 95% of the sample was slower |
| 198 ms | 90th | 90% of the sample was slower |
| 212 ms | 75th | 75% of the sample was slower |
| 229 ms | 50th | 50% of the sample was slower |
| 248 ms | 25th | 25% of the sample was slower |
| 266 ms | 10th | 10% of the sample was slower |
| 278 ms | 5th | 5% of the sample was slower |
| 301 ms | 1st | 1% of the sample was slower |
Why a browser cannot be exact about this
Between the moment the page turns the screen green and the moment your eyes see it, the display adds a delay. Your mouse or keyboard adds another before the click gets back to the page. Together that inflates a browser measurement by tens of milliseconds — commonly 10 to 50 on a computer with a monitor, and considerably more on some televisions.
So the absolute number is only meaningful as “this device, this browser, today”. Comparing your own runs with each other is the honest use of this test: the device delay is the same every time, so it cancels out. That is what the history under the test is for.
What the number cannot tell you
One run is a noisy sample. Sleep, coffee, time of day, how warm your hands are and what your laptop is doing in the background all move it, sometimes by more than the gap between two percentiles.
Results also drift slower with age across large samples, so measuring yourself against a mixed-age group is rough by design. A slow result is not a finding about you: this test cannot detect, screen for or rule out anything.
Questions
What is a good reaction time?
In the sample used here, adults aged 18 to 65 averaged 231 milliseconds, with most results between roughly 190 and 280. Anything in that band is unremarkable. Browser scores tend to sit a little above laboratory ones for the same person.
Why is my result slower than on another site?
Sites start and stop their clocks at different points, and some subtract an assumed hardware delay while others do not. Comparisons between sites mean very little. Comparisons between your own runs on one device mean a lot.
Is the keyboard faster than the mouse?
Usually the difference is a few milliseconds either way, well inside the noise of a single run. Pick one and keep using it if you want to compare runs.
Can I get faster with practice?
Scores usually improve over the first few runs, mostly because you learn the task, and then flatten. We make no claim beyond that, and we do not sell training.
Do you store my results?
Your attempts stay in your own browser, so they survive a reload and never leave your device. Nothing is sent to a server and there is no account to create.
Further reading
Where this comes from
- Woods et al., Frontiers in Human Neuroscience, 2015 — typical simple visual reaction time of 231 ms and its spread, from 1,469 adults aged 18 to 65; the percentile rows on this page are fitted to those figures. Checked 4 August 2026.
- Human Benchmark, reaction time test notes — the size of the delay a computer and monitor add to a browser measurement — commonly 10 to 50 ms, and much more on some televisions. Checked 5 August 2026.