Lower on the line means slower. Each row is one test: five rounds, grouped from timestamps (a gap over 90 seconds starts a new one, and a set of fewer than five is listed as partial and never plotted). Dates are UTC. Getting familiar with the task lowers your times by itself — a downward line is your test performance, not a measured change in your nerves. How the numbers are computed is written out on our methodology page.
Reaction time test
How fast are your reflexes? Tap to arm this online reaction time test → the panel goes red, keep your finger off → the instant it flashes green, tap. Five rounds, averaged, then scored against published laboratory norms for your age group. False starts and tab-switches void the round.
Tap to start
5 rounds · random 1.5–4.5 second wait each time
—All-time best (this device)
—This round's average
0 / 5Rounds done
Your reaction profile
Band · Finish 5 rounds to get a band
stored in this browser only
- Average
- —
- Median
- —
- Fastest
- —
- Slowest
- —
- Consistency
- —
Against the laboratory reference
- Normal for your group
- —
- Distance from normal
- —
- Estimated percentile
- —
—
—
Estimated, not measured: the Estimated percentile wording rounds to one person in ten, because it is a normal-curve estimate built from that study's mean and spread (±26.8 ms) and placed on a five-round browser score. Consistency is the round-to-round spread of your own five taps; every step, including how much of the delay is the screen rather than you, is written out on our methodology page.
Reference: Woods DL, Wyma JM, Yund EW, Herron TJ, Reed B (2015), Factors influencing the latency of simple reaction time, Frontiers in Human Neuroscience 9:131 — 1,469 adults, 120 trials each, laboratory rig. Ages outside 18–65 are left blank rather than guessed.
Your history on this device
—
What is human reaction time?
Average reaction time is the gap between a stimulus and your response, measured in milliseconds. In the largest recent laboratory study of it — Woods et al. (2015), 1,469 adults aged 18 to 65, 120 trials each on a laboratory rig — the whole-sample average for a simple visual "see green, press" task was 231.0 ms, rising from 217.9 ms in the 18–24 group to 239.1 ms in the 59–65 group.
A browser score sits above those laboratory figures by design, not by error: your screen, its refresh rate and your touch input add a disclosed 20–60 ms between the moment the page paints green and the moment your tap is registered. That is why an ordinary reflex reads in the mid-200s to high-200s here, and why this page compares you with the lab average for your age group instead of with a single internet number.
How this reaction time test works
Tap the panel to arm it. It holds red for a random 1.5 to 4.5 seconds, then flashes green; the clock starts the instant the green frame is painted and stops at your first tap, read out in milliseconds. Five rounds are averaged, because one lucky tap is noise and one tired tap is noise too.
Three kinds of round are thrown out rather than scored: tapping before green (a false start), switching tab or window mid-wait (the wait was never real), and a timer that stalls for seconds because the browser throttled a background tab. A tap that lands late after green is kept — the panel genuinely was on screen that long, so that is a slow reaction, not a broken measurement.
Your result: what it means
Finish five rounds and the box above fills in: the laboratory average for your age group, your five-round average, and the distance between them, plotted on a bar that spans the study's own spread of about ±27 ms around its mean.
Is 250 ms a good reaction time?
On a browser it is a quick hand: take the 20–60 ms of screen, refresh and input path back off and a 250 ms tap lands around 190–230 ms on the laboratory scale, at or below the study's 231.0 ms whole-sample average. What counts as normal, fast or slow for your own number — including 100 ms, 200 ms and 300 ms — is set out on the normal reaction time page.
Average human reaction time by age
These are measured laboratory values, not a survey of what other web tests report. Woods et al. (2015) gave 1,469 adults aged 18–65 one hundred and twenty trials of a simple visual reaction task on a laboratory rig — see Woods DL, Wyma JM, Yund EW, Herron TJ, Reed B (2015), Factors influencing the latency of simple reaction time, Frontiers in Human Neuroscience 9:131.
| Age group | People | Measured average | Typical on a browser |
|---|---|---|---|
| 18–24 | 86 | 217.9 ms | ≈ 258 ms |
| 25–31 | 115 | 221.0 ms | ≈ 276 ms |
| 32–38 | 201 | 224.8 ms | ≈ 265 ms |
| 39–45 | 273 | 227.7 ms | ≈ 268 ms |
| 46–51 | 276 | 233.6 ms | ≈ 274 ms |
| 52–58 | 272 | 236.4 ms | ≈ 276 ms |
| 59–65 | 246 | 239.1 ms | ≈ 279 ms |
Last column is our estimate, not the paper’s: the measured average plus the ~40 ms this test adds on top of a laboratory rig: 10–40 ms of screen, refresh and input path, and the un-practiced early trials that any five-round set still contains. The source prints the sixth group as “51–58”; we show 52–58 so the bands do not overlap. Ages outside 18–65 are left out on purpose — the study did not measure them, and a value copied from a younger band would be a guess.
For the full age picture and every millisecond your device adds, read average reaction time by age and device.
Frequently asked
- What is a good reaction time?
- In the laboratory study cited above, adults averaged 217.9 ms (18–24) to 239.1 ms (59–65). On a browser, add ~40 ms of screen, input and warm-up delay: under 250 ms is good, under 200 ms is uncommon.
- Does this test measure reaction time accurately?
- Timing starts when the browser paints green and stops at your first tap, measured to the millisecond. Refresh rate and touch sampling add a few milliseconds of overhead, so it is fair for comparing runs and similar devices, not for lab work.
- Can you improve your reaction time?
- Somewhat. Sleep, warmth, caffeine and familiarity all move the number. Big improvements in games usually come from prediction rather than from faster raw reflexes.
- Where do your normal reaction time values come from?
- From one peer-reviewed laboratory study, cited on the page: Woods et al. (2015), 1,469 adults aged 18–65, 120 trials each on a laboratory rig. The "typical on a browser" column in the table is our own disclosed estimate of the 20–60 ms your screen and input add — it is not the paper's.
- Are my scores uploaded anywhere?
- Nowhere but your own browser. There is no backend, no account and no analytics on this site — the comparison above is made against published laboratory data on your device, and the privacy policy says exactly that.