Reaction Time Test

How to take the reaction time test

Reaction time is the delay between something happening and your response to it. This test measures simple reaction time, the fastest kind: one signal, one response, no decision to make.

1. Choose Visual, Sound or Choice, and how many rounds to play. Five is the default.
2. Click, tap or press Space on the box to start a round, then wait while it is red.
3. The moment it turns green, click, tap or press Space as fast as you can.
4. Start the next round the same way. After the last one you get your average, median, best and worst time.

The wait before the green is random, between 1.5 and 5 seconds, so you cannot time it. Pressing early counts as a false start: that round is not counted and you play it again. Rest a finger on the mouse button, the screen or the Space bar before you start, and use the Fullscreen button for a bigger target with fewer distractions. Esc stops a test at any time.

In the Sound mode the box stays dark and you react to a short beep instead. In the Choice mode one of two or four pads lights up and you press its key (F or J, or D, F, J and K) or tap it. If you use a screen reader, the Sound mode works best, because it needs no visual cue.

What is a good reaction time?

The most careful large study of simple reaction time in recent years tested 1,469 adults aged 18 to 65 (Woods and colleagues, 2015). They reacted to a black-and-white target on a 60 Hz screen with a fast gaming mouse, after 20 practice rounds. Their average was 231 ms. The researchers measured their own equipment and found that the screen and mouse added 17.8 ms, so the reaction itself averaged about 213 ms. A second group of 189 people aged 18 to 82 averaged 237.8 ms.

Use these numbers as a rough guide, not a pass mark. Your browser result also contains your own screen, input device and operating system delays, which nobody calibrated, so it usually reads higher than the lab figures, sometimes by a lot. That is why the most meaningful comparison is with yourself: the same device, the same mode, a few tests on different days. The Progress chart above shows exactly that.

Very fast single rounds deserve suspicion rather than pride. A press under 100 ms is counted as a false start, because a real reaction that quick is so rare that it is almost always a lucky guess. For a sense of the extremes: in the 1880s Francis Galton recorded average visual reaction times of 181 to 189 ms in people aged 18 to 60, shorter than any recent large study has found.

Visual, sound and choice reaction time

Reacting to a sound

People usually react faster to a sound than to a light. A study of 120 first-year medical students found auditory reaction time significantly faster than visual reaction time for both men and women. Sprint starts rely on this: the false-start rule in sprinting assumes that a reaction to the gun is rarely faster than 100 ms, although a study of starting blocks found some athletes could react a little quicker.

The Sound mode plays a 200 ms, 1,000 Hz beep through the Web Audio API and works out when it leaves your audio device. Your speakers or headphones add their own delay, and wireless headphones can add far more than wired ones without the browser knowing, so only compare Sound results made with the same audio setup.

Choice reaction time and Hick’s law

A choice reaction needs a decision as well: which of several responses is the right one. It is slower than a simple reaction, and it gets slower as the number of choices grows. In 1952 W. E. Hick showed that the time grows with the logarithm of the number of equally likely choices, now called Hick’s law:

RT = a + b × log₂(n + 1)

n = number of equally likely choices
a = time that does not depend on the choice (seeing, moving)
b = extra time per bit of uncertainty

Hick added 1 to n for the possibility that no signal comes at all; a year later Ray Hyman wrote the same law with log₂ n. Either way, every doubling of the choices adds roughly the same amount of time. The table shows the log term for the modes in this test:

The constants a and b differ from person to person and from task to task, so the tool does not predict your choice time. Run Visual, then Choice with 2 pads, then Choice with 4, and compare your own averages. The pads sit in the same left-to-right order as their keys, a natural match between signal and response; practice and such a match both reduce the extra time that more choices cost. A large survey of 7,130 adults also found that choice reaction time slows throughout adult life, while simple reaction time changes little until about 50.

How this test measures your reaction time

A reaction time test in a browser is easy to build badly. These are the choices that keep this one honest:

• The press, not the release. The test listens for the moment a mouse button, finger or key goes down. A click event only fires after you let go, which would add the time your finger stays down.
• The event’s own timestamp. Browsers stamp each input event with the time it happened, on the same high-resolution clock as the stimulus. The test uses that stamp, so a moment of slowness on the page does not lengthen your time. The standard rounds these clocks to 0.1 ms or coarser, far finer than any device delay.
• The first frame that can show the change. The box changes color inside an animation frame, and the clock starts at the next frame, the earliest moment the green can be on your screen. Starting at the code’s request instead would count a frame of waiting against you.
• The sound at the speaker. In Sound mode the beep is scheduled ahead, and the Web Audio API reports when that sample reaches your audio output, so the clock starts when the sound does.
• An unpredictable wait. Each wait is drawn at random from 1.5 to 5 seconds with the browser’s cryptographic random generator, so you cannot learn the rhythm.

Every press is then scored with the rules of the psychomotor vigilance test, the standard reaction test of sleep research:

The test is also built to be safe for people with photosensitive epilepsy. The box changes color once per round, at least 1.5 seconds apart, and never by itself in a loop, well within the accessibility rule of no more than three flashes in any one second. The change from red to green is also a change in brightness, text and symbol, so it does not depend on telling red from green.

Device lag: what your time includes

No browser can see the moment light leaves your screen or the instant your finger lands. Everything between those moments and the browser’s clock is added to your time. These are the main sources:

In the 2015 study the display and mouse together added 17.8 ms, and the authors note, citing earlier work, that hardware and software delays can inflate measured reaction times by up to 100 ms. Nobody has measured your own setup that way, which is why a browser test can never be a lab test: treat its numbers as yours, not as lab values.

The screen refresh is the one delay you can easily see. A change to the screen has to wait for the next refresh, so on average it waits about half a frame:

The test estimates your screen’s refresh rate from the browser’s animation frames and shows it under your history. Some phones and laptops lower the refresh rate to save battery, which is one reason a phone result can be slower than the same person’s desktop result.

Average, median, best and standard deviation

One round tells you little: your times naturally vary by tens of milliseconds from press to press. That is why the test runs several rounds and summarizes them. Here is a five-round test worked through, with times of 268, 241, 255, 312 and 249 ms:

standard deviation = √( Σ (time − average)² ÷ (rounds − 1) )

The deviations from the 265 ms average are 3, −24, −10, 47 and −16 ms; their squares add up to 3,150. Dividing by one less than the number of rounds is the usual sample formula, and it gives about 28 ms. A small standard deviation means you are consistent; a large one usually means a few rounds where your attention slipped.

The median is often the better headline. One slow round, like the 312 ms above, pulls the average up but hardly moves the median. The Distribution chart groups every saved time for the current mode into 25 ms steps, so you can see where most of your reactions land and how often you are slow.

What affects reaction time

• Age. In the 2015 study, simple reaction time rose by about 0.55 ms per year, roughly 5.5 ms per decade, and the oldest group (59 to 65) was 21.2 ms slower than the youngest (18 to 24). A survey of 7,130 adults found simple reaction time changes little until around 50, while choice reaction time slows throughout adulthood.
• Sleep. A review of 70 studies of short-term sleep loss found its largest effect on simple attention, above all on lapses: moments when a response comes very late or not at all. Simple reaction tests are a standard tool in sleep research for exactly this reason.
• Fatigue during the test. Over 120 rounds, the 2015 study’s participants slowed from 228.1 ms in the first 20 rounds to 237.2 ms in the last 20. Short tests avoid this.
• The wait before the signal. In the 2015 study, reactions were about 15% slower when a signal came only 1 second after the one before than after longer gaps. Here each wait is random, from 1.5 to 5 seconds, and you start every round yourself.
• The kind of signal. Reactions to sound are usually faster than to light, and decisions between several responses are slower (see Hick’s law above).
• Physical activity. In the study of 120 medical students, those who exercised regularly reacted faster than those who did not. That shows a link, not proof that exercise caused it.
• Sex. Findings are mixed. The 2015 study found no difference between men and women, while the student study and the survey of 7,130 adults did find differences, especially in how variable choice reaction times were.

Not medical advice. This is a self-test for curiosity and practice, not a medical or fitness-to-drive assessment. A slow or variable result in a browser can come from your device, your setup or a moment of distraction, and it does not diagnose anything. If you notice a real, lasting change in your reactions, talk to a doctor.

How to improve your reaction time

• Sleep enough. In the review of sleep-loss studies, simple attention was the skill hit hardest, so a good night is the most reliable way to a faster result.
• Warm up first. Lab tests give practice rounds before the real ones; the 2015 study gave 20. Run a short test to warm up, then the one you count.
• Keep it short. Five or ten rounds at a time avoids the slowdown that comes with fatigue.
• Stay active. Regular exercise is linked with faster reactions, and it is good for you anyway.
• Practice choice tasks. With practice, the extra time that each added choice costs gets smaller.
• Remove device lag. A wired mouse or keyboard, a high refresh rate, wired headphones, fullscreen mode and closing heavy tabs all make your result closer to your real reaction. They improve the measurement, not you, so note your setup when you compare.

Private by design

• Nothing is uploaded. The page’s Content-Security-Policy only allows connections back to this site, and the test makes none. Open your browser’s Network panel and play a test: no request appears.
• Your history stays on your device. Settings and finished tests are kept in your browser’s local storage under one key, up to the last 100 tests. Download them as a CSV file, or clear them with the Clear history button.
• No account and no cookies. Once the page has loaded the test even works offline.

Sources

• 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 (full text on PubMed Central): averages, the age table, hardware delays, fatigue and waiting-time effects.
• Der G, Deary IJ (2006). Age and sex differences in reaction time in adulthood: results from the United Kingdom Health and Lifestyle Survey. Psychology and Aging 21(1):62–73 (PubMed).
• Jain A, Bansal R, Kumar A, Singh KD (2015). A comparative study of visual and auditory reaction times on the basis of gender and physical activity levels of medical first year students. International Journal of Applied and Basic Medical Research 5(2):124–127 (PubMed Central).
• Pain MTG, Hibbs A (2007). Sprint starts and the minimum auditory reaction time. Journal of Sports Sciences 25(1):79–86 (PubMed).
• Basner M, Dinges DF (2011). Maximizing sensitivity of the psychomotor vigilance test (PVT) to sleep loss. Sleep 34(5):581–591 (PubMed Central): the 100 ms false-start and 500 ms lapse rules.
• Lim J, Dinges DF (2010). A meta-analysis of the impact of short-term sleep deprivation on cognitive variables. Psychological Bulletin 136(3):375–389 (PubMed Central).
• Hick WE (1952). On the rate of gain of information. Quarterly Journal of Experimental Psychology 4(1):11–26.
• Hyman R (1953). Stimulus information as a determinant of reaction time. Journal of Experimental Psychology 45(3):188–196.
• Proctor RW, Schneider DW (2018). Hick’s law for choice reaction time: A review. Quarterly Journal of Experimental Psychology 71(6):1281–1299 (PubMed).
• W3C, High Resolution Time (Working Draft, 1 September 2026): the clock behind event timestamps, and its 0.1 ms resolution floor.
• W3C, Web Audio API 1.1 (Working Draft, 22 September 2026): getOutputTimestamp() and outputLatency.
• W3C, Web Content Accessibility Guidelines (WCAG) 2.2, Success Criterion 2.3.1 Three Flashes or Below Threshold.
• MDN Web Docs: Event.timeStamp, pointerdown and requestAnimationFrame().

Reviewed by microplugins · Last updated October 2026