Reaction Time Test — Measure Your Reflexes
Reaction Time Test
Timing uses performance.now, which is monotonic and unaffected by a system clock change. The delay before each round is randomised between 1.2 and 5 seconds because a predictable rhythm lets you anticipate the signal instead of reacting to it. A typical adult result sits between 200 and 300 milliseconds.
What Is
Reaction time is the interval between the arrival of a stimulus and the beginning of a response, and it is one of the most reproducible measurements in experimental psychology precisely because it is so short and so tightly constrained. It is not a test of intelligence, but it is a genuine index of central nervous system processing speed. A simple visual reaction time in a healthy adult typically lands between 200 and 300 milliseconds, and that figure decomposes into identifiable stages: roughly 60 to 100 milliseconds for the signal to travel from the retina to the visual cortex, about 100 milliseconds for cortical processing and the decision that the button should be pressed, and the remaining 60 to 120 milliseconds for the motor command to reach the forearm muscles and for the finger to complete its movement. Because the peripheral component is comparatively slow, no amount of concentration can push a visual reaction time much below the 200 millisecond floor on most hardware. Measuring that floor honestly requires controlling for several confounds, and this tool handles the most important ones.
The first is anticipation. If the wait were a fixed interval, the best strategy would be to click in time with the interval rather than to respond to the signal, and the resulting score would measure nothing but metronome-following. The randomised 1.2 to 5 second delay removes that regularity. The second confound is the false start, which is handled by forfeiting rather than by recording a meaningless short time. The third is measurement precision, handled by performance.now().
Finally, a single trial is worthless. Reaction times are a distribution, not a quantity, and they are right-skewed: most trials cluster near the fast end while a minority run much slower because of a lapse of attention, a blink, or a brief noise in the system. That is why the summary reports the median alongside the mean, since the median is far more resistant to those outliers, and why the standard deviation is reported at all, since a spread of a few milliseconds suggests a steady, well-rested tester while a spread of a hundred suggests fatigue or distraction. A visually guided clicking test like this also belongs to a family of paradigms built on signal detection theory, in which the real question is never the raw speed but where you set your criterion for acting given that you cannot be certain the signal has occurred.
How to Use
- Set the Number of rounds dropdown to 3, 5, 7 or 10 before you start. The control locks as soon as the first result is recorded, so the whole run stays consistent.
- Click the large panel once when it says Click to begin. The panel changes to blue and starts an unpredictable wait of between 1.2 and 5 seconds.
- Sit still and watch the panel rather than the clock. When it flashes green and shows Now!, press it again as fast as you can — that second press is the one being timed.
- If you click before the panel turns green, that round is forfeited. It is logged as an early click and the run continues from where it left off rather than restarting.
- Use the Restart button to clear all results and begin a completely fresh run, or the Next round button to skip a pause and start the next trial immediately.
- Read the per-round Results list as you go to spot a warm-up effect, since your first trial is usually the slowest and the last two or three are usually the fastest.
- Once every round is in, study the Summary table: compare Best against Worst, and use Mean and Median together to see whether an outlier skewed the average.
- Check Standard deviation and Rounds forfeited before celebrating a low mean. A fast mean with a large spread usually means two or three lucky clicks, not a genuinely fast nervous system.
Examples
Input: 5 rounds, all valid: 218, 244, 231, 402, 226 ms
Process:
Sort to 218/226/231/244/402; median = 3rd value; mean = 1321 / 5; sd from squared deviations / 5Result: Best 218 ms, Worst 402 ms, Mean 264 ms, Median 231 ms, SD 69 ms, 0 forfeited
Input: 10 rounds with a warm-up, all valid: 196, 203, 189, 215, 198, 210, 201, 199, 188, 194 ms
Process:
Mean = 1993 / 10 = 199.3; median = 6th of 10 sorted values; sd = sqrt(652.1 / 10)Result: Best 188 ms, Worst 215 ms, Mean 199 ms, Median 199 ms, SD 8 ms, 0 forfeited
Input: 3 rounds where round 2 was clicked before the green signal
Process:
The 0 ms entry is stored as an early click and dropped, leaving 2 valid trials: mean = 514 / 2, median = 2nd of 2 sorted valuesResult: Best 249 ms, Worst 265 ms, Mean 257 ms, Median 265 ms, SD 8 ms, 1 round forfeited
Input: 5 rounds: 233, 0 (too early), 251, 268, 240 ms
Process:
Valid set = 233/240/251/268; mean = 992 / 4 = 248; median = 3rd of 4 sorted values = 251; sd = sqrt(698 / 4)Result: Best 233 ms, Worst 268 ms, Mean 248 ms, Median 251 ms, SD 13 ms, 1 round forfeited
Related Searches
People also search for: reaction time, reaction time test, measure reaction speed, human reaction time, reflex test, simple reaction time.
reaction timereaction time testmeasure reaction speedhuman reaction timereflex testsimple reaction timevisual reaction timems reactionclick speed testreflexesresponse timeneuron reactionpsychomotor speedattention reactionbrain speedfree reaction testhuman benchmarkreaction in milliseconds
Frequently Asked Questions
What is a normal reaction time for a human being?
A simple visual reaction time in a healthy adult normally falls between about 200 and 300 milliseconds. Around 60 to 100 milliseconds of that is signal travel to the visual cortex, roughly 100 milliseconds is cortical processing and the decision to move, and the last 60 to 120 milliseconds is the motor command reaching the forearm and the finger completing the tap. Because the peripheral stages are comparatively slow, the floor is hard to beat no matter how focused you are, and any score far below 180 milliseconds should be treated as an anticipation artefact rather than a superhuman result.
Why does the test wait a random amount of time before turning green?
Because a predictable interval lets you stop reacting and start guessing. If the panel always flashed after exactly two seconds, the optimal strategy would be to count and click on the third beat, and the resulting score would measure your metronome-following ability rather than your reflexes. Randomising the delay between 1.2 and 5 seconds removes the rhythm, so the only winning move is a genuine response to the signal as it appears.
Is a lower score always better, or can being too fast be a problem?
Yes, lower is better only if the response is legitimate. Clicking before the panel turns green does not produce a fast score; it forfeits the round, which is logged as an early click and excluded from the best, worst, mean, median and standard deviation figures. It is counted separately as a forfeited round so that impatience is visible rather than hidden inside an unfairly low average.
Why does the summary report both a mean and a median?
Reaction times are right-skewed, so a small number of unusually slow trials can drag the arithmetic mean noticeably upward. The median, which depends only on the order of the values, stays put. If Mean and Median are close your run was consistent and the mean is trustworthy; if the mean sits well above the median, at least one trial was an outlier and the median is the more honest description of your typical response.
What does the standard deviation tell me?
It describes how spread out your trials were. A standard deviation in the low tens of milliseconds means you were steady and the mean represents you well. A standard deviation approaching a hundred milliseconds means your attention drifted between rounds, which frequently happens when people take more than 10 trials in one sitting. Because it is computed as the population standard deviation, it uses every valid trial rather than a sample correction, which is the right choice for a short personal test.
Does this test need an internet connection or send my results anywhere?
Neither. The whole test is implemented client-side in the browser, and the timings come from the browser's own high-resolution monotonic clock rather than from any server. Nothing is uploaded, no account is required, and no score is persisted. The only caveat that matters is that this measures the speed of your response to a visual cue on a particular display and input device, not the pure neural component of your reaction time.
How can I legitimately get a better score?
Get some sleep, sit in a quiet room, use a mouse rather than a trackpad, and keep your finger resting on the panel so only the final press is timed. Do not lock your wrist rigidly, because that adds a mechanical component to the measurement. Run several sessions of 5 rounds rather than one long session of 10, since attention lapses accumulate and the early rounds of any session act as warm-up.