What "simple reaction time" actually measures
Reaction time research splits into a few distinct tasks, and this tool measures the simplest one. Simple reaction time is the gap between a single, expected stimulus and a single response: one signal, one action, no decision to make. That is different from choice reaction time (respond differently depending on which of several signals appears) and further still from the kind of split-second decisions a driver or athlete makes mid-play, which layer judgment on top of raw detection speed.
That narrower definition is exactly why simple reaction time is useful as a baseline: there is nothing to think about, so the number mostly reflects how fast a signal travels from your eye through your nervous system to your finger. It is also, as the next section explains, a number that depends heavily on exactly how it gets measured.
How the test times each trial
Each trial waits a random 1.2 to 3.5 seconds before the box turns green, so there is nothing to count down and anticipate. Click before the box turns green and it counts as a false start: the trial resets and does not count toward your average, but it is tallied separately so you can see how many times you jumped the gun. Click after the box turns green and the elapsed time is measured with the browser's performance clock (performance.now()), which reports sub-millisecond timestamps, from the moment the color changes to the moment your click registers.
Running several trials instead of one matters because a single click is noisy: attention lapses, a finger already mid-motion, or a moment of hesitation can all swing one result by 50 ms or more. Averaging trials, and separately reporting the best, the median, and the spread (standard deviation) between them, gives a steadier picture than any single click.
Why the "right" number depends on how you measure it
A literature review by Robert Kosinski at Clemson University notes that for roughly 120 years, the textbook figure for simple visual reaction time in college-age adults was about 190 ms, a number tracing back to Francis Galton's studies in the 1800s. The same review cites Eckner et al. (2010), who timed the same NCAA football players two ways: 203 ms on a falling-meter-stick test and 268 ms on a computer test, and notes that computer-measured results at Clemson typically land close to that higher figure. The gap is not really about people getting slower. It is about the test.
Woods et al. (2015) tested a community sample of 1,469 people ages 18 to 65 with a calibrated computer-based test and measured a 231 ms average, or 213 ms once they subtracted the delay their own hardware and software introduced. Their paper's conclusion is direct: processing speed in a modern sample is not slower than a century ago, once you account for the fact that computer equipment, theirs included, adds its own lag on top of a person's true reaction time. This tool runs in a browser, on whatever device you happen to be using, so it inherits that same kind of measurement lag, and then some.
Why your own trials vary, and what to do about it
Fatigue, a distracting notification, or simply not being warmed up can each add tens of milliseconds to a single trial, which is exactly why the tool reports median and best alongside the average: the median is less pulled around by one slow outlier, and the best trial shows what you are capable of when everything lines up. A large gap between your best and worst trial (a high standard deviation) says more about how consistent your attention was during the test than a single average ever could.
Reaction time is also known to change with age. Woods et al. (2015) measured it lengthening by about 0.55 ms per year of age across their 1,469-person sample, and traced most of that increase to slower motor output rather than slower mental processing. This tool has no way to know your age and does not attempt to adjust for it.