What is a good reaction time, really?

The short answer is that a median around 250 milliseconds is unremarkable for an adult on a visual click test, most people fall between 200 and 300, and under 200 is genuinely fast.

The longer answer is more useful, because a large and variable part of that number has nothing to do with you.

What is actually being timed

A visual reaction time is usually broken into three parts, and only the middle one is what people mean by “reflexes”.

Perception. Light reaches your retina and the change is registered. This is fast, but it is not instant, and it depends on how large and how high-contrast the change is. A full screen turning green is detected faster than a small dot appearing.

Decision. Recognising that this is the signal and selecting the response. On a simple test with one signal and one response this is minimal. Add a second signal — respond to green but not to red — and it grows substantially, which is why choice reaction times run 100 milliseconds or more slower than simple ones.

Motor execution. The signal travels to your hand and the muscle moves. Nerve conduction alone accounts for a meaningful slice of the total, and it is one of the parts that slows gradually with age.

A simple click test like ours measures all three together and cannot separate them. That is fine for tracking yourself over time and useless for drawing conclusions about which part is slow.

How much of your score is your equipment

More than most people expect, and this is the single biggest caveat on any web-based measurement.

Added together, a laptop with a 60Hz panel and a Bluetooth mouse can contribute 30 to 40 milliseconds before your nervous system is involved. That is most of the distance between “average” and “fast”, which means a table of global averages is comparing hardware as much as people.

It also means one specific thing: a phone result and a desktop result are not comparable, and averaging them together produces a number that describes neither.

What genuinely moves it

Sleep, by a wide margin. Sleep restriction slows reaction time reliably and it does so before people notice any subjective effect — which is the dangerous part, because it also degrades your judgement of your own state. The difference between a rested and a sleep-deprived result can dwarf every other factor on this page.

Time of day. Most people are measurably slower in the early afternoon and shortly after waking. If you want a fair comparison across a fortnight, take the test at roughly the same time each day.

Caffeine, modestly, and mostly by offsetting tiredness rather than by lifting a rested person above baseline.

Practice, for the first few dozen attempts, and then it plateaus quickly. You are learning the task — where to look, when to expect the change, how to hold the mouse — rather than rebuilding anything underlying. That early improvement is real but it is not a change in your reflexes.

Age, gradually. Reaction time is typically fastest in the early twenties and slows slowly after that. Across a decade the effect is smaller than one bad night’s sleep, which is worth keeping in proportion when a test result looks disappointing.

Why the median beats the average

A five-round test where four rounds land near 250ms and one lands at 600ms because something distracted you has an average of 320. It has a median of 250.

The 600 is not a reaction time. It is a lapse of attention, and including it produces a number that describes neither your reflexes nor anything else in particular. Every result on our reaction time test reports the median first for this reason, with the mean shown alongside it rather than instead.

The same logic applies to early clicks. An anticipated click can come in at 80 milliseconds, which is faster than the signal can physically travel, so it is not a measurement at all. It is counted separately and the round is repeated, rather than being averaged in.

Simple, choice, and why the distinction matters

The test on this site is a simple reaction task: one signal, one response. There is nothing to decide once the panel turns green.

A choice reaction task adds a decision — respond to green but not to red, or press a different key depending on which shape appears. Choice times run substantially slower, typically 100 milliseconds or more, and the gap grows with the number of alternatives.

This is worth knowing for two reasons. It explains why a driving-reaction figure quoted at around 1.5 seconds and a browser result of 250 milliseconds are not contradictory: the first includes perceiving a hazard, deciding what to do and moving a foot between pedals, and the second includes none of that.

It also means a fast simple reaction time predicts less than people assume about real performance, because almost nothing in life is a simple reaction task.

Reading your own history properly

Three habits make a browser test genuinely useful rather than merely diverting.

Take it at the same time of day. Most people are measurably slower in the early afternoon and shortly after waking, so a 9am result against a 3pm one confounds time of day with everything else you might be trying to see.

Ignore your first few sessions. Early improvement is real but it is task learning — where to look, when to expect the change, how the mouse feels — and it plateaus within a few dozen attempts. The interesting variation starts after that.

Watch the spread, not just the median. Tiredness tends to show up first as occasional very slow rounds rather than as a uniformly slower average. If your median is stable but your worst rounds are getting worse, that is the signal.

What a low score does not mean

It does not mean anything is wrong with you. Reaction time is used in clinical assessment, in sports science and in occupational testing, but those settings use calibrated hardware, controlled lighting, standardised instructions and normative data collected on comparable equipment. A browser tab has none of that.

If you take the test once, on a laptop, at eleven at night, and score 340 milliseconds, the only supportable conclusion is that you scored 340 milliseconds on a laptop at eleven at night.

How to use it properly

Take it at the same time of day on the same device, and look at the trend across a fortnight rather than at any single result. That comparison — you against yourself, on unchanged equipment — is the only one a web test can support, and it is genuinely informative: the effect of a bad night’s sleep is usually large enough to see in your own history.

The daily round exists for exactly this, combining reaction time with seven other tests into one score a day so there is a single line to watch rather than eight.