Average reaction time

Every run submitted here

One bar per band of equal width, tallest where most people land. The three signals and the two input devices are counted apart, and the chart always says which of the six cuts you are looking at.

Tap target · Mouse or keyboard

—

median

—

10th percentile

—

90th percentile

—

runs in this cut

The F1 start is predictable on purpose, so it measures anticipation. Tap and sound give no warning, so they measure response. Three different quantities — never one chart.

The fastest runs taken here

All-time, one cut at a time, and long runs only. A summary of 5 attempts carries enough luck that the top of a board built from short runs would be sorted by luck rather than by speed — so the long run of 15 is the way onto this table, and the short one stays exactly as welcome in the chart next door.

Tap target · Mouse or keyboard

Go to that cut

There are no names on this board and no box to type one into. Every run gets a short label this site works out for it — that is the Run column — so you can find your own row again. It belongs to the run and not to a person: take the test twice and you get two unrelated labels, and nothing here can turn either of them into a name.

How this is measured

Six decisions stand between a run and a bar on that chart. None of them is neutral, so each one is written down.

Why the median and not the average

One lapse of attention adds a couple of hundred milliseconds to a single attempt and drags an average with it. The median ignores it. That is why your number barely moves between two honest runs, and why an average would have moved a lot — the run that reports it would be reporting your worst moment, not your speed.

Why 5 attempts is a warm-up and 15 is a measurement

The same person varies by roughly 30–50 ms from attempt to attempt, while the whole interesting range between one person and another is about 180–320 ms. The noise is the same size as the difference we are trying to see, so a summary of 5 attempts wobbles enough that two honest runs disagree visibly. Run 15 and it settles. Both lengths are counted here, and they are not the same distribution — a median of a short run is drawn from a wider one.

Why the three signals are three charts

The F1 start counts down to a moment you can see coming: five lights, then darkness. What that measures is anticipation, and it is faster. The tap target and the beep arrive after a wait you cannot predict, so they measure response. Putting all three on one chart would prove only which signal is easiest, and the number underneath it would describe nobody.

Why a finger and a mouse are counted apart

A touchscreen has its own delay before it reports a touch, and so does a mouse, and the two are not the same size. That gap sits between two visitors to this same page — one on a phone, one on a laptop — so splitting by device is the honest cut, and splitting by where the run came from would not be. Which device a run used is read from the events that produced it, never guessed from the screen.

What is thrown away, and what is only flagged

An attempt faster than 120 ms is not a reaction — nerve and muscle cannot do it — so a run containing one is refused outright and stored nowhere. A false start costs you the attempt; it never becomes a fast time. Runs that are merely hard to believe are a different case: they are kept, marked, and left out of every figure on this page, because a metronome and a remarkably steady person look identical from five numbers, and refusing them would only tell whoever wrote the script what to fix.

Why a cut sometimes shows nothing at all

Below a fixed number of runs there is no chart and no figure here — not a rounded one, not a hedged one. A median of nine runs is noise wearing the costume of a measurement, and a thin slice is also where one person’s result stops being anonymous. The page says how many it has instead.