The faint line is the raw instantaneous speed — the gap between two keystrokes, turned into a rate. It is violently noisy, because one hesitation between two letters is a rate of 12 wpm and a fast digraph is 300. Every typing test you have used smooths this before showing it to you. Here you choose how.
A fitted distribution, not a survey — a lognormal pinned to two widely published anchors: a median around 40 wpm, and roughly the top 1% above 100. Treat the percentile as indicative. It rests on two numbers.
Median time from the previous key to this one, per character. Medians rather than means, so one interruption while you looked away does not brand a letter as slow.
This page has changed in 11 archived releases, but each of those was a site-wide pass, so none is attributable to this tool on its own and none is listed here. That is not a claim that the tool never changed — a release that reworked many pages at once may well have altered this one too. The changelog has them.
Net speed docks a whole word for each uncorrected mistake in a minute. But correcting one is not free either — the backspacing and retyping eat into the gross figure as well. Accuracy is almost always the cheaper improvement, and it is the one that survives contact with real work.
The fastest instant on the chart is noise: one 100 ms gap reads as 120 wpm. The number worth quoting is the one that holds up over a whole passage, which is exactly what the smoothing controls are there to show you — drag the window and watch how much of your peak was never really there.
That table sorts by median delay alone, so an awkward character you hit twice can outrank one you hit forty times. The count printed beside each key is the other half of the story: the key worth drilling is a slow one that appears often, not the slowest one on the list.
Fifteen seconds measures a sprint you can hold through one clean phrase. Two minutes measures the thing you actually type with, including the moment you lose your place and recover. If two results disagree, the longer one is the honest one.
Not an average word — a fixed five characters, spaces included. Every typing speed you have ever been quoted is (characters ÷ 5) ÷ minutes, and that convention exists so that two people typing completely different text can be compared at all.
It is also, for English prose, about 14% generous. A real English word averages around 4.7 letters, and with the space after it that is 5.7 characters. So a 60 wpm result is really about 53 English words a minute. Nobody adjusts for this, because a fixed divisor is the only thing that makes scores comparable — but it is worth knowing that the unit is a convention rather than a measurement.
Switch this test to code or to numbers and watch what happens. The same fingers doing the same work produce a different wpm, because the divisor stopped matching the material. That is the rule of thumb failing in front of you, and it is the reason a typing speed quoted without saying what was typed does not mean very much.
Gross speed counts everything you typed. Net subtracts one word for every uncorrected error, per minute — five mistakes in a minute costs five words. It is a harsh convention and a deliberate one: the alternative, simply not counting wrong characters, flatters people who type fast and badly.
An instantaneous rate from one gap between two keystrokes is almost pure noise. A 100 ms gap is 120 wpm; a 500 ms pause to think is 24. Neither is your typing speed. So every live readout is filtered, and the choice of filter is a real trade you can drive here:
A moving average is smooth and predictable, and it lags by half its window — a 9-keystroke average shows you where you were 4 keystrokes ago. An exponential average weights recent keys more and never fully arrives at a new level, but responds sooner; an alpha of 0.2 has the same average data age as a 9-point window. A median is the only one of the three that a single outlier cannot move at all — one long pause shifts a 9-point mean by a third of the pause and the median by nothing — and it needs only half its window to accept a genuine change of pace.
The percentile comes from a distribution fitted to two published anchor points, not from a survey of anyone, and certainly not from other people's results on this page — nothing you type here is sent anywhere or stored. Copying text, which is what this measures, is substantially faster than composing it, so this is not your writing speed. Nothing here accounts for keyboard, layout, language or whether you can see the keys. And a test this short is a noisy estimate of anything: the variation between two runs of your own is usually larger than the difference between two adjacent bands.