Typing Speed Test

What is a typing speed test?

A typing speed test gives you a passage, times how long you take, and reports your speed in words per minute. It is the standard way to measure keyboard fluency, and it turns up in job applications, transcription work and school assessments.

The catch is in the word. Typing tests do not count words. The convention is that one word is five characters, including the space after it, so your WPM is simply everything you typed divided by five and then by the minutes you took. That is why a test can score you before you have finished a sentence, and why two tests with completely different passages can agree on your speed.

This one runs the timer in your browser, gives you a choice between the common-word list nearly every test uses and real prose from twelve public-domain books, and then tells you what your score means for the text you were actually given.

How to use the typing speed test

1. Pick your material and your length. Common words is what other tests give you; real prose is a whole sentence from a novel, with the punctuation and the long words that come with it. Thirty or sixty seconds.
2. Start typing. The timer begins on your first keystroke, not before. Correct characters stay dark, wrong ones are struck through as you go, and you can keep going or fix them — the score counts what ends up on screen.
3. Read the bottom panel, not just the big number. It shows your real words per minute, the characters per word of the text you got, and whether that made your score run high or low against the standard five.

The five-character word is doing more work than people think

The usual complaint about WPM is that it is not really words per minute. That is true, and measuring it turns the complaint into a defence. Your hands move at some character rate and they have no idea what the words are — so if the score counted real words, the same typist would get a different number on every passage. Across twelve books, a typist at a constant 300 characters a minute genuinely produces between 51.4 and 61.4 words a minute, a swing of 19.5% from the text alone. Over 25-word passages it is worse: 47.2 to 66.4 at the fifth and ninety-fifth percentiles, a spread of 41%.

Divide by five instead and the reported score is identical on all of them. The convention is a normaliser, and it is the reason a typing test measures the typist rather than the passage. Anyone proposing to fix WPM by counting real words would be making the score noisier, not more honest.

There is a second finding that surprised us. Drawn evenly, the 200 commonest English words average 4.96 characters including the space — within 1% of the standard's flat 5.0. So on the material typing tests actually use, your reported WPM is very close to your real words per minute. The convention is well calibrated for word lists, which is presumably not a coincidence.

Real prose is where it drifts. Our twelve books average 5.404 characters per word, so a score measured on prose runs about 8% high: a reported 60 WPM is really 55.5 words a minute. Switch this tool between its two modes and the bottom row moves while the big number stays put, which is the whole argument in one control.

What an error costs, and why it is exactly five times

Most tests report two numbers. Gross WPM is everything you typed over five, per minute. Net WPM subtracts one word for every uncorrected error. Because a word is defined as five characters, that penalty has a closed form: net equals gross times one minus five times your error rate.

So an uncorrected error costs five times its share of your keystrokes. Get 2% of characters wrong and you lose 10% of your score. Get 5% wrong and you lose a quarter. At a 20% error rate the formula reaches exactly zero — not because anyone chose that as a floor, but because five times a fifth is one. That is arithmetic falling out of the definition, and it is worth knowing before you decide whether to go back and fix a typo mid-test.

It also means accuracy is worth more than speed at the margin. Typing 10% faster gains you 10%. Cutting a 4% error rate to 2% gains you the same, and it is usually the easier of the two.

Honest limits

A single test of thirty or sixty seconds is a small sample and your score will bounce around by several words a minute between attempts on the same material. Treat one run as an estimate, not a measurement, and take the median of a few if you want a number to quote.

The comparison is scored position by position rather than by a diff, which is what the WPM formulas assume. One consequence is worth knowing: if you insert a character early on, everything after it counts as wrong, because the text no longer lines up. That is standard behaviour for typing tests and it is why they usually want you to correct as you go.

The prose figures come from twelve nineteenth-century books, which is the corpus this site uses throughout. It is real English but it is not your email, your code or your chat messages, all of which have their own character-per-word profile. The tool tells you the figure for the passage you got, so you can see how far your material sits from the standard rather than trusting an average.

Finally, this measures typing, not writing. Copying text you can see is a different task from composing it, and nobody drafts prose at their copy-typing speed.

Why is it free?

Because it costs us nothing to run. The timer, the scoring and the word list all live in your browser, and nothing you type is uploaded, stored or seen by us — there is no server involved to receive it.

So there is no account, no sign-up and no limit on attempts. Run it twenty times if you want to see how much a single test really varies; it is exactly as free the twentieth time as the first.