How WPM is calculated

Words per minute sounds like it should mean "how many words you typed in a minute". It doesn't quite. In typing tests, a word is a fixed unit of five characters, and the number you see depends on how your mistakes are counted. Once you know the formulas, it is easy to see why your score changes from site to site, and you can check any result yourself with a calculator.

A "word" is five characters

Real words vary a lot in length. "I" is one letter and "responsibility" is fourteen. If a typing test counted actual words, a passage full of short words would produce a much higher score than a passage full of long ones, even for the same typist on the same day.

To get round this, typing tests use a standard word of five characters. Spaces and punctuation count as characters. So does every letter. Ordinary English words average somewhere around four to five letters, which makes five characters, including the space, a convenient stand-in for one word. The convention goes back to typewriter-era testing and almost every typing site still uses it.

That gives us the basic formula:

WPM = (characters typed ÷ 5) ÷ minutes

A quick example. In a 30-second test you type 175 characters, spaces included. That is 175 ÷ 5 = 35 standard words, in half a minute. 35 ÷ 0.5 = 70 WPM.

Notice what this means in practice. At 60 WPM you are typing 300 characters a minute. If the text is full of long words, you will have typed fewer than 60 dictionary words, and that is fine. The unit is characters, dressed up as words.

Gross WPM: everything you typed

Gross WPM, sometimes called raw WPM, counts every character you typed, whether it was right or wrong. It tells you how fast your fingers are moving, but nothing about whether the result is any good.

Gross WPM = (all characters typed ÷ 5) ÷ minutes

Worked example

A 1-minute test. You typed 265 characters in total, including spaces and including some mistakes.

• 265 ÷ 5 = 53 standard words
• 53 ÷ 1 minute = 53 gross WPM

Gross speed is useful as a diagnostic. If your gross WPM is much higher than your net WPM, you are typing faster than you can type accurately, and slowing down a touch will probably raise your real score.

Net WPM: only what you got right

Net WPM is the honest number. It is what employers usually mean when they ask for a typing speed, and it is the main score on most typing sites, including ours. There are two common ways to calculate it.

Method 1: subtract errors per minute

This is the classic textbook formula. Take your gross WPM and subtract one word for every uncorrected error per minute.

Net WPM = gross WPM − (uncorrected errors ÷ minutes)

Using the same 1-minute test as above, with 4 uncorrected errors:

• Gross WPM = 53
• Errors per minute = 4 ÷ 1 = 4
• Net WPM = 53 − 4 = 49 WPM

A longer test works the same way. Say you take a 2-minute test, type 610 characters and leave 8 errors uncorrected:

• Gross WPM = 610 ÷ 5 ÷ 2 = 61
• Errors per minute = 8 ÷ 2 = 4
• Net WPM = 61 − 4 = 57 WPM

Each error costs a full word per minute, which is a stiff penalty. Four mistakes in a minute knock 4 WPM off your score, and that is before counting the time they cost you.

Method 2: count only correct words

Many modern typing sites, including Presspace, Monkeytype, Typeracer and 10FastFingers, use a variation of this approach. Instead of subtracting a penalty, they only give you credit for words you typed correctly.

Net WPM = (characters in correctly typed words + one space per correct word) ÷ 5 ÷ minutes

Worked example. A 1-minute test where you attempted 54 words:

• 50 words were typed correctly. Together they contain 214 letters.
• Add one space for each correct word: 214 + 50 = 264 characters.
• 264 ÷ 5 = 52.8 standard words.
• 52.8 ÷ 1 minute = 52.8, shown as 53 WPM.

The 4 words with mistakes in them earn nothing at all, however close they were. That is why a single typo in a long word hurts more than one in a short word: you lose credit for the whole word.

Both methods give similar results for most people. Method 2 is a little stricter on long words and a little kinder on short tests with a single slip.

Accuracy, corrected and uncorrected errors

Errors come in two kinds, and they affect your score differently.

• Uncorrected errors are mistakes still in the text when the test ends. They lower your net WPM directly.
• Corrected errors are mistakes you fixed with Backspace. They do not appear in the final text, so they do not lower net WPM directly, but they cost you time, and that time comes out of your score anyway.

Accuracy captures both. On Presspace it is calculated per keystroke:

Accuracy = correct keystrokes ÷ all keystrokes × 100

Worked example. During a test you pressed 300 keys. 279 of them were the right character in the right place, and 21 were wrong (some you fixed, some you didn't).

• 279 ÷ 300 = 0.93
• 0.93 × 100 = 93% accuracy

That same test would show a raw WPM of 60 (300 keystrokes ÷ 5 ÷ 1 minute). Put next to a net score of 53, the picture is clear: the fingers are moving at 60 but only 53 words a minute of usable text came out. The gap is the price of the mistakes.

Converting CPM, KPH and WPM

Some tests, especially in Europe and for data entry work, report characters per minute (CPM) or keystrokes per hour (KPH) instead of WPM. Because a standard word is five characters, converting is simple arithmetic.

• CPM to WPM: divide by 5
• WPM to CPM: multiply by 5
• CPM to KPH: multiply by 60
• WPM to KPH: multiply by 300 (5 characters × 60 minutes)

Two worked conversions:

• A test reports 425 CPM. 425 ÷ 5 = 85 WPM.
• A data entry job asks for 10,000 KPH. 10,000 ÷ 60 ≈ 167 CPM, and 167 ÷ 5 ≈ 33 WPM.

Be careful with that last one. Data entry KPH figures are often measured on numbers, forms and the numeric keypad, sometimes counting keys like Tab and Enter, and they rarely penalise errors the way a WPM test does. Treat the conversion as a rough guide only. Our article on typing tests for jobs goes into how employers measure this.

Consistency: how steady your speed is

Two typists can both score 60 WPM with very different typing. One holds a steady pace. The other races through easy words and stalls on hard ones. Presspace shows a consistency score to tell these apart.

We sample your speed every three seconds during the test, then measure how spread out those samples are:

Consistency = 100 × (1 − standard deviation ÷ mean)

The standard deviation divided by the mean is called the coefficient of variation. It describes spread as a fraction of your average speed, so a fast typist and a slow one can be compared fairly.

Worked example. Five 3-second samples come out at 60, 66, 54, 60 and 60 WPM.

• Mean = (60 + 66 + 54 + 60 + 60) ÷ 5 = 60
• Differences from the mean: 0, 6, −6, 0, 0. Squared: 0, 36, 36, 0, 0. Average of those: 72 ÷ 5 = 14.4
• Standard deviation = √14.4 ≈ 3.79
• Coefficient of variation = 3.79 ÷ 60 ≈ 0.063
• Consistency = 100 × (1 − 0.063) ≈ 94%

If the samples had been 40, 80, 50, 70 and 60, the mean would still be 60 but the consistency would drop to about 76%. Low consistency usually points to a rhythm problem, and our guide on how to type faster has drills for smoothing it out.

Why different sites give different numbers

Even with the same five-character rule, you can get noticeably different scores on different tests. The usual reasons:

• Net method. Subtracting errors per minute and crediting only correct words give slightly different results.
• When the clock starts. Most modern tests, ours included, start timing on your first keystroke. Older tests sometimes start the moment the page loads, which costs you your reaction time.
• Whether errors must be fixed. Some tests block you until you correct a mistake. That turns every error into lost time instead of a penalty.
• Test length and text. Short tests and common words produce higher numbers. A quote with punctuation and capitals will be slower than a list of simple words.
• Rounding. Small, but worth knowing. 52.8 might be shown as 52 or 53 depending on the site.

None of these makes one test right and another wrong. They just measure slightly different things. For tracking your own progress, use one test format and stick to it. For context on what your number means, see what is a good typing speed.

Check a result yourself

If you ever want to verify a score by hand, here is the whole process:

1. Note the test length in minutes. A 30-second test is 0.5, a 2-minute test is 2.
2. Count the characters in the words you typed correctly, adding one for each space.
3. Divide by 5, then by the minutes. That is your net WPM.
4. Count every keystroke, divide by 5 and by the minutes. That is your raw WPM.
5. Divide correct keystrokes by all keystrokes and multiply by 100. That is your accuracy.

Try it now: take a 30-second test, copy down the stats on the results screen, and run the numbers. Once you have done it once, every score you see afterwards will make more sense.

Questions people ask

Read next