Number typing test
Loading the test…
This is a typing test with numbers: figures of one to four digits, some with decimals, are mixed into the normal word stream, so you practise reaching up to the number row and coming back to the home row without looking down. You get net WPM, raw WPM, accuracy and a per-key heatmap that shows exactly which digits you fumble. It's free, it runs in the browser and there's nothing to sign up for.
How the numbers test works
Numbers appear inside running text, the way they do in addresses, dates, prices, order numbers and reference codes. The test is built for the number row above the letters, because that's where your hands already are when numbers turn up in a sentence. If you prefer the numeric keypad you can use it: the test checks the character you typed, not which key produced it.
It's a one-minute timed test. Add the punctuation toggle in the ribbon if you want commas, periods and capitals as well, which is close to what real forms and reports look like.
Scoring follows the same rules as every other test here. A number is a word: "2048" followed by a space is five characters, exactly one standard word. One wrong digit makes the whole number wrong, and wrong words add nothing to net WPM. Raw WPM counts every character you typed. Accuracy counts every character keypress, including the ones you corrected. The definitions with a worked example are on the scoring methodology page.
Why numbers slow almost everyone down
Most people learned letters by touch and digits by sight. The number row is the furthest row from home, the keys are small targets, and nothing in everyday typing forces you to learn them properly. So a typist who never looks at the letters will often glance down for every 7 or 8, and each glance costs far more than the keystroke itself, because your eyes then have to find their place in the text again.
There's also less room for error. In a word, a typo is often a neighbouring letter you'd notice at once. In a number, 3816 and 3826 look almost the same on screen, and there's no spelling to tell you something is off. That's why accuracy usually drops when the numbers toggle is on, and in net WPM a number with one wrong digit scores zero.
Number row fingering
Standard touch typing gives each digit to the finger directly below it on the home row. If you've been using two fingers for the whole number row, this table is where to start.
| Digit | Finger | Reach from |
|---|---|---|
| 1 | left little | A |
| 2 | left ring | S |
| 3 | left middle | D |
| 4 | left index | F |
| 5 | left index | F (reach right) |
| 6 | right index (many typists use left) | J (reach left) |
| 7 | right index | J |
| 8 | right middle | K |
| 9 | right ring | L |
| 0 | right little | semicolon key |
The 5 and 6 are the hard ones because they are two rows up and one column in. Pick one finger for 6 and stick with it; switching between hands from day to day is what makes it unreliable.
Technique: reach and come back
Keep the other fingers touching the home row while one finger reaches up. If your whole hand lifts off to type a 9, it has to find home again afterwards, and that's where the pauses come from. Your index fingers should stay on the bumps of F and J whenever they aren't the ones reaching.
Read numbers in chunks. A four-digit number is easier to hold in your head as two pairs: 4512 as "45, 12". Glance at the number once, say the pairs to yourself, then type without looking back at the screen between digits.
Slow down to stay accurate. On letters, typing at the edge of your speed costs you an occasional word. On numbers, the same pace produces a steady trickle of wrong digits, and each one zeroes the number. Aim for clean numbers first; speed follows once the reaches are automatic.
Watch the heatmap. After each run the results screen lists the keys you missed most. If the same two digits keep appearing, that's your drill. Practice builds a short session around the keys you actually miss, but its drills are made of real words, so digits are best drilled here.
A worked example: how much the numbers cost you
Say you type 60 WPM net on plain words, which is 300 characters a minute. Now take a stretch of text where every fifth word is a four-digit number. If you type each number at half your letter speed, that number takes twice as long as an ordinary five-character word. Out of every five words, four take one unit of time and the number takes two, so the same 25 characters now take six units instead of five. Your speed falls to five-sixths of 60, which is 50 WPM, and that's before a single wrong digit.
Add one wrong digit per ten numbers and you lose another word in fifty, which takes you to about 49 WPM. The lesson is that speed on the number row matters, but a wrong digit costs you the whole number, so reliable reaches are worth more than fast ones.
Number row or numeric keypad?
For numbers inside text, the number row wins: your hands never leave the letters, and a figure in the middle of a sentence is over before you'd have moved to the keypad and back. For long columns of figures with little or no text, the keypad is faster. All ten digits sit under one hand, the 5 has a raised bump for orientation, and data-entry staff on numeric work generally use it.
This test is aimed at the first case. If your job is mostly numeric entry, the KPH test reports your speed in characters per minute, which multiplies straight into keystrokes per hour, the unit data-entry roles usually state their requirements in, and you can type its digits on the keypad if you like.
Numbers in data-entry and exam tests
Government typing exams in India mix figures into their passages, and some state their requirement in key depressions rather than words. The Staff Selection Commission's own documents give two useful benchmarks: the Data Entry Operator skill test asks for 8,000 key depressions per hour, and the CGL Data Entry Skill Test sets a passage of about 2,000 key depressions in 15 minutes, which works out to the same 8,000 per hour. Divided by 300 (five characters per word, sixty minutes per hour), 8,000 key depressions an hour is about 27 WPM.
That sounds modest, but those passages carry figures and punctuation, and the exam's own counting rules differ from ours, so don't treat our score as your exam score. Our net WPM is stricter: a single wrong digit cancels the whole number. Rules also change from one notification to the next, so check the current notice. The SSC CGL DEST preset sets up a 15-minute test with numbers and punctuation, and the exam practice page lists the other presets.
On AZERTY and other layouts
On a French AZERTY keyboard, the digits on the number row need Shift, and the unshifted keys give accented letters and symbols. That makes number-heavy text noticeably harder on AZERTY, and many French typists use the keypad for any run of figures. If that's your keyboard, practise here with your system layout as it is, or try the AZERTY typing test to work on the layout itself.
Frequently asked questions
Do I have to use the number row?
No. The test checks the digit you typed, so the keypad works too. The number row is usually faster for numbers mixed into text, and the keypad for long columns of figures.
How long are the numbers?
Between one and four digits, mixed in among ordinary words, and about one number in seven has a decimal part of one or two digits. A four-digit number plus its space is exactly one standard five-character word.
What is a good speed on a number typing test?
Compare it with your own plain-word score rather than with other people. Run the 1-minute typing test and this test back to back. If your numbers score is within about 20% of your plain score with high accuracy, your number row technique is in good shape. A bigger gap means the reaches still need work.
Why did one wrong digit cost me so much?
Net WPM counts only words typed exactly right, and a number is one word. 1957 typed as 1975 is as wrong as a misspelled word, so it adds nothing to your net figure, though raw WPM still counts it.
Can I add punctuation too?
Switch on the punctuation toggle in the ribbon. Numbers and punctuation together come close to real data-entry material, and the punctuation typing test covers the marks on their own.