How to improve your typing speed
Nearly every guide opens by telling you to learn touch typing. When researchers put motion-capture markers on people's hands and measured it, that turned out not to be the dividing line at all.
Correction, August 2026. An earlier version of this article contained a quotation attributed to “Typing research, Feit et al., 2016” that those researchers never wrote — it was invented. It also gave a week-by-week improvement schedule with no basis. Both have been removed. The article has been rebuilt around what the cited papers actually report, with the figures reproduced so you can check them.
The touch-typing assumption, tested
In 2016, Feit, Weir and Oulasvirta recruited 30 people, fitted their hands with motion-capture markers and their heads with eye-tracking glasses, and recorded exactly how they typed. About 43% described themselves as touch typists; the rest were self-taught.
| Group | Mean speed | Std. dev. | Uncorrected errors |
|---|---|---|---|
| Trained touch typists | 57.83 WPM | 15.25 | 0.76% |
| Self-taught, non-touch | 58.93 WPM | 10.82 | 0.47% |
The self-taught group was marginally faster, and the difference was not statistically significant (Mann–Whitney U = 103, p = 0.38). The paper puts it plainly: typists using as few as one or two fingers per hand reached rates “normally attributed to touch typists.”
The number of fingers you type with does not predict how fast you type. What you do with them consistently does.
This does not mean touch typing is useless. It means the standard advice — that learning the home-row system is the single change that unlocks speed — is not supported by the measurement. Plenty of fast typists never learned it. If you have been avoiding practice because you type “wrong,” that excuse does not survive the data.
The three things that did predict speed
The same study identified what separated the fast typists from the slow ones, regardless of technique. All three are trainable, and none is about finger count.
1. Unambiguous finger-to-key mapping
Fast typists press a given letter with the same finger every time. Slow typists vary — sometimes the index finger, sometimes the middle, depending on what came before. Each variation is a decision, and decisions cost milliseconds that compound across every keystroke.
This is the most actionable finding in the whole literature, and it is not “use the correct
fingers.” It is use the same finger consistently, whichever one you have settled on.
If you hit B with your right index finger and the textbook says left, that costs you nothing
— as long as it is always the right index finger.
2. Active preparation of upcoming keystrokes
Fast typists move a finger toward the next key while the current one is still being pressed. The 2018 follow-up study of 168,960 people found this is more prevalent than expected: rollover, where the next key goes down before the previous one comes up, is common rather than exceptional.
This is what “reading ahead” actually means physically. You are not reading ahead for comprehension; you are buying time for your hand to travel.
3. Minimal global hand motion
Fast typists keep their hands anchored and move fingers. Slow typists move whole hands around the keyboard, which means re-locating the anchor position after every excursion. Anchoring is exactly what the home row is for — which is why home-row practice helps even though the touch-typing label does not.
Sources. Feit, A.M., Weir, D. and Oulasvirta, A. (2016). How We Type: Movement Strategies and Performance in Everyday Typing, CHI '16. Dhakal, V., Feit, A.M., Kristensson, P.O. and Oulasvirta, A. (2018). Observations on Typing from 136 Million Keystrokes, CHI '18. Both datasets are publicly released.
Accuracy is not a trade-off
The usual framing is that you choose between speed and accuracy. The 168,960-participant dataset does not show that. Faster typists made fewer uncorrected errors, not more — the two move together.
The mechanism is unglamorous. An error costs you twice: once to type the wrong character, once to backspace and retype it. At 60 WPM a corrected mistake costs roughly a second, and three mistakes a minute is a measurable fraction of your score. Accuracy is not a virtue you trade speed for; it is where a chunk of your speed is currently going.
Practically: if your accuracy is below about 95%, that is the constraint, and pushing for a higher peak WPM will not move your average. TypeCrt reports raw WPM alongside net WPM precisely so you can see how much errors are costing you — the gap between the two numbers is the cost.
Why targeted practice beats general practice
If speed lives in consistent finger mapping and prepared transitions, then it lives in specific letter pairs, not in general fluency. The 2018 study supports this directly: letter pairs typed by different hands or different fingers were more predictive of overall speed than, for instance, letter repetitions.
That has a sharp implication for how you practise. Typing ordinary English text drills the letters you already type well, because those are the frequent ones — the top ten letters make up roughly 70% of written English. Your weak keys are, almost by definition, the ones you rarely encounter, so general practice barely touches them.
- Run a normal test on TypeCrt and open the per-key heatmap afterwards.
- Note which keys are slowest. They are usually a handful, and usually not the ones you would guess.
- Switch to KeyForge, which weights generated words toward exactly those keys, and watch the same panel over a week.
This is the one thing we built TypeCrt to do that a plain typing test does not, and it follows from the research rather than from a marketing decision.
A practice session built on this
Fifteen minutes, structured around the three predictors rather than around volume.
| Time | What | Which predictor |
|---|---|---|
| 5 min | KeyForge on your weakest keys, deliberately slowly, using the same finger every time | Consistent mapping |
| 5 min | Timed tests at a speed where accuracy holds above 97% | Preparation and rollover |
| 5 min | Quote or custom text — real prose with punctuation and capitals | Transfer to real typing |
Do the slow, deliberate portion first. Consistency of finger mapping is a motor habit, and habits are formed at the speed you can execute them correctly, not at the speed you wish you could.
The ordinary advice, kept honest
These are conventions rather than findings. They are reasonable and widely repeated, and we are labelling them as such rather than dressing them up as science.
- Stop looking at the keyboard. Follows directly from predictor 3 — you cannot keep your hands anchored if your eyes are recalibrating them. TypeCrt's blind mode hides your input so you cannot check mid-test.
- Practise punctuation and capitals. Most people are far slower with shift combinations and symbols than with letters, and real writing is full of them. TypeCrt's punctuation mode mixes them in.
- Sit so your wrists are neutral. No study we found links posture to WPM. It is about not hurting yourself over years of typing, which is a better reason anyway.
- Short daily sessions over occasional long ones. Widely recommended and consistent with motor-learning research in general, though we have not found a typing-specific study that measures it.
- Track the trend, not the test. Single-test variance is large — the population standard deviation was over 20 WPM, and your own scatter between tests is not trivial. TypeCrt's profile charts a moving average for this reason.
What we will not tell you. How many weeks it takes to reach 80 WPM. We could not find a credible study, the honest answer varies enormously between people, and the confident schedules you will find elsewhere are invented. Measure your own trend line instead — it is the only figure that describes you.
Find your weak keys
One test produces the per-key heatmap. That tells you what to practise, which is most of the problem.
Start a typing test →