How to Measure Your Typing Speed Accurately

Take a typing test twice in a row and you'll likely get two different numbers — sometimes five or more WPM apart. Neither is wrong; typing speed is a distribution, not a constant, and a single test is one sample from it. If you're using tests to track improvement, a little measurement discipline separates real progress from noise — and explains most of the mysteries, like why your score "dropped" overnight or why a friend's number from a different site isn't comparable to yours.

Why single scores bounce around

Several sources of variance stack on every run. Warm-up is the biggest: first tests of the day run noticeably slower than the third or fourth, because motor programs need a few minutes to come fully online. Fatigue pushes the other way — scores sag at the end of a long day or a long session. The text itself varies: a run of short common words scores faster than one that happens to include "rhythm" and "bureaucracy." And attention wobbles for reasons that have nothing to do with skill. A five-WPM swing between consecutive tests carries almost no information. Treat any single score as weather; skill is the climate.

Test length changes what you're measuring

Duration isn't just preference — different lengths measure genuinely different things. A 15-second test measures burst speed: no stamina, no sustained focus, and one stumble swings the score wildly, but it's a fine tool for quick, frequent reps. A 60-second test measures the opposite — the ability to hold pace and concentration — and typically lands a few WPM below your short-test scores, which surprises people the first time. It shouldn't: holding speed is a harder skill than reaching it. The 30-second test is the practical middle, long enough to smooth out single stumbles, short enough to repeat several times in a session. None of them is the "true" number. They're different questions — just never compare a score from one duration to a score from another and call the difference progress.

Why scores differ between sites

Two tests can watch identical typing and report different numbers, because scoring rules differ. A "word" is standardized as five characters, but tests diverge on everything after that: whether errors subtract from speed (net WPM) or don't (raw), how corrections are treated, whether the timer runs wall-clock or pauses, and what happens to the word you're mid-way through when time expires. This site scores net WPM — only correct characters count toward speed — on wall-clock time, and corrections fix your text without repairing your accuracy score, so accuracy always reflects your first-attempt error rate. Full details are in our guide on how WPM is calculated. The practical rule: cross-site comparisons are meaningless, and that includes comparing your number to a friend's from somewhere else. Compare your scores on one test, one duration, over time — that comparison is solid.

Establishing a real baseline

To find your actual speed, take five or six tests at the same duration, spread over two or three days, after at least one warm-up run each time. Ignore the best and the worst; the middle scores are your baseline. This takes fifteen minutes total and gives you a number stable enough to be worth comparing against next month. Skipping it is how people conclude they "got worse" after a week of practice, when they actually compared a lucky early run against a tired later one.

Tracking progress without fooling yourself

Improvement hides inside the noise, so watch aggregates rather than single runs. Weekly averages are the honest signal: five WPM of week-over-week average movement is real progress even when individual tests still bounce. Watch accuracy alongside speed — accuracy climbing toward 98% at level speed is genuine improvement that a speed-only reading misses, and it usually converts into WPM within a couple of weeks. Personal bests are worth celebrating, but they're your luckiest sample, not your level; expect most runs to land under a fresh best for a while. That's regression to your average, not decline.

A measurement routine that works

In practice: warm up with a throwaway run before any test you intend to count. Test at a consistent duration and, roughly, a consistent time of day. Check weekly averages on your history page rather than reacting to single runs — this site keeps your last 50 results in your browser for exactly that purpose. Re-baseline once a month with the multi-test routine above. It's less exciting than chasing a new best every session, but it means that when your numbers say you've improved, you actually have — and knowing that is what makes the practice worth it.