Typing speed test: method

What it measures

Transcription typing speed and keystroke accuracy on the visitor's own device; a perceptual-motor skill, not a cognitive ability test.

Procedure

Short everyday passages written for this test are shuffled by the run's seed and joined to at least 1,600 characters (more than anyone types in the time). Mode: 30 or 60 s (default 60). The clock starts at the first key and stops at the time limit or at the end of the text. Correct letters turn dark, wrong ones red; Backspace corrects; Space moves to the next word. Pasting and dropping text are blocked. On a physical keyboard, an input event that adds more than 2 characters counts as paste or autofill; text committed by an input method (IME) is allowed. On phones, swipe typing and word suggestions are allowed. The page detects the input class (physical keyboard or touch). Leaving the browser tab ends the run. A friend's challenge link gives the same text.

Parameters

Scoring

Target and typed text are split at spaces. WPM = (characters of correctly typed finished words, plus one space each, plus the correct start of the word in progress) / 5 / minutes, where minutes = the shorter of elapsed time and the mode's duration (at least 1 s). Raw WPM = all typed characters / 5 / minutes. Accuracy = characters that matched the text when typed / all characters typed (corrected mistakes still count). Uncorrected errors = wrong or missing letters in finished words. Primary: WPM; higher is better. Compared with other players as the share of players beaten, only within the same input (keyboard or touch), the same duration and the same text language, from 100 comparable players. Player comparisons use only first runs (no earlier run by the same visitor, account, or IP address and device) without quality flags, with the same scorer version and the same runner settings.

Valid-run rules

A run is rejected if the duration is not 30 or 60 s, the elapsed time is longer than the mode plus 1.5 s, the keystroke counts do not add up, word times go backwards, or the typing goes past the end of the text. A run is shown but not counted for comparisons when net WPM is above 300 or, on a physical keyboard, an input event added more than 2 characters. The server rebuilds every run from the raw log (and the run's seed where there is one) and computes the score itself; anything the page says about right or wrong is ignored. A log that breaks the game's rules is rejected, and nothing is stored. Stored runs can carry quality flags that are never shown and never block a result: 10 or more runs from one IP address, a missing visitor id, tab switches, and automation signals (automated browser, impossible speed, no input events, machine-regular timing).

Differences from the classic task

Typing research and typing websites use different texts, durations and error rules (for example gross versus net WPM, whether errors must be corrected). This test counts only correctly typed words (net), lets errors stand, and uses its own passages, so numbers from other tests and from the large online datasets (Dhakal et al. 2018; Palin et al. 2019) are context only.

Limitations

Results depend strongly on the device, the keyboard layout, autocorrect and the text. Passages differ in difficulty between languages, so scores are not comparable across languages. Thirty seconds is a quick snapshot. Not a measure of intelligence or a test for any learning difficulty. For languages written without spaces between words, the word-based scoring may not work as intended (see notes).

Languages and review status

Own passages written for this test in English and 53 other languages (54 in total). Pages in other languages use the English text.