A score you can understand

The scoring rules used by this version of Typing On.

The evaluated unit

Committed text is normalized to NFC and line endings to LF, then separated into grapheme clusters for the selected language. One visible unit can contain several code points. Spaces, case, punctuation and meaningful accents count; physical key presses are not the measured unit.

Committed insertion attempts

An evaluated attempt is a committed grapheme insertion or replacement. Partially composed input is deferred. Unchanged text and deletion alone do not add attempts; retyping after deletion does.

Speed and accuracy

Calculations use actual elapsed minutes.

Gross WPM = A ÷ 5 ÷ T

CPM = A ÷ T

Net WPM = max(0, Gross WPM − U ÷ T)

Accuracy = C ÷ A × 100

A
evaluated insertion attempts
C
correct insertion attempts
U
uncorrected error units
T
actual elapsed minutes

With no input, accuracy displays 100% and speed is zero.

What corrections mean

Correcting a mismatch reduces remaining errors but does not erase earlier evaluated mistakes. Completed passages are locked: you cannot go back and edit them. Their attempts remain in the cumulative score.

A worked example

In exactly one minute, 300 evaluated attempts with 294 correct attempts and two uncorrected units give 60 gross WPM, 58 net WPM and 98% accuracy. Corrected mistakes still belong to the total attempts.

Timing and new passages

The clock begins with committed input. Practice can pause; test and exam clocks keep running when you leave the tab. New passages preserve the original timer and totals. Input after the time limit is rejected, and the result uses actual elapsed time.

Passage selection

Language and difficulty stay fixed while a passage is selected. If an exact topic has no match, another topic in the same language and difficulty is used with a notice. A filtered pool may repeat after exhaustion.

Charts and limits

Charts use actual performance samples from your session. Consistency is a clamped comparison of sample-speed variation to mean speed, not a comparison with other people. Results are independent self-assessments; an employer may use different units and penalties.

Further reading

MDN Intl.Segmenter · MDN CompositionEvent

TEST IN PROGRESS

Leave this attempt?

Your current typing progress will be discarded. Saved results stay in this browser.