How the Coleman-Liau readability index works
The letters-per-word logic behind the Coleman-Liau index, how to read its grade output, and where a character-based score helps or misleads.
A readability score built for machines
Meri Coleman and T. L. Liau published their index in 1975 with a practical goal, scoring text automatically without a human counting syllables. Older formulas leaned on syllable counts, which are slow and error prone for a computer. By switching to letters per word the pair produced a measure a simple program could compute reliably. That design choice is why the tool can score your text the instant you stop typing, with no dictionary lookups or guesswork.
The two ingredients, L and S
The formula rests on two ratios. L is the number of letters per 100 words, a proxy for how long the average word is, and S is the number of sentences per 100 words, a proxy for how short the sentences are. The index is 0.0588 times L minus 0.296 times S minus 15.8. Longer words raise L and push the grade up, while more, shorter sentences raise S and pull the grade down. The constants were fitted to match graded reading passages.
Reading the grade output
Treat the result as a US school grade. Around 6 to 8 is comfortable for a broad audience, 9 to 12 is typical high school prose, and anything past 13 signals dense, college level writing. The plain-English band under the card translates the number for you. Because the pangram scores only 3.8, you can see that a short sentence of short common words reads well below high school level even though it uses every letter of the alphabet.
Where a character count helps and where it misleads
Counting letters makes the score exact and repeatable, which is ideal for comparing drafts or checking that content meets a target grade. The trade-off is that it treats a long technical term the same as a long but familiar word, so heavy jargon can inflate the grade even when readers know the terms. It also ignores meaning and structure entirely. For a rounded view, read the Coleman-Liau figure next to the Flesch and Gunning Fog scores shown on the same panel rather than trusting any single number.