Boneyard Tools

Readability formulas compared

How Flesch, Gunning Fog, Coleman-Liau, ARI and SMOG differ, why they disagree, and how to read a panel of grade scores together.

Why one score is not enough

Every readability formula reduces writing to a few countable signals, and each one chooses a different set. Because of that, a single index can flatter or punish a passage depending on which signal it happens to weigh. Running six formulas over the same word, sentence and syllable counts turns a lone opinion into a panel. When the panel agrees you can trust the grade, and when it splits the disagreement itself tells you what to fix.

Syllable formulas versus character formulas

Flesch Reading Ease and Flesch-Kincaid Grade both hinge on syllables per word, so they react sharply to polysyllabic vocabulary. Coleman-Liau and the Automated Readability Index instead count letters per word, sidestepping syllable estimation entirely, which makes them steadier on names and technical terms that fool a syllable counter. Gunning Fog and SMOG take a third path, counting only the share of words with three or more syllables. A passage full of long but familiar words can therefore read as easy on ARI yet difficult on Gunning Fog.

How sentence length skews the picture

Most of these formulas fold in words per sentence, so a wall of run-on sentences inflates almost every grade at once. SMOG is the exception, leaning almost wholly on complex-word density, which is why it can stay calm while the others spike. If your scores jump the moment you paste a long paragraph, split the sentences first and re-check. Sentence length is usually the fastest lever to move the whole panel in one edit.

Turning the numbers into edits

Read the spread, not just the average. If the character-based indices sit low but the syllable and complex-word indices sit high, your vocabulary is the culprit, so swap in plainer words. If every index rises together, your sentences are too long. For general web content, target a Flesch Reading Ease near 60 to 70, which lines up with roughly grade 7 to 9 on the grade indices, and re-run after each change to confirm the panel moved the way you intended.

Frequently asked questions

Do professional tools use these same formulas?

Yes. Flesch-Kincaid is built into many word processors, and Gunning Fog, Coleman-Liau, ARI and SMOG are standard in editing and plain-language tooling. The formulas here are the same public equations, run in your browser.

Which grade should I report if they disagree?

Report the cluster the majority fall into, and note the range. If four indices land near grade 9 and one says grade 14, grade 9 is the fair summary while the outlier flags a specific weakness to address.

Are these formulas accurate for every language?

No. They were tuned for English and its spelling patterns. Running them on other languages, code or heavy markup produces numbers that are not meaningful, since the syllable and letter assumptions no longer hold.