Channel 7 of 7 · Reading grade

Flesch vs Flesch-Kincaid vs Gunning Fog: readability scores compared

Four common readability formulas, the arithmetic behind them, and why none of them was built for a 30-word post.

Short answer

All four formulas measure the same two things, sentence length and word length, and weigh them differently. Flesch Reading Ease gives a 0 to 100 score where higher is easier; Flesch-Kincaid, Gunning Fog and SMOG give a US school grade. For short social posts, Flesch-Kincaid grade is the most practical, because it uses every word and reads as a familiar grade number.

Key takeaways

  • Every formula here combines average sentence length with some measure of long words.
  • Flesch Reading Ease runs the opposite way to the grade formulas: higher means easier.
  • Gunning Fog and SMOG count words of three or more syllables, so jargon pushes them up fast.
  • On a single post, any formula is unstable; use it to find the heavy sentence, not as a target.

Paste the same paragraph into four readability checkers and you may get four different numbers: a 64, a grade 8.2, a fog index of 10.5 and a SMOG of 9.8. They aren't contradicting each other. They're four recipes using the same two ingredients.

Here's how each one works, and which is the most useful for a post on X.

What do readability formulas actually measure?

All four formulas measure sentence length and word length, and none of them measures meaning.

Readability formulas date from the mid-1900s, when researchers wanted a quick way to match texts to readers. They found that two easily counted features predicted reading difficulty fairly well:

  1. How many words are in each sentence on average.
  2. How long the words are, counted in syllables.

The formulas differ in how they weight these two and how they define a "long" word. None of them can tell whether a short word is unfamiliar or whether a sentence is confusing for some other reason. "The cat sat" and "The fief fell" score the same.

How is Flesch Reading Ease calculated?

Flesch Reading Ease is 206.835 minus 1.015 times the average sentence length, minus 84.6 times the average syllables per word.

Rudolf Flesch published it in 1948. The result usually falls between 0 and 100, and higher means easier.

Score Rough feel
90 to 100 Very easy
60 to 70 Plain English
30 to 50 Difficult
Below 30 Very difficult, academic or legal

The reversed direction trips people up. With Reading Ease you want a higher number; with every grade formula you want a lower one.

How is the Flesch-Kincaid grade calculated?

Flesch-Kincaid grade is 0.39 times the average sentence length, plus 11.8 times the average syllables per word, minus 15.59.

J. Peter Kincaid and colleagues derived it in 1975 for the US Navy, which wanted to check training manuals against the reading skills of enlisted personnel. It uses the same two ingredients as Reading Ease, rescaled so the result reads as a US school grade.

This is the number the Twittermeter post meter shows on its reading grade dial. Grade 6 to 8 suits most general audiences. Our guide to writing readable posts covers how to bring a high grade down.

How is the Gunning Fog index calculated?

Gunning Fog is 0.4 times the sum of the average sentence length and the percentage of words with three or more syllables.

Robert Gunning introduced it in the early 1950s for business and newspaper writing. Instead of averaging syllables over all words, it counts "complex" words, those of three or more syllables, with some exceptions such as proper nouns and familiar compound words in the original method.

Because it counts long words as a share, Fog reacts sharply to jargon. A post with "implementation", "functionality" and "optimization" jumps several grades even if the sentences are short.

How is the SMOG grade calculated?

SMOG grade is 1.043 times the square root of polysyllables scaled to 30 sentences, plus 3.1291.

G. Harry McLaughlin published SMOG in 1969. The original method samples 30 sentences and counts words of three or more syllables. The formula used today is:

SMOG = 1.043 × √(polysyllables × 30 ÷ sentences) + 3.1291

SMOG is popular in health communication, partly because it tends to predict full comprehension rather than partial understanding. It was designed for at least 30 sentences, so on a three-sentence post the result is little more than a guess.

How do the four compare side by side?

They agree on direction most of the time and differ on scale, sensitivity to jargon, and minimum text length.

Formula Output Long-word measure Best text length Most sensitive to
Flesch Reading Ease 0 to 100, higher is easier Syllables per word 100+ words Both factors
Flesch-Kincaid grade US grade Syllables per word 100+ words Sentence length
Gunning Fog US grade % of 3+ syllable words 100+ words Jargon
SMOG US grade Count of 3+ syllable words 30 sentences Jargon

Which formula should you use for posts on X?

Use Flesch-Kincaid grade, and treat it as a pointer to the problem sentence rather than a score to hit.

A post is usually 20 to 50 words and two or three sentences. Every one of these formulas was designed for longer passages, so all of them wobble on posts. Flesch-Kincaid tends to wobble less than Fog and SMOG, because it averages syllables across every word instead of counting only the long ones. It also gives a number people understand without a conversion table.

A practical way to use it:

  1. Check the grade. If it's over about 10, something is heavy.
  2. Find the longest sentence and split it.
  3. Find the longest word and see if a shorter one does the job.
  4. Check again and stop when the post reads well aloud.

What do none of the formulas catch?

They miss unfamiliar short words, unclear structure, missing context, and whether the post makes sense to its audience.

A post can score grade 5 and still confuse people, for example if it uses an internal acronym or refers to a conversation the reader never saw. Read the post as someone outside your team would see it in a feed. Plain language habits, covered in plain language for social posts, catch what the arithmetic can't.

Questions

Which readability score does Twittermeter use?

The meter shows the Flesch-Kincaid grade level. It removes links and @handles before scoring and waits for at least a dozen words before coloring the dial, because very short text gives unstable results.

What Flesch Reading Ease score is good?

Scores of 60 to 70 are usually described as plain English that most adults read easily. Above 80 is very easy. Below 50 starts to feel like dense business or academic writing.

Why do two tools give different grades for the same text?

Mostly because they count syllables differently. Counting syllables by computer is approximate, and tools use different rules for silent e, -ed endings and abbreviations. Sentence splitting also varies, especially around line breaks, emoji and URLs.

Twittermeter is independent and not affiliated with X Corp. X changes its limits and features from time to time. If something here doesn't match what you see in the app, X's help center has the final word.