Readability Scores, Explained
After reading this you will know what a Flesch score of 60 or a Gunning fog of 12 actually measures, how each formula turns sentence length and word length into a grade, and when to ignore the number entirely.
What a readability score is
A readability score is a single number that estimates how hard a piece of English prose is to read. Almost every classic formula uses just two facts about your text: how long the sentences are, and how long the words are. Longer sentences and longer words push the difficulty up. That is the whole idea.
Here is the hook. Take this sentence: "The cat sat on the mat." Six words, one sentence, every word one syllable. Now compare: "The domesticated feline positioned itself comfortably upon the rectangular floor covering." Twelve words, one sentence, and several words run to three or four syllables. Both describe the same scene. The first scores around a 2nd grade reading level; the second scores around college level. The formulas cannot see meaning. They see length, and length correlates with difficulty well enough to be useful.
The tool reports six formulas at once: Flesch Reading Ease, Flesch-Kincaid Grade, SMOG, Gunning fog, Coleman-Liau and the Automated Readability Index (ARI). It also shows you the raw inputs each formula saw, so you can check the arithmetic yourself.
When to use it, and when not to
These formulas were built for graded English prose. Rudolf Flesch published Reading Ease in 1948. J. Peter Kincaid and colleagues recalibrated it for the U.S. Navy in 1975. G. Harry McLaughlin published SMOG in 1969. Robert Gunning's fog index dates to 1952. Every one of them was fit to running sentences of ordinary English.
Use a score to compare drafts, to check that a public notice or a help article is not drifting into legalese, or to keep a consistent target across a team. A drop from grade 14 to grade 9 after editing is real evidence you simplified something.
Do not score code, bullet lists, tables, headings, or non-English text. The formulas divide by sentence count. A page of five-word bullet points has almost no punctuation, so the sentence detector either finds one giant "sentence" or none, and the grade swings wildly. A score of "grade 3" on a bulleted spec sheet means nothing.
The formulas and the intuition behind them
Start with the two inputs shared by most formulas. Let W be the number of words, S the number of sentences, and Y the total number of syllables. The average sentence length is W/S and the average syllables per word is Y/W.
Flesch Reading Ease turns these into a score where higher is easier:
The constant 206.835 sets the ceiling. Every extra word per sentence costs about 1.015 points. Every extra tenth of a syllable per word costs about 8.46 points, so word length dominates. A typical score runs from 0 (very hard) to 100 (very easy), though the formula can go outside that range.
Flesch-Kincaid rearranges the same two inputs into a U.S. school grade:
A result of 8.0 means an average 8th grader (about 13 years old) should manage the text.
Gunning fog counts "complex" words, meaning words of three or more syllables. Let C be that count:
SMOG works from the same complex-word count but takes a square root, and it was calibrated to expect at least 30 sentences:
Coleman-Liau and ARI skip syllables entirely. They use characters per word instead, which a computer counts exactly with no heuristic. Let L be letters per 100 words and St be sentences per 100 words for Coleman-Liau:
ARI uses characters Ch, words and sentences directly:
The two character-based formulas are worth trusting more when syllable counting is shaky, for example on text with many proper nouns or invented product names.
A worked example
Scoring three plain sentences
Take this sample text: "The cat sat on the mat. It watched the door. A bird flew past the window and the cat jumped."
- Count sentences:
3. - Count words: 6 + 4 + 11 =
21words. - Count syllables. Every word here is one syllable except "window" (2) and "jumped" (1). Total syllables Y = 22.
- Average sentence length W/S = 21/3 = 7.0.
- Average syllables per word Y/W = 22/21 \approx 1.048.
- Flesch Reading Ease: 206.835 - 1.015 \times 7.0 - 84.6 \times 1.048 \approx 206.835 - 7.105 - 88.66 = 111.07. Anything above 100 means "extremely easy", which fits.
- Flesch-Kincaid Grade: 0.39 \times 7.0 + 11.8 \times 1.048 - 15.59 \approx 2.73 + 12.37 - 15.59 = -0.49. Round up to grade 0, meaning below 1st grade.
The negative grade is not a bug. The formula is linear and can produce values below zero for very short, simple text. Read it as "as easy as English gets".
Coleman-Liau and ARI would give slightly different grades for the same sentences because they count letters, not syllables. That is expected. The six formulas rarely agree to the decimal. Look at the cluster, not one number.
How word length moves the score
Because word length carries the heavy coefficient in Flesch, small changes in syllables per word swing the score fast. The chart below fixes sentence length at 15 words and varies only the average syllables per word.
Reading and interpreting the numbers
Translate Reading Ease into a grade band with the ranges Flesch published. The higher the ease, the lower the required grade.
| Reading Ease | Difficulty | Grade level |
|---|---|---|
| 90 to 100 | Very easy | 5th grade |
| 80 to 90 | Easy | 6th grade |
| 70 to 80 | Fairly easy | 7th grade |
| 60 to 70 | Plain English | 8th to 9th grade |
| 50 to 60 | Fairly hard | 10th to 12th grade |
| 30 to 50 | Hard | College |
| 0 to 30 | Very hard | College graduate |
The grade formulas already speak in grades, so a Flesch-Kincaid of 9.2, a Gunning fog of 10, and a Coleman-Liau of 8.5 together tell you the text sits near the end of middle school. Most general-audience writing aims for grade 8 to 10. Plain-language guidance for public documents often targets grade 6 to 8.
Treat a difference of one grade as noise. A difference of three or four grades between two drafts is a real signal.
Common mistakes
The first mistake is trusting the last decimal. These formulas were fit by linear regression to reading tests decades ago. Their standard error is measured in grade levels, not hundredths. Report "about grade 9", not "grade 9.14".
The second mistake is writing to the score. You can lower any grade by chopping sentences and swapping in short words, but you can produce choppy, unclear prose that a human finds harder to follow. The formula rewards "He ran. He fell. He cried." over a smoother sentence, even when the smoother sentence reads better. The number is a smoke detector, not a style guide.
Syllable counting is a heuristic: it finds vowel groups and applies a few corrections for silent e and common endings. It miscounts names, acronyms and unusual words. "Queue" (1 syllable) and "cache" (1 syllable) trip up many counters. Expect the syllable-based formulas (Flesch, Flesch-Kincaid, SMOG, Gunning fog) to differ from a careful hand count by a few percent. Coleman-Liau and ARI avoid this by counting letters.
The third mistake is scoring too little text. SMOG assumes 30 sentences. Applied to three sentences it extrapolates and becomes unstable. For any of these formulas, aim for at least 100 words, and preferably several hundred, before you trust the grade.
Related tools
To get the raw counts these formulas depend on, the Word & Character Counter gives you words, characters, sentences and paragraphs directly. If a score surprises you, the Word Frequency & N-gram Analyzer shows which long or repeated words are driving the difficulty up. When a syllable count looks wrong, the Unicode Inspector reveals hidden characters or odd punctuation that can confuse the sentence detector. To measure whether an edit actually lowered the grade, compare the before and after with the Text Diff Checker.
Frequently asked questions
Which readability formula should I use?
For general writing, Flesch-Kincaid Grade is the most widely cited and the easiest to explain. If your text has many proper nouns or coined terms, prefer Coleman-Liau or ARI, because they count letters instead of guessing syllables. Better still, read all six and look at whether they cluster.
Why do the six formulas disagree?
They were calibrated on different reading tests with different sample texts and different definitions of a "hard" word. Gunning fog and SMOG count words of three or more syllables; Flesch weights average syllables per word; Coleman-Liau and ARI use characters. Disagreement of one to two grades is normal.
What is a good Flesch Reading Ease score?
For a general audience, aim for 60 to 70, which maps to about 8th or 9th grade. Technical documentation often lands at 30 to 50. Children's books run 90 and above. There is no single "good" number: it depends on who will read the text.
Can I use these scores for languages other than English?
No. The constants were fit to English. Other languages have different syllable structures and word lengths, so an English formula will report a meaningless number. Use a formula built for that language instead.
Does the tool send my text anywhere?
No. The scoring runs entirely in your browser. Nothing you paste leaves your device.
Related tools
Keyboard Layout Analyzer
Paste your text and see QWERTY, Dvorak and Colemak fight it out — finger heatmaps, travel and effort.
Text Diff Checker
Compare two texts line by line, with word highlights.
Word & Character Counter
Words, characters, sentences, paragraphs and reading time.
Word Cloud Generator
Paste any text and get a word cloud in seconds — shapes, fonts, colours, stop words in 8 languages, PNG/SVG export.
Anagram & Word Finder
Words from your letters, exact anagrams and c?t patterns.
Repeated Phrase Finder
Find the phrases you repeat too often in a draft or manuscript.