Flesch Reading Ease Explained (The Score Our Word Counter Shows You)
The Flesch Reading Ease score runs from 0 to 100: 60 to 70 is plain English for a general adult audience; below 30 is academic or legal prose. The word counter on the homepage calculates it live. What the score cannot tell you is whether your writing is good — that gap matters before you optimize for it.
Where the formula came from
In 1948, Rudolf Flesch published "A New Readability Yardstick" in the Journal of Applied Psychology (vol. 32, pp. 221–233; DOI: 10.1037/h0057532). Flesch was a readability researcher and writing teacher who had spent years arguing that most American writing — particularly government and corporate prose — was needlessly inaccessible. His formula was an attempt to make that argument quantitative.
The formula looks like this:
206.835 − 1.015 × (words ÷ sentences) − 84.6 × (syllables ÷ words)Two variables move the score: average sentence length and average syllables per word. Both pull the number down when they grow. The constants — 206.835, 1.015, 84.6 — were derived empirically by calibrating against English text samples at known reading difficulty levels. Flesch's goal was a score that would track practical reading ease for American English, and for that narrow purpose the formula holds up reasonably well.
Microsoft Word bundles both the Flesch Reading Ease test and a second test, the Flesch–Kincaid Grade Level, in its built-in readability statistics. Microsoft's support documentation confirms it uses the same formula Flesch published in 1948 and recommends a score between 60 and 70 for standard documents.
Flesch Reading Ease vs. Flesch–Kincaid Grade Level
These two tests are often confused because they share a name and use the same inputs, but they produce completely different outputs.
Flesch Reading Ease gives a score from 0 to 100, where higher is easier. Flesch–Kincaid Grade Level, developed in 1975 by J. Peter Kincaid for the U.S. Navy, gives a U.S. school grade level instead — a score of 8.0 means an eighth grader can understand the text. The two scales run in opposite directions: a text that scores 65 on Reading Ease should score somewhere around 8.0 on Grade Level. Green Eggs and Ham scores −1.3 on Grade Level because its vocabulary is so simple the formula undershoots the scale.
Microsoft Word reports both when you run its readability check. The Reading Ease score and the Grade Level score are correlated inversely — improving one improves the other — but they emphasize different things rhetorically. Use Reading Ease when you want to talk about accessibility; use Grade Level when your audience thinks in educational terms (health literacy guidelines, plain-language mandates, etc.).
How to read the number
90 and above corresponds to very simple English — sentences often under 8 words, mostly one-syllable vocabulary. Beginning readers and simple consumer instructions land here. Adult prose that scores this high tends to feel choppy rather than clear; the connective tissue that links ideas across sentences is absent.
60 to 70 is the target range for most writing aimed at a general adult audience. Sentences average around 15 to 20 words; words are mostly one or two syllables. Standard journalism, edited blog posts, and consumer-facing documentation cluster here. Reader's Digest has historically scored around 65 on this scale.
30 to 50 covers college-level reading. The sentences are longer, more subordinated, and the vocabulary includes more technical or multi-syllable terms. This is where most business reports and analytical writing land.
Below 30 is the territory of academic journals, legal contracts, and government regulations. The Harvard Law Review scores in the low 30s. This is not inherently bad writing — it is writing calibrated for readers who share a specialized vocabulary and can hold a long sentence in working memory. The problem arises when text written at this level is aimed at a general reader.
Three adjustments that move the score
The formula responds to two levers: sentence length and syllable count per word. Every meaningful adjustment you can make will target one or both.
Split long sentences. Any sentence that runs more than 25 words is a candidate. Break it at the most natural logical boundary. "The report, which was submitted late due to a series of unforeseen technical complications affecting the entire department, contained recommendations the board found difficult to implement" becomes two sentences and the score rises immediately because average sentence length drops. You can verify the change by pasting both versions into the word counter on the homepage and comparing the live Flesch score.
Replace multi-syllable words with plainer equivalents. "Utilize" → "use"; "facilitate" → "help"; "approximately" → "about"; "demonstrate" → "show"; "commence" → "start." These swaps reduce average syllable count without changing meaning. The formula weights syllables at 84.6 times the sentence-length penalty of 1.015, so polysyllabic vocabulary hits the score harder than sentence length does.
Cut qualifiers and hedges. Phrases like "it is worth noting that," "one might reasonably argue," and "to a certain degree" add syllables and words with no corresponding information gain. Removing them tightens sentence length and syllable count simultaneously.
After each revision, paste into the word counter on the homepage and watch the number move. The comparison across drafts is more useful than the absolute score.
Where the formula breaks down
Flesch Reading Ease was designed for continuous expository prose in English. It gives unreliable readings for several common content types.
Short sentences with short words score extremely well regardless of meaning. "Mike eats your cat gun now" scores close to 100; it is incoherent. The formula has no access to semantics — it counts syllables and sentence boundaries, nothing else. This is why a genuine improvement in clarity often produces only a modest score gain while a passage of well-crafted nonsense can outscore a sophisticated but readable argument.
Code snippets break the syllable counter. An identifier like calculateReadabilityScore() has no agreed syllable count, and tokenizing it as a word produces garbage input to the formula.
Bullet lists and headers skew the sentence-length average. Fragments and single-line headers lower the average artificially, pushing the score upward past what the running prose would earn on its own.
The Wikipedia article on Flesch–Kincaid readability tests notes that the tests "neglect between-reader differences and effects of content, layout and retrieval aids" — which covers almost everything that actually determines whether a specific reader will understand a specific piece of writing.
Non-English text produces numbers that are not comparable to the English scale. Flesch calibrated his constants against American English; applying them to French, German, or Spanish yields a score on the same 0-to-100 dial that points in an entirely different direction.
Use the score as a directional audit tool. A 65 does not certify your writing is accessible; a 40 does not mean it is incomprehensible. But if your target reader is a general adult and your score is 20, that gap is worth investigating even if the fix is not simply shortening sentences.