How to Use This Tool
Paste your text. You get both Flesch scores, and a range showing how much they could move if the syllable count is a little off — which it always is.
The formulas
These are the published 1948 formulas, unchanged:
- Reading ease = 206.835 − 1.015 × (words ÷ sentences) − 84.6 × (syllables ÷ words). Higher is easier; 60–70 is described as plain English.
- Grade level = 0.39 × (words ÷ sentences) + 11.8 × (syllables ÷ words) − 15.59. The output is a US school grade.
Only two things go in: sentence length and syllables per word. Vocabulary difficulty, abstraction, sentence structure and whether the text makes sense are all invisible to it. A page of grammatical nonsense in short words scores as highly readable.
Why there is a range
Look at the weights. Syllables per word is multiplied by 84.6, and that ratio is usually somewhere near 1.5, so the term is worth roughly 127 points of the score. A small proportional error in the syllable count therefore moves the result a lot.
And the syllable count is definitely a little wrong. No rule-based counter handles every (two or three, depending who you ask), poem (two), business (two, not three) and chocolate (two or three) all correctly at once. Getting these right needs a pronunciation dictionary, which is megabytes.
So the tool shows the score you would get if the counter were 5% high or 5% low. If that band spans two readability categories, the categories are not telling you anything.
Sentence splitting moves it too
Words per sentence carries a weight of 1.015, so every fabricated or missed sentence break shifts the average and drags the score with it. Splitting on every full stop — which most readability tools do — turns e.g., Dr. and 3.5 into sentence boundaries, inflating the sentence count and making the text look easier than it is.
The splitter here holds a list of common abbreviations and ignores full stops inside numbers. The sentence list is shown below so you can check what it decided rather than trusting the count.
What the score is and is not for
It is a rough flag for “this is denser than my audience expects”. It works because long sentences and long words really do correlate with difficulty.
It is not a target. Optimising for it rewards chopping sentences at random and swapping precise words for vague short ones, which makes writing worse while the number improves. The grade levels are also calibrated on mid-twentieth-century American school material, so treating “grade 8” as a statement about any particular reader today is reading far more into it than it can support.
