The Raygor Readability Graph: grading text by letter count instead of syllables

Anyone who has scored a readability formula by hand knows where the time goes: syllable counting. Is "fire" one syllable or two? Does "readable" have three or four? Different counters get different answers, and on a 100-word sample those small disagreements add up to real differences in the final grade. Alton Raygor, a professor at the University of Minnesota, published his readability estimate in 1977 as a direct response to that problem. His graph works almost exactly like Edward Fry's famous readability graph — plot two numbers from a 100-word sample, read the grade level off the chart — but it replaces the syllable count with something nobody argues about: the number of long words.

That single substitution is the whole idea, and it's worth understanding what it buys you and what it costs.

How the Raygor estimate works

Take a 100-word sample of the text you want to grade. Count two things:

• Sentences per 100 words — how many complete sentences fit in the sample, counting a partial sentence at the end as a fraction. Long, winding sentences push this number down; short, punchy ones push it up.
• Long words per 100 words — in Raygor's method, the count of words of six or more letters. No syllables, no pronunciation judgment calls: you look at each word and count its letters, something two people will always agree on.

Plot the intersection of those two counts on Raygor's printed graph and the region it lands in gives an estimated U.S. grade level, from the early elementary grades up through college and professional-level text. Like Fry's chart, the graph has corner regions where results are considered invalid — combinations of sentence length and word length that don't correspond to any normal graded text.

The Raygor Readability Graph Calculator automates the plotting. It takes the first 100 words of your text, counts the long words and the (fractional) sentence count in that sample, and maps the resulting point onto an approximation of Raygor's published grade regions, returning an estimate on a scale that runs from 1st grade up to 17 — roughly "beyond college." Because the method is built around a 100-word sample, the tool needs a reasonable amount of text to say anything meaningful; give it very little and it will tell you to add more rather than guess from a fragment.

Letters versus syllables: what the trade buys

The obvious win is speed and consistency. Counting letters is mechanical; counting syllables is phonetic. When the original audience for these graphs was a teacher or librarian with a pencil, that difference meant Raygor samples could be scored in roughly half the time of Fry samples, with less variation between scorers. Studies comparing the two graphs on the same passages generally found they produce similar grade estimates through the ranges that matter most for classroom material, which is exactly what you'd hope: word length and syllable count are strongly correlated in English, so swapping one for the other preserves most of the signal.

There's a subtler benefit for software, too. Automatic syllable counting in English is genuinely error-prone — silent e's, vowel teams, and proper nouns all trip up the standard heuristics. A letter count has no such failure mode. If you've ever seen two readability tools give different Flesch-Kincaid grades for the same text, dueling syllable counters are usually the reason. A Raygor-style measure sidesteps that entire class of disagreement, which is why it remains useful when you need results that are reproducible down to the digit.

What it costs

The cost of the trade is precision at the extremes. "Long" is a blunter instrument than "many-syllabled": strength is a long word with one syllable, while idea is a short word with three. Texts with unusual vocabulary — technical writing full of short jargon, or children's writing full of long-but-familiar compound words — can drift a grade or so away from where a syllable-based measure would put them. In practice the two approaches agree far more often than they disagree, but when they do disagree, it's worth looking at the actual words driving the count rather than trusting either number blindly.

Like every mechanical readability measure, Raygor also has no idea whether a long word is familiar. "Wonderful" and "phlebotomy" both clear the six-letter bar, but only one of them slows a young reader down. If word familiarity is the thing you actually care about, pair the graph with the Dale-Chall readability calculator, which checks words against a list most fourth-graders know rather than measuring their length.

Raygor or Fry?

If you're choosing between the two graphs, the honest answer is that for most English prose they'll tell you nearly the same story, and the choice comes down to what you're optimizing for. Choose the Fry graph when you want the more established measure — it's the one textbook adoption committees historically standardized on. Choose Raygor when you want the more objective count, when your text has vocabulary that confuses syllable counters (heavy on names, acronyms, or technical terms), or when you're comparing many samples and want to eliminate counting noise between them.

Better yet, don't choose. Because both estimates are cheap to compute, running them side by side is a useful sanity check: when they agree, you can trust the grade band; when they diverge, the disagreement itself tells you something about your text — usually that its word-length profile and its syllable profile have come apart, which is exactly the kind of vocabulary quirk worth knowing about. A readability consensus score formalizes that idea across several formulas at once.

Try it on your own writing

Paste a passage into the Raygor Readability Graph Calculator and look at the two counts it reports, not just the grade. Many long words with short sentences means your difficulty is coming from vocabulary; few long words strung into marathon sentences means it's coming from structure. The grade level is the headline, but the two axes are the diagnosis — and knowing which one is driving your score tells you exactly what kind of edit will bring it down.