FORCAST: the readability formula that ignores sentences entirely

Almost every readability formula leans on one assumption: that your text is made of sentences. Flesch, Gunning Fog, SMOG, ARI — all of them divide by sentence count somewhere in their arithmetic. So what happens when you point them at a job application form, a multiple-choice test, a parts list, or a set of safety instructions written as bullet points? The sentence count is meaningless, and the grade level they hand back is garbage. FORCAST is the formula built for exactly that situation: it never counts a sentence at all.

This piece explains how FORCAST pulls that off, why counting single-syllable words turns out to be a surprisingly good proxy for difficulty, and the specific kinds of text where it's the right call over the famous formulas.

Where FORCAST came from

FORCAST was developed in 1973 by Caylor, Sticht, and colleagues for the US military, which had a very particular problem: it needed to grade the readability of job-related reading material that often wasn't prose. Forms, technical manuals, questionnaires, and tests are full of fragments, headings, and lists — text with no dependable sentence structure. The researchers wanted a formula that would still produce a sensible grade level on that kind of material, and that an enlisted clerk could compute by hand. The name is a loose acronym, and the design brief was blunt: drop sentence length, keep only what you can count reliably.

The formula

FORCAST works from a 150-word sample and a single measurement: how many of those words have exactly one syllable. Our FORCAST calculator takes the first 150 words of your text (scaling proportionally if you give it fewer), counts the single-syllable words, and applies:

grade = 20 − (single-syllable words per 150 ÷ 10)

That's the whole thing. No sentence count, no syllable averaging, no character tallies — just a count of one-syllable words. The logic is inverse: the more single-syllable words your sample contains, the easier it is, and the lower the grade. A sample with 150 single-syllable words (every word monosyllabic) would score 20 − 15 = 5; a dense technical sample with only 100 single-syllable words scores 20 − 10 = 10. Fewer simple words means harder text means a higher grade.

The output is a US grade level, read the same way as the other formulas: 8 is eighth-grade reading, 12 is late high school, and the general-audience comfort zone sits around grades 7–10. Because the formula has no sentence term, the achievable range is naturally narrow — FORCAST rarely returns very low or very high grades, which is a deliberate consequence of measuring only one thing.

Why single-syllable words are a good proxy

Counting one-syllable words sounds crude, but it leans on a robust pattern in English: the core, high-frequency vocabulary everyone knows — the, run, work, fast, good, get — is overwhelmingly monosyllabic, while specialized and abstract vocabulary tends to pile on syllables (specialized, abstract, vocabulary). So the proportion of one-syllable words tracks how much plain, everyday language a passage uses. FORCAST doesn't need to know which words are hard; it just needs to know how many are obviously easy, and it infers the rest.

That single-measurement design is also what makes it sentence-agnostic. Because the formula never asks "where does this sentence end?", it doesn't matter whether your text is flowing paragraphs or a column of checklist items. Feed it a numbered list of instructions and it produces the same kind of grade it would for prose — which is precisely the point.

When to choose FORCAST

Use FORCAST when your text isn't built from normal sentences. Application forms, intake questionnaires, multiple-choice and fill-in-the-blank tests, equipment instructions, recipe steps, slide bullets, and UI copy all confound sentence-based formulas but sit comfortably within FORCAST's wheelhouse. If you've ever watched Flesch–Kincaid swing wildly on a bulleted document because it can't find reliable sentence boundaries, FORCAST is the steadier instrument.

It's also a fine sanity check on ordinary prose when you specifically care about vocabulary load rather than sentence complexity — though for prose you usually have better-calibrated options.

Where FORCAST goes wrong

• It's blind to sentence length. The same trait that makes it work on forms means it completely misses run-on sentences. A passage of monstrously long sentences built from simple words will score as easy — because by FORCAST's only measure, it is. On flowing prose, that blind spot is a real weakness, so pair it with a sentence-aware formula.
• The output range is compressed. With one input and no sentence term, FORCAST won't distinguish a children's book from a comic strip very well at the low end, or two dense technical manuals at the high end. It's a coarse instrument by design.
• It depends on syllable counting. Deciding whether a word is single-syllable means estimating pronunciation, so proper nouns and invented words can be miscounted — the same caveat that applies to every syllable-based formula.
• It's English-only. The monosyllabic-word heuristic was calibrated on English and doesn't carry to other languages.

Putting it to use

Reach for FORCAST when you're checking the readability of text that isn't really sentences — forms, tests, instructions, lists, and other non-narrative material where the sentence-based formulas can't get a foothold. Paste a representative 150-word stretch into the FORCAST calculator, read the grade it returns, and aim for the 7–10 band for a general adult audience. Because FORCAST deliberately ignores sentence length, it pays to pair it with a sentence-aware measure on anything resembling prose: Flesch–Kincaid for the standard grade, or a readability consensus that averages several formulas so no single measure's blind spot steers you wrong.