Flesch-Kincaid Explained: What the Score Means and How to Use It

If you’ve ever pasted a draft into a “readability checker” and watched a number appear, that number was almost certainly a Flesch-Kincaid score. It powers Microsoft Word’s readability stats, the Hemingway Editor, the Yoast SEO plugin, and most online “grade level” tools. Despite being old enough to drink in any country, it remains the default measurement of how hard your sentences are to read.

This post unpacks what the score actually computes, why it caught on, the cases where it gives the right answer, and the cases where it lies.

The formulas, demystified

Flesch-Kincaid is two formulas, not one. They share an input and disagree about what to optimize for.

Flesch Reading Ease rates text on a 0-100 scale where higher means easier:

206.835 - 1.015 * (words / sentences) - 84.6 * (syllables / words)

A score of 100 means a child could read it; a score of 30 means a grad student would slow down. Most popular fiction lands between 70 and 90. Insurance contracts dip into single digits.

Flesch-Kincaid Grade Level maps the same inputs to a US school grade:

0.39 * (words / sentences) + 11.8 * (syllables / words) - 15.59

A score of 8.0 means an average eighth-grader can follow it. Newspapers target around 7-9. Academic papers float in the 14-18 range.

Both formulas hinge on two ratios: average sentence length and average syllables per word. Long sentences and long words make scores worse. That’s the entire model.

If you want to see your own writing scored against both formulas at once, paste it into the Readability Checker . It runs the calculation locally and shows the same numbers Word would, plus a few related metrics.

Why this specific formula won

There are a dozen other readability formulas (Gunning Fog, SMOG, Coleman-Liau, ARI, Dale-Chall) and most of them agree most of the time. Flesch-Kincaid won because:

1. The US Navy adopted it in 1975 for technical manuals. That gave it a federal mandate, which trickled into education and contract law.
2. It only needs syllable counts and sentence breaks, both of which were easy to compute on a 1980s desktop.
3. The grade-level output is intuitive. Telling an editor “aim for grade 8” is a more specific instruction than “score 65 on a 0-100 axis.”

The result is that a lot of style guides quietly assume Flesch-Kincaid. The US Department of Defense, the IRS, and many state plain-language laws cite it explicitly.

What a good score looks like

A few rough benchmarks, gathered from public corpora:

• The Bible (KJV): grade 12, ease 70.
• Reader’s Digest: grade 8, ease 65.
• Time magazine: grade 11, ease 52.
• Scientific American: grade 13-15, ease 35-50.
• A typical software EULA: grade 16+, ease under 30.

For most public-facing writing, a Flesch-Kincaid grade between 7 and 9 is a safe target. Below 6 starts to feel choppy; above 11 starts to lose general readers.

If you’re writing for a technical audience, push higher: grade 11-13 is normal for engineering blog posts and not a problem. The issue isn’t difficulty, it’s whether the difficulty matches the reader.

Counting syllables is harder than it looks

The biggest source of disagreement between readability tools is how they count syllables. There’s no perfect algorithm for English. Common heuristics:

• Count vowel groups, then subtract a silent e at the end.
• Treat consecutive vowels (oa, ea, ou) as one syllable.
• Hard-code exceptions for common suffixes (-ed, -es, -ing).

Even with all of that, “fire” might count as 1 or 2 syllables depending on the dialect, and “really” as 2 or 3. Most readability scorers settle on a quick approximation that’s wrong by ~3% on average and never bother to be more accurate, because the 3% noise rarely changes the grade-level conclusion.

This is why a single paragraph can get slightly different scores across tools. The formula is identical; the syllable counter is what diverges.

What the score is good for

• Sanity-checking long passages. If a 600-word section comes in at grade 14, something is off (probably a wall of long sentences) and rewriting will help.
• Comparing two drafts. Same writer, same topic, different versions: a 2-grade improvement is real.
• Hitting a regulatory target. Some US states require consumer contracts under grade 8. Some federal agencies require plain language scoring above ease 60.
• Setting expectations. A blog post at grade 12 tells you “this is for a technical audience.” That alone is worth communicating to your editor.

What the score is not good for

• Comparing different writers. Flesch-Kincaid rewards short sentences and short words. A skilled writer using domain vocabulary correctly will look “harder” than a sloppy writer using simple words to say nothing.
• Comparing different languages. The constants (1.015, 84.6, etc.) were trained on English. The formulas don’t transfer to German or Japanese in any meaningful way.
• Judging clarity. “Lorem ipsum dolor sit amet” scores well. Real garbage can hit grade 6 if it’s short enough.
• Optimizing for SEO. Google has said many times that readability is one signal among many. Hitting grade 8 will not magically rank you. Hitting grade 18 won’t keep you out of the index.

A practical workflow

If you write for a general audience and want a sane workflow:

1. Draft normally. Don’t think about scores.
2. Run the finished draft through a readability checker.
3. If the grade is more than 2 above your target, find the longest sentences and split them. That’s almost always the fix.
4. If the grade is right but the prose feels lifeless, ignore the score. The number was a tool, not a verdict.

For step 2, the Readability Checker is a good starting point because it shows several scores side by side. Different formulas weighting the same text differently is informative; if Flesch-Kincaid says grade 12 and SMOG says grade 8, look at why.

For step 3, you may also want a word counter so you can target specific paragraphs without re-pasting the whole document.

A short history

The original “Flesch Reading Ease” was published by Rudolf Flesch in 1948, in a book called The Art of Readable Writing. He was reacting to a generation of academic prose that had become genuinely impenetrable. His goal was simple: a single number that would tell an editor when a sentence was too long.

The “Kincaid” half was added in 1975 by Peter Kincaid, working with the Navy’s Personnel Research Center. The Navy needed to know whether enlisted sailors could read their training manuals. The grade-level reformulation made the score actionable: “aim for grade 9” was a sentence a writing instructor could give.

That dual origin is part of why the score feels split between casual and bureaucratic uses. It is, equally, a tip from a 1948 magazine editor and a 1975 Navy contract spec.

The bottom line

Flesch-Kincaid is a reasonable cheap heuristic for “are my sentences too long and my words too dense?” It is not a measure of quality, persuasiveness, or correctness. Use it the way a writer would use a stopwatch: to confirm a hunch, not to set the artistic direction.

If your writing scores well and reads poorly, the score lied. If it scores poorly and reads well, the score lied differently. The score is a tool; the writing is the thing.