How to Count Chinese Character Strokes (and Why the Number Matters)

When I started learning Chinese, the first thing that broke my brain was that a character is not a picture you sketch freehand. Every Chinese character is built from a fixed number of strokes, and a stroke is a single continuous pen movement — you put the pen (or brush) down, drag it in one motion, and lift it off. That count is not decoration. It is the address a character lives at in old dictionaries, the rhythm a teacher uses to grade handwriting, and the number a grandparent checks before a baby gets named. Once I understood that the stroke count was a real, agreed-upon integer for each character, a lot of Chinese stopped feeling random.

This guide explains what counts as one stroke, why the number gets used everywhere, and the one trap that catches almost every beginner: simplified and traditional forms of the "same" character often carry completely different counts.

What counts as one stroke

A stroke is one unbroken movement of the writing instrument. If your pen never leaves the paper, that is one stroke, no matter how much it bends partway through.

That last part trips people up. Take the character 口 (mouth). It looks like four straight lines forming a box, so you would guess four strokes. It is actually three. The top and the right side are written as a single bent motion — down and then across without lifting — so that corner counts once. The teaching standard groups these motions into a handful of basic stroke types: the horizontal (一), the vertical (丨), the left-falling and right-falling diagonals (丿 and ㇏), the dot (丶), and the various "turning" strokes where one continuous movement changes direction.

A few rules that clear up most confusion:

A bend does not add a stroke. The horizontal-then-vertical hook in 口 or 国 is one stroke, not two.
A hook on the end of a line (the little flick in 小 or 丁) is part of the stroke it ends, not a separate one.
Modern counting uses one rule everywhere: each visible written motion is one stroke. Older calligraphy manuals sometimes split "compound" strokes like 乙 into several, but the unified convention used in textbooks and dictionaries today counts it as one.

If you want the canonical drill, the character 永 (forever) is the classic teaching example — it contains eight distinct stroke motions in five strokes' worth of types, and Chinese students practice "the eight principles of 永" precisely because it packs the core shapes into one small character. 永 is 5 strokes.

Why the stroke count actually matters

Stroke count is not trivia. It is load-bearing in at least four places.

Dictionary lookup. Before phonetic input, you could not type a character you did not know how to pronounce. So paper dictionaries are organized by radical and then by residual stroke count: you find the character's radical, then scan the sub-list ordered by how many strokes remain. A 《新华字典》 radical index is unusable if you cannot count. Even modern handwriting-recognition apps fall back on stroke order and count when they are unsure.

Writing practice. Teachers grade beginners partly on stroke count and stroke order, because a character written with the wrong number of strokes is wrong even if it looks roughly right. When I build practice sheets, I sort the week's characters from light to heavy — 人 is 2 strokes, 我 is 7, 警 is 19 — so the hardest ones never bunch up on one day.

Name numerology. This is the big one in Chinese-speaking culture. In traditional 姓名学, the stroke counts of a surname and given name combine into five "grids" (天格, 人格, 地格, 外格, 总格), and each total maps to a fortune reading. Whether or not you believe a number can be lucky, this system is consulted constantly for baby names and brand names across the mainland, Taiwan, Hong Kong, Korea, and Japan. Swap one given-name character for another with a different count and the whole reading changes.

Font and signage design. A character with 23 strokes and one with 7 strokes do not behave the same when you shrink them. A heavy character can turn into an ink blob at a 40px favicon size or dominate a horizontal plaque. Designers count strokes to predict which characters will hold up small.

Simplified versus traditional: the count is not the same

Here is the trap. People assume that 體 and 体 are "the same word in two outfits," so they must have the same stroke count. They do not. They are, for counting purposes, two different characters.

Worked example. The word for "body" is tǐ:

Traditional 體 — 23 strokes. The left side is 骨 (bone) and the right is 豊, and all of it has to be written.
Simplified 体 — 7 strokes. The simplification replaced the whole right-and-left structure with 亻 (person) plus 本 (root).

That is a 16-stroke difference for one syllable. If you are checking a name against the five-grid system, plugging 体 into a calculation that assumes the traditional form will hand you a wrong total and a wrong reading.

It is not always that dramatic, but the principle holds across the simplified set. 龍 (dragon, traditional) is 16 strokes; 龙 (simplified) is 5. 學 (study) is 16; 学 is 8. The number always belongs to a specific written form, not to the abstract meaning. So the first question before you count is always: which form is this text actually using?

A second, subtler source of disagreement is regional convention. A handful of characters, like 那, are broken down slightly differently in mainland, Taiwan, Hong Kong, and Japanese standards, so two honest dictionaries can list different numbers for the same glyph. The fix is to pick one standard and stay consistent — mainland convention for simplified, Taiwan's 教育部 standard for traditional — rather than mixing sources mid-name.

Counting a whole passage without going cross-eyed

Counting one character by hand is fine. Counting a 60-character classical excerpt to budget your copying practice is not — you will miscount, and you will count the punctuation by accident, which quietly inflates your average. This is exactly the chore I built a tool for. The Chinese Stroke Counter takes any pasted text — simplified, traditional, mixed, or interleaved with English and punctuation — and returns the per-character count, the total, and the average, with only CJK ideographs included in the math so your average is honest.

It keeps the two forms separate exactly as described above: 體 returns 23 and 体 returns 7, never one silently mapped to the other. For a rare variant it does not have authoritative data on, it returns "?" rather than inventing a number — and you should treat that "?" as "go look this one up by hand," not as a zero, especially inside a five-grid total.

If your text mixes traditional and simplified and you want to normalize it before counting, run it through the Traditional / Simplified Converter first, then count the version you will actually write. Pairing the two means the number you read always matches the characters in front of you — which, after all the rules above, is the only number worth trusting.

Stroke count is one of those quiet fundamentals: invisible once you have it, confusing until you do. Learn what counts as one stroke, respect the simplified-versus-traditional gap, and the rest of the writing system gets a little less mysterious.

Made by Toolora · Updated 2026-06-13