The Thirty-Six Stratagems are usually introduced as ancient Chinese military history: thirty-six moves, each earned on a battlefield, handed down for two thousand years. That framing is mostly right. Where it’s wrong is where people quote it most.
We publish a page for every one of the thirty-six, and each page states where its stratagem comes from. Reading all thirty-six of those origin texts end to end — which nobody has reason to do, because a reader arrives at one page and leaves from it — the collection turns out to rest on four quite different kinds of ground.
What twenty-seven of them look like
Twenty-seven of the thirty-six name a dated episode with named people and present it as fact. These are not vague gestures at antiquity. 围魏救赵, besiege Wei to rescue Zhao, is the Qi strategist Sun Bin advising against marching to the besieged city and striking the enemy's own capital instead — 353 BCE. 关门捉贼, shut the door to catch the thief, is Bai Qi cutting the Zhao army off from its supply line at Changping in 260 BCE. 假道伐虢 is Duke Xian of Jin buying passage through Yu with horses and jade in 655 BCE, then taking Yu on the way home.
That last one is worth pausing on, because the name and the event are the same thing: the stratagem is called "borrow passage to conquer Guo" because someone borrowed passage and conquered Guo. When a stratagem works that way, its title is a citation.
Four of them are named from literature, not war
Four are not citations at all. Their names come from poems and anecdotes, and a battle was attached to them afterwards.
擒贼擒王 — to catch the bandits, first catch their king — is a line from a poem by Du Fu, written in the eighth century, roughly a millennium after the Warring States campaigns the collection mostly draws on. 李代桃僵, the plum dies in place of the peach, comes from an old folk poem about a plum tree whose roots are eaten by insects while the peach beside it survives. 抛砖引玉, toss a brick to draw jade, comes from a story about two Tang poets and a wall at a temple. 打草惊蛇, beat the grass and startle the snake, is a marginal note written by a corrupt Tang magistrate whose clerks had been accused of graft — a bureaucratic joke, with no soldiers in it whatsoever.
All four have perfectly good military illustrations on their pages. But the illustration arrived second. The name came from letters.
Two are proverbs with no incident behind them
顺手牵羊 — to lead away a goat that comes to hand — and 混水摸鱼, to fish in muddied water, claim no anchoring event and never did. They are proverbs describing a habit, promoted into a list of stratagems because the habit is strategic. Our own pages say so plainly: one is "proverbial rather than tied to one battle", the other simply "proverbial".
Two more hedge. 瞒天过海 opens by admitting it has "no single verified battle", offering instead a much-retold Tang story about screening an emperor from the sight of the sea. 欲擒故纵 rests on Zhuge Liang capturing and releasing the chieftain Meng Huo seven times — an episode our page introduces with "is famously said to have"—the careful writer's phrase for a strong tradition with weak evidence.
The final tally for the collection: twenty-seven documented episodes, four names from literature, two proverbs, two hedged traditions, and one that historians rejected.
And the most famous one is the least supported
Then there is 空城计, the empty fort. Zhuge Liang, hopelessly outnumbered, throws open the gates of Xicheng, sits on the wall playing his qin, and the suspicious Sima Yi withdraws rather than walk into what must be a trap. It is the best-known stratagem in the collection and one of the best-known scenes in Chinese literature.
It is also the one entry whose famous telling a named historian examined and rejected: Pei Songzhi, the fifth-century annotator of the Records of the Three Kingdoms, looked at the claim and dismissed it as implausible. The version most readers know comes from the Romance of the Three Kingdoms, a novel written more than a thousand years after the events it dramatises.
The interesting part isn't whether the stratagem works—it does—but where its celebrity comes from. The same page offers a soundly historical instance: the Han general Li Guang, caught far from support by a large Xiongnu force with a hundred horsemen, ordered his men not to run but to dismount and unsaddle. The Xiongnu, certain that no one would do that without an ambush waiting, withdrew.
So the empty fort is real and Li Guang did it. Zhuge Liang on the wall with his qin is the part that comes from a novel. If you cite the stratagem, you are on firm ground; if you cite the scene, you are citing fiction that a historian had already declined to believe.
What this count does and does not license
A tidy number invites overclaiming, so let's be plain about the limits of this count.
This measures our pages, not the historical record. "Twenty-seven of thirty-six name a dated event" is a fact about what those texts assert. It is not a finding that twenty-seven stratagems are historically verified — establishing that would take source criticism of each episode, weighing the Records of the Three Kingdoms against the Zuo Commentary against later compilations, which is a scholar's job and not this one.
The count is still useful, for a slightly unusual reason: it works because the pages are honest about their own sources. A page that opens "this stratagem has no single verified battle" — as 瞒天过海 does — is doing the work that makes counting possible at all. If every page had asserted a battle with equal confidence, there would be nothing to tally and no way to know it.
What to do with it
Use the collection the way it is actually built. Twenty-seven entries give you a documented case you can look up and reason from. Four give you a phrase from a poem and a battle chosen later to fit it, which is fine as long as you know which part is load-bearing. Three rest on tellings the sources themselves hedge or reject.
The practical advice is simple. When a stratagem is doing real work in an argument—a board paper, a negotiation, a strategy deck—check which of those four kinds you’re leaning on. The move can be sound while the story is decorative. What you should not do is put the empty fort in a slide as something that happened.
One footnote from the exercise. Reading all thirty-six origins in sequence turned up two dates on our own pages that were wrong: one gave the Battle of Guandu as "200 BCE... 200 CE", and another attached 207 CE to that battle when 207 belongs to a later episode and Guandu was 200. Both were English-only — the Chinese was correct in both cases. Both are now fixed. Nothing catches an error like that except reading the whole thing in one pass, which is the same reason the count exists.
