Stratagem 13 · Attacking Stratagems

打草惊蛇

dǎ cǎo jīng shé

Beat the grass to startle the snake

A deliberate probe to flush out what's hidden — or a warning against the careless move that does it by accident. Poke the grass and watch which snakes rise; but never poke it when surprise was your only advantage.

Reeds parting in a sudden line as something unseen moves away. Artist’s impression, drawn for RECATOOLS. No illustration of the Thirty-Six survives from antiquity — the collection itself only surfaced in the 1940s. How the Thirty-Six reached us →

The principle

This stratagem has two faces. As a probe, you make a small, deliberate disturbance — a feint, a test order, a leading question — to provoke a reaction that reveals a hidden enemy's position, strength, or intent before you commit. As a caution, it names the opposite error: acting rashly and alerting a quarry you could have taken by surprise. The same act is wisdom or blunder depending on whether concealment was your asset. Decide first what surprise is worth to you, then choose whether to disturb the grass at all.

Historical origin

The idiom comes from a Tang-era anecdote about a corrupt magistrate, Wang Lu. When his own clerks were accused of graft, he wrote in the margin of the complaint: 'You have merely beaten the grass, but I am already the startled snake' (汝虽打草，吾已惊蛇) — the exposure of his subordinates had warned him. The phrase came to mean both an accidental alarm and, in strategy, a deliberate one used to make a hidden enemy show himself. Recorded in the Tang collection Youyang Zazu.

Applied today

Business

To learn a competitor's plans, run a small public test — a limited launch, a floated price — and read their reaction. Their scramble to respond reveals what they were quietly protecting.

Everyday

The caution matters more day to day: don't tip your hand. Asking pointed questions or making threats before you're ready only warns the other side to cover the very thing you were after.

In practice

Invented scenarios showing this stratagem at work — illustrations, not reports of real events.

AI & Technology a red-team review of a public assistant

Patched before the paper

A researcher posted a thread of working jailbreak prompts a fortnight before submitting the full write-up. The behaviours were patched within days. When the paper appeared it described a system that no longer existed, and the release notes credited nobody. The disclosure had been generous and early, which was exactly what made it useless as evidence.

Use it Capture reproducible evidence and agree a disclosure timeline before demonstrating a flaw publicly — a patched system cannot be studied.

Business & Strategy a two-person analytics startup

The question that gave it away

Hoping to secure distribution, the founders asked a much larger platform whether it would consider exclusivity in a category the platform had not yet entered. The answer was a polite no. Nine months later the platform shipped that category itself. The question had contained the whole plan, and it had been put to the one party best placed to act on it first.

Use it Work out what a question reveals before asking it — probing early can hand your plan to whoever can move faster.

Career & Work an analyst considering a move

Mentioned six months early

Long before he was ready to act, an analyst told a friendly manager that he was thinking about a change. Nothing hostile followed. He was simply left off the next long project, then the one after that, on the reasonable grounds that continuity mattered. The disclosure had been honest, and its effect was a very quiet year.

Use it Signal a move only once you can act on it — an early disclosure changes how everyone else plans around you.

Cybersecurity a penetration test at a payments company

The scan that warned them

A tester began with a broad, noisy sweep of the network. Alerts fired within minutes, the response team rotated credentials, and every subsequent attempt met a hardened target. The report that followed was thin. The lesson was recorded for the defenders instead: their detection worked, and it worked because someone had disturbed the ground before moving.

Use it Probing carelessly alerts your target — and conversely, deliberate noise is a cheap way to test whether anyone is watching.

🔤 Look up this idiom: 打草惊蛇 (成语词典)