← All 36 stratagems
Stratagem 18 · Attacking Stratagems

擒贼擒王

qín zéi qín wáng
To catch bandits, first catch their leader

Break the whole by taking its centre. Don't grind down a mass of followers — find the leader, the keystone, the source of cohesion, and remove it; the rest loses its shape and its will.

Loose arrows scattered on the ground and one banner pole standing stripped.
Loose arrows scattered on the ground and one banner pole standing stripped. 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

A large force is often held together by a small centre of gravity — one commander, one relationship, one node without which the whole loses coordination. Attacking the mass wastes strength against numbers; attacking the centre collapses the numbers for free. The work is to identify what actually holds the thing together, which is rarely the biggest or loudest part, and to reach it directly. It applies far beyond battle: the key decision-maker in a negotiation, the linchpin dependency in a system, the one belief holding an argument up. Find the keystone, and the arch falls on its own.

Historical origin

The maxim comes from a poem by Du Fu (8th c.): 'to shoot a rider, first shoot his horse; to catch the bandits, first catch their king.' A vivid application is Zhang Xun's defence of Suiyang (757 CE): unable to identify the enemy commander Yin Ziqi among his troops, Zhang Xun had his men fire stalks of straw as arrows; enemy soldiers, thinking the defenders were out of real arrows, ran to report to Yin Ziqi — revealing him. Zhang Xun's archer then shot Yin Ziqi in the eye, and the leaderless assault faltered. Recorded in the New Book of Tang.

Applied today

Negotiation

In any multi-party deal, find the one person who truly decides and win them — persuading the rest without the decision-maker is effort spent on people who cannot say yes.

Everyday

To change a system, target its keystone dependency, not its symptoms. Fix the one bottleneck, convince the one influencer, remove the single point of failure — the rest shifts on its own.

In practice

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

AI & Technology a product with dozens of AI features

One paragraph, forty-one features

Forty-one features called the model and complaints arrived about most of them. The team spent a month tuning them individually before anyone traced the calls and found thirty-six inheriting the same shared system prompt, written in a hurry a year earlier and never revisited. One paragraph was rewritten. Most of the complaints stopped, including several nobody had connected to it.

Use it Trace shared prompts and defaults before tuning individual features — the common ancestor explains more failures than the leaves do.

Career & Work an internal proposal

The one signature

The proposal was socialised with eleven people over two months, each of whom improved it slightly and none of whom could approve it. When it finally reached the person who could, she asked one question nobody among the eleven had raised, and answering it took another quarter. The consultation had been genuine and had tested nothing that mattered.

Use it Find the approver early and rehearse against their question — broad consensus is not a substitute for the one decision.

Cybersecurity a post-incident review

They went for the keys, not the doors

The review expected a slow walk from server to server. What had happened was quicker: two hours of quiet reconnaissance, then one administrator of the identity system, and after that no further breaking in was necessary. Every other control was intact and none of them mattered. The rest of the estate had been reached by asking politely with the right credential.

Use it Harden and watch the identity system first — an attacker does not enumerate your assets, they take whoever can grant them.

Negotiation a supplier pitching a hospital group

Six meetings, wrong room

A sales team ran six enthusiastic meetings over four months, each with people who liked the product and none of whom controlled the budget. A seventh meeting, arranged with the person who signed, resolved the question in twenty minutes — unfavourably, but decisively. The months were not spent persuading; they were spent not identifying who needed persuading.

Use it Find who actually decides before investing effort — enthusiasm from people without authority costs time and buys nothing.

🔤 Look up this idiom: 擒贼擒王 (成语词典)

Related stratagems

Advertisement
After content · AD-W1 Responsive