Picture a man chasing a thief down a narrow alley. He is winning — the thief is slower, stumbling, almost in reach. The closer the prize gets, the harder the chaser runs, and the less he looks at the walls on either side. He does not see that the thief is not slow. The thief is leading. And in the doorways the chaser blew past without a glance, men are waiting. The chase did not weaken the enemy's army. It weakened the enemy's attention, and his own hunger held the leash. He ran himself into the trap, and every stride he gained on the bait was a stride away from the men who might have saved him.1
That is the oldest unorthodox move in the Chinese record. You do not break the enemy's army. You let his appetite break it for you. The ambush and the feigned retreat are one weapon with two parts: the retreat baits the greed, and the greed delivers the army to the trap.
A Ming-dynasty text, Essential Strategies from a Grass Hut, calls the ambush the very Tao of deceit: "those who excel at ambushes can be expected to be victorious, those caught in them to suffer defeat."1 That is a strong claim — ambush is not a deceptive tactic but the model of deceptive tactics. And Sawyer's documentary point reinforces it: though ambushes were used constantly from earliest antiquity, they were considered unorthodox throughout the whole of Chinese military history. The move never normalized. It stayed deviant for three thousand years.
The feigned retreat is the ambush's twin. An ambush alone needs the enemy to walk into a fixed spot. The feigned retreat drives him there. You show him a fleeing, beatable force, and his desire to finish it pulls him off his order, off his guard, and onto the prepared ground. The two halves compose into a single machine: a lure that moves and a jaw that waits.
But Sawyer's deeper claim — the one that separates this page from a generic "set a trap" tactic — is about where the vulnerability lives. The ambush does not primarily exploit terrain. It exploits the enemy's internal bonds. Specifically: greed plus the absence of mutual rescue. An army whose men will not save each other is already half-destroyed; the ambush just collects on a weakness the enemy carried in with him.
Read the first recorded case closely and the mechanism is laid out almost as a formula. The Duke of Cheng, facing the Northern Jung in 714 BCE, is worried about a physical problem: the Jung are infantry, his force is chariots, and infantry can maneuver around behind chariots and strike from the rear.1 That is a terrain-and-mobility fear. Prince T'u's answer ignores the terrain entirely and goes after the enemy's character.
His reading of the Jung: "They act hastily and are not well ordered. They are greedy and lack close feelings. In victory they will not yield to each other, in defeat will not rescue each other."1 Every clause is a bond the Jung do not have. No order. No restraint. No generosity in victory. And the decisive one — no rescue in defeat.
The plan falls straight out of that diagnosis. Send "courageous but irresolute men" to provoke the Jung and then flee. The greed pulls the Jung vanguard forward fast and ragged, chasing gain. When the vanguard hits the three waiting ambushes and runs, the rear will not come for them — because they never rescue each other. So the conflict cannot be sustained. The front is isolated and destroyed before the back even decides to care.
This is the engine: the ambush converts an internal defect into a fatal one. Greed gets the front of the column committed past the point of return. Absence of rescue guarantees the back will not unwind the commitment. The terrain — the three ambush sites — only finishes what the enemy's own missing bonds set up. Change the enemy's character and the same terrain fails. The vulnerability was never in the ground. It was in the relationships inside the pursuing army.
The ambush-and-retreat is a node in the broader unorthodox grammar.
Sawyer gives two Tso Chuan cases, and laying them side by side shows the concept maturing from a single ambush into a sustained campaign of manufactured contempt.1
714 BCE — the Northern Jung in Cheng (the pure ambush). The full sequence: the Jung invade, the Duke fears their infantry will loop behind his chariots, Prince T'u prescribes the provoke-flee-and-ambush plan keyed entirely to the Jung's greed and lack of mutual rescue. The Duke heeds it. The Jung advance forces hit the ambushes and run. Chu Tan pursues, isolates them, and strikes front and rear at once — "utterly destroying them." The remaining barbarian contingents flee in haste. Clean confirmation: every prediction in Prince T'u's diagnosis came true, because the diagnosis was about the enemy's bonds, not his numbers.
Sawyer pairs this with a second, smaller Cheng case — a coalition force moving east to attack Cheng in reprisal for an earlier clash at Pi. Cheng, having anticipated the line of march, ambushed and defeated them, ending the invasion. He links it to the Hundred Unorthodox Strategies' chapter on "Harm": establish ambushes at critical points along a contested border and the enemy "will not recklessly advance," because, as the Art of War says, "what can cause the enemy not to come forth is the prospect of harm." The ambush here works as more than a battle-winner; it becomes a deterrent — a standing threat that shapes whether the enemy moves at all.
611 BCE — Ch'u and the state of Yung (the campaign of manufactured arrogance). This is the richer case, and it elevates the feigned retreat from a single lure into a sustained psychological operation.
The setup: Ch'u is gutted by an extensive famine. Sensing weakness, enemies converge from every side. The Jung attack in the southwest, reaching Mount Fou. More attacks come in the southeast. Officers from Yung lead a collection of indigenous Man clans in revolt; the Chün assemble the Pai-p'u for a strike. The whole coalition is betting on one assumption: a famine-struck Ch'u cannot field an army.
Ch'u's officials want to flee — shift the capital to Fan-kao. Wei Chia kills the idea with one line: "We cannot, for wherever we can go, the invaders can also go."1 Relocation solves nothing because mobility is symmetric. His counter-proposal inverts the whole logic: do not run, attack — specifically attack Yung. His reasoning targets the coalition's founding assumption: "the Chün and Pai-p'u attacked us on the assumption that the famine will preclude our fielding an army. If we send our army forth they will certainly be frightened and withdraw. The various Pai-p'u dwell separately and will race off to their own towns."1 The same insight as Prince T'u's, one register up: the coalition has no real cohesion. It is a collection of separate clans held together only by the belief that Ch'u is helpless. Break that belief with a sudden army and the coalition disassembles itself — each clan racing home to defend its own town. Within fifteen days the Pai-p'u desisted.
Then comes the refinement that makes this case famous. Ch'u presses on toward Yung. Tzu-yang Ch'uang is captured, escapes after three nights, and reports back that the Yung are numerous and reinforced by several Man clans — he counsels falling back to the main force. Shih Shu overrules him with a different plan entirely: "When we encounter them again in a little while, let's make them arrogant. If they become arrogant while our forces are enraged, we will be able to conquer them. This was the way our ancestral ruler Fen Mao subjugated the Hsing-hsi."1
This is the engine stated outright: make them arrogant, then strike. And the method is the feigned retreat, repeated as a course of treatment. Over the next several encounters Ch'u's forces all run. Only the Pi, T'iao, and Yü bother to pursue. The men of Yung draw the intended conclusion — "Ch'u wasn't worth fighting with" — and stop making defensive preparations. The contempt has been cultivated, fed deliberately, encounter after encounter, until the enemy disarms himself psychologically.
Then the jaw closes. The earl of Ch'u races forward, assembles his army at Lin-p'in, and splits into two operational columns — Tzu Yüeh approaching Yung from Shih-hsi, Tzu Pei from Jen. Troops from Ch'in and Pa join Ch'u's main force; a number of Man clans come to accord with the earl. Ch'u extinguishes the state of Yung.
The arc is the lesson. A single feigned flight is a lure. The Ch'u campaign is a program — repeated feigned weakness used not to bait one charge but to lower the enemy's guard permanently, until "Ch'u wasn't worth fighting with" becomes a settled belief and the defenses come down. The retreat stopped being a tactic and became a slow re-education of the enemy's estimate.
You are weaker, and the enemy knows it — that is the fact you cannot argue away, so you stop trying to argue it. Instead you decide to feed it. When his scouts come, your men break and run, sloppy, unconvincing-looking, leaving gear behind. He pursues a little; you let him have the small win. Next week, again — your column scatters at the first push, your banners go down, your courageous-but-irresolute men make a show of panic and bolt. He pursues farther this time, more careless, and again you give it to him. You are not losing. You are teaching. Encounter by encounter you watch the thing you want take root in his camp: the word goes around his fires that you are not worth fighting, that the famine broke you, that there is no army here worth preparing against. You can see it in how his pickets thin, how his clans start drifting back to their own towns, how the guard simply comes down. And the night his contempt is total — when he has stopped preparing because you have personally convinced him there is nothing to prepare for — your enraged main force, held back and concealed and burning to move, splits into two columns and comes from two directions at once into a man who has filed you under harmless. The trap was never the ground. The trap was the belief you spent a month planting in his head.
The move fails in recognizable ways, and the failures are instructive because they invert the engine.
The evidence is Sawyer's reading of the Tso Chuan's earliest battle narratives, cross-referenced against later theoretical texts — the Essential Strategies from a Grass Hut, the Hundred Unorthodox Strategies, the Art of War.1 The 714 BCE and 611 BCE cases are well-attested set pieces; the framing of them as "unorthodox" is, as with the animal cases, a retrospective application — the Tso itself never uses the category.
The central tension Sawyer surfaces — and it is a real one — is doctrine against practice. The feigned retreat is among the most prized unorthodox moves in the tradition, and the same tradition's foundational text warns against falling for it: do not pursue feigned retreats, do not swallow the bait army. So the move is simultaneously the thing the smart commander deploys and the thing the smart commander fears. This is not a contradiction to resolve. It is the natural shape of any deception that becomes well-known: the more famous the trap, the more the prey is taught to suspect it, which raises the skill required to spring it. The tactic and its counter co-evolve. Sawyer presents Sun-tzu's warning as the initiation of consciousness of this problem in the military writings — the moment the tradition started teaching its members to distrust the very move it prized.
Flag Sawyer's standing bias. His project argues for a distinctively Chinese sophistication in unorthodox, deception-centered warfare, and a tactic this old, this theorized, and this self-aware is excellent evidence for that thesis. The corrective question: ambush and feigned retreat are near-universal in early warfare — Hannibal at Lake Trasimene, countless steppe and forest peoples. What Sawyer is really claiming is not that China invented the move but that China theorized it earliest and most explicitly. That narrower claim is more defensible, and the reader should hold him to the narrower one.
The page is single-source, but Sawyer is again arbitrating among his own texts, and the sharpest pairing is internal to the tradition: the Tso Chuan practitioners versus Sun-tzu. Prince T'u and Shih Shu do the feigned retreat with total confidence — it is their best weapon. Sun-tzu, codifying a century-plus later, warns against it — treats the chasing of a fleeing enemy as a discipline failure that gets armies swallowed. Both are right, and the gap between them is the whole story. The field commanders are speaking as the predator who has found a reliable kill. Sun-tzu is speaking as the theorist who has watched too many predators become prey. The tension reveals what neither states alone: a deception's value is inversely proportional to its fame, and the feigned retreat is so good and so old that by Sun-tzu's day defending against it had become as important as deploying it. The convergence underneath: both sides agree the decisive variable is the enemy's self-control under the prospect of gain. Prince T'u exploits its absence; Sun-tzu prescribes its presence. They are describing the same hinge from opposite seats.
Strip away the chariots and the Man clans and the move is bare: you offer an adversary a cheap-looking win, his desire for it pulls him off his discipline, and the position he abandons to chase it is where you destroy him. The bond exploited is his own greed; the asset spent is something you were willing to give up. That structure is not military. It runs anywhere two minds compete and one can be made to overreach.
Cross-domain (Japanese strategy): Sutemi — Sacrifice Strategy is the same logic in one body. Sutemi is the deliberate offering of something real — your balance, your position, an apparent opening — as bait, accepting genuine risk to draw the opponent into a committed attack you have already prepared to answer. The "courageous but irresolute men" sent to provoke the Jung and flee are sutemi at army scale: a force deliberately exposed and apparently sacrificed to pull the enemy onto prepared ground. The parallel unlocks what the military case states only implicitly — that the bait must carry real cost to be convincing. A fencer's feigned opening that risks nothing is read instantly as a feint; the provocateurs who flee must actually be in danger, or the Jung smell the trap. The insight neither field generates alone: the credibility of a lure is purchased with genuine vulnerability, which is why the safest-looking deceptions fail and the ones that cost something work. You cannot bait greed with a free gift — the greedy mind, of all minds, knows a thing of value is never free.
Cross-domain (Japanese strategy): Sen — Initiative and Attention Gaps explains the timing. Sen is the seizing of initiative through the gaps in an opponent's attention — the instant his focus narrows onto one thing, the moment he commits. The pursuing Jung vanguard, eyes locked on the fleeing prize, has opened exactly such a gap: total attention forward, none to the flanks where the ambushes wait. The feigned retreat manufactures the attention gap that sen exploits — the lure does not merely move the enemy, it narrows his perception to a single point so that everything off that axis goes unwatched. The insight: greed acts as more than a motive — it is an attention-collapsing force; the chaser does not see the doorways because desire has spent his whole field of attention on the back of the fleeing man. The ambush works in the perceptual blind spot that the bait itself drilled open.
Behavioral-mechanics: Deceptive Positioning is the human-influence version of the 611 BCE campaign. Cultivating a target's contempt — letting yourself be underestimated, repeatedly, deliberately, until the target stops defending against you — is exactly Shih Shu's "make them arrogant" program transposed from war to maneuver among people. The Ch'u campaign adds what the influence framing often leaves implicit: that contempt is built over time, encounter by encounter, not produced in a single move. The men of Yung did not decide in one day that "Ch'u wasn't worth fighting with" — they were taught it, run after run, until the belief was load-bearing enough that their defenses came down. The insight: the most powerful deceptive position is not a single false impression but a durable false estimate patiently installed, because the target who has concluded you are harmless will rationalize away the evidence when you finally move. By then his contempt is doing your concealment for you.
The Sharpest Implication If the ambush's real target is the enemy's internal bonds — greed plus no mutual rescue — then the tactic is a diagnostic instrument before it is a weapon. You cannot decide to spring this trap. You can only spring it on an enemy who carries the defect. Prince T'u did not choose the feigned retreat and then hope the Jung were greedy; he read the Jung's missing bonds first and derived the tactic from the defect. This inverts the usual order of planning. You do not pick your move and look for an enemy to use it on. You read the enemy's character, find the bond he lacks, and let the missing bond name the move. Which means the deepest skill in this tradition is not setting traps — it is seeing what an enemy cannot do for himself, and the commander who cannot read internal bonds is firing his best weapon blind.
Generative Questions