Most of Law 47 warns you against your own swelling head. This passage identifies a third motive, and it is the one that never feels like a warning sign because it feels like loyalty: "People who go past the mark are often motivated by a desire to please a master by proving their dedication. But an excess of effort exposes you to the risk of making the master suspicious of you."1
Nothing in that state resembles arrogance from inside. You are not celebrating; you are working late. You are not taking more for yourself; you are producing more for someone else. Every internal check the chapter recommends — am I elated, am I overconfident, is this my ego — returns clean.
And the outcome is the same. An excess of effort is the phrase, and it is a strange thing to be able to write in a book about power. Effort is supposed to be the safe variable, the one with no downside. This says it has a ceiling, and that the ceiling is set by someone else's comfort rather than by your capacity.
Greene's illustration is fourth-century B.C. and unusually blunt. A captain under the Chinese general Wu Ch'i "charged ahead before a battle had begun and came back with several enemy heads. He thought he had shown his fiery enthusiasm, but Wu Ch'i was unimpressed. 'A talented officer,' the general said with a sigh as he ordered the man beheaded, 'but a disobedient one.'"2
The captain succeeded. That is the part to hold onto. He did not exceed his orders and fail; he exceeded them and won, and brought back proof.
Wu Ch'i's sentence separates the two axes explicitly — talented but disobedient — and in doing so states which one is being scored. Talent was acknowledged and was not the currency. The crime was not excess and not failure. It was that the success was unauthorised.
What the captain misread is the master's actual demand. He believed he was being paid for outcomes, and produced a good one. An army is not buying outcomes from individual officers; it is buying predictability — the ability to know where a unit will be when the plan requires it. An officer who generates good results without instruction supplies uncertainty at the same time, and uncertainty is priced far above the marginal value of a few heads.
The Macedonian version is stated as policy: "On several occasions, generals under Philip of Macedon were disgraced and demoted immediately after leading their troops to a great victory; one more such victory, Philip thought, and the man might become a rival instead of an underling."3
Immediately after. The punishment tracks the victory itself, with no intervening misconduct.
Read from the subordinate's side, this is unfalsifiable. There is no conduct that avoids it, because the trigger is the achievement. Which means the instruction Greene draws — measure your victories carefully4 — has a logical terminus he does not name: at some point, measuring your victories means producing fewer of them than you could, deliberately, on purpose, and calling it prudence.
Flag that rather than smoothing it. The advice is real and the cost is real. An organisation in which subordinates ration their output to stay below a superior's threshold is losing exactly the thing it hired them for, and Greene's framing puts the entire adaptive burden on the person with less power. He does not price what the practice does to the enterprise, because the corpus's unit of analysis is one operator's survival.
🚩 SINGLE SOURCE — Greene names no instance, no general, and no source for the Philip claim. The best-documented Macedonian case runs the other way: Parmenion served through decades of victories without demotion, and was killed much later for other reasons.
The second half of Greene's instruction cuts against something the same chapter argues elsewhere: "It is also wise to establish a pattern of strict obedience to earn his trust."4
Set that beside the Keys to Power, twenty lines earlier, which says the successful "vary [patterns] at will" and that repeating what worked is the trap.5 One passage says patterns kill you; the next says build one deliberately.
Both are right, and the reconciliation is about audience rather than about patterns. A pattern is a liability with an opponent and an asset with a superior, because the two are reading you for different things. An opponent reads you to anticipate — legibility is what they extract, and it costs you. A superior reads you to decide how much rope to give — legibility is what they require, and its absence is what gets subordinates cut back regardless of results.
That resolves a genuine confusion in the chapter, and it produces a working rule: vary the moves you make against people who are trying to counter you; hold rigid the behaviours by which people who depend on you predict you. See The Pattern That Worked Is the Trap for the other half.
Wu Ch'i ordered the execution "with a sigh."2 Greene includes the detail and does not use it, and it changes what the anecdote is about.
A sigh means the general did not want to do it. He rated the officer — a talented officer — and killed him anyway, which tells you the decision was not a preference but a rule he judged himself unable to suspend. And a rule he could not suspend in this instance is a rule that might have been suspendable in another.
That reframes "establish a pattern of strict obedience to earn his trust"4 as something more specific than good conduct. Trust of this kind is not a verdict rendered on your character; it is a balance, accumulated by compliance and drawable exactly once per unusual act. The captain had no balance. He was new to the general's judgment, and his first withdrawal was a large one.
Which supplies the timing rule the passage lacks. The initiative is not forbidden — it is expensive, and the price is paid out of a stock built beforehand by doing ordinary things exactly as instructed for longer than felt necessary. Spend it on something worth it, once, having first checked that the balance exists.
The captain's error, on this reading, was not that he charged. It was that he charged before he had bought the right to.
The last piece of the passage is about favours rather than victories, and its mechanism is the sharpest thing in it: "It is a dangerous mistake to ask for more. You will seem insecure—perhaps you feel you did not deserve this favor, and have to grab as much as you can when you have the chance, which may not come again."6
The ask is not read as greed. It is read as a self-assessment, and that is why it costs so much.
Someone who believes they are on a rising trajectory has no reason to press — more will come, and pressing now would be a poor trade against the relationship. Someone who presses is revealing a private forecast in which this was the peak. You have not asked for a second favour; you have informed your superior that you consider the first one anomalous.
Greene's prescription follows from that: "The proper response is to accept the favor graciously and withdraw. Any subsequent favors you should earn without having to ask for them."6 The withdrawal is the signal. It says the forecast is fine.
The unstated dependency is worth naming: this only works on a superior who is paying attention. Against one who is not, silent excellence is invisible and the advice quietly rewards the wrong person. Greene never mentions the condition, and it is the condition under which most people actually work.
It is worth putting the halves of this passage side by side, because a reader taking away one lesson will get the other one wrong.
The Wu Ch'i captain acted without authorisation. His error was taking initiative that was not granted.
The favour-asker requested authorisation for something not offered. His error was asking rather than acting.
So stay in your lane and don't ask are being recommended simultaneously, and together they describe a very narrow channel: do exactly what was assigned, exceptionally well, and then wait. Greene presents this as tactical wisdom for serving a master and it is more accurately a description of what serving a master costs.
The channel does have an exit, and the chapter's own Observance is standing in it. Madame de Pompadour did not ration her output — she moved it into domains her superior had not assigned and could not feel rivalled in. See Madame de Pompadour and the Substituted Asset. The way out of the channel is not smaller victories; it is victories in a different currency.
You have done something well for someone you report to, and it went better than expected. The instinct is to follow it immediately with more.
Before you do, work out what you were actually being bought for. Was it the result, or was it the reliability? You can tell by asking what would have happened if you had produced the same result a different way, unannounced. If the answer is that would have been fine, you are being paid for outcomes. If there is a flicker of well, they'd have wanted to know, you are being paid for predictability, and an unauthorised win costs more than it earns.
Then check the register of your next move. Anything that requires them to revise their picture of what you are — a new capability, a new relationship, a visible constituency of your own — is the expensive kind, however good it is. That is the move to run past them first, in advance, framed as a question.
If you have just been given something — a promotion, a budget, an introduction — take it and stop. Do not attach anything to the acceptance. The second request is not read as ambition; it is read as your own estimate that this was the ceiling, and you will not be able to withdraw it.
And if you find you are rationing your output to stay comfortable, notice what that is. It is a real strategy and it is also a signal about the position. Sometimes the correct response is a different master.
Evidence quality: [POPULAR SOURCE] throughout. The Philip of Macedon claim is 🚩 SINGLE SOURCE and uncited, with no instance named; the best-documented Macedonian case runs contrary. The Wu Ch'i anecdote belongs to the Chinese military-classic tradition (the Wuzi and Sima Qian's Shiji) and is plausible, but 🚩 SECONDARY WITHOUT PRIMARY — Greene cites nothing.
Tension: "establish a pattern of strict obedience" directly contradicts the same chapter's instruction to vary your patterns. Preserved and reconciled above by audience, but Greene does not notice the collision.
Open questions: the favour-withdrawal strategy assumes a superior who observes and rewards unrequested excellence. What is the correct play under one who does not? The chapter's advice, applied there, produces invisibility, and Greene never states the precondition.
This is Law 47's version of a theme that runs from Law 1 onward — never outshine the master — and it is the darkest statement of it, because here the master punishes the subordinate for succeeding at the assigned task. The earlier laws warn against displaying superiority. This says the display is unnecessary; the result is sufficient.
That is also where the corpus becomes internally awkward. Mastery is a long argument that excellence compounds and eventually cannot be denied. Law 47 says excellence in service to another person has a ceiling set by that person's tolerance, and that exceeding it is punished immediately rather than eventually rewarded. Both books are addressed to the same reader and they imply different careers.
The convergence is with the Machiavelli marginal in the same chapter, which supplies the mechanism this passage assumes. See The Prince Discredits the Victory — Machiavelli explains why the master turns, and this passage explains what the subordinate should do about it. Greene never links them, and they are eighty lines apart in the same chapter.
Approval-Seeking Pathways — the psychological material describes what a person actually does when their standing depends on another's regard, and it supplies the motive Greene names in one clause and never examines. A desire to please a master by proving dedication is not a strategic miscalculation; it is a stable pattern with its own escalation logic, in which each round of approval raises the effort required to feel the same security, because the approval is being used to settle a question that is not settleable by evidence. That explains the otherwise puzzling behaviour of the Wu Ch'i captain — charging before a battle had begun, unprompted, at real risk — which makes no sense as a bid for advancement and perfect sense as a bid for reassurance. The insight neither frame reaches alone: Greene's remedy is calibration, measure your victories, which presupposes the subordinate is optimising an output and merely has the target wrong. If the behaviour is approval-driven, calibration is unavailable, because the quantity being maximised is not achievement but the master's visible response, and there is no level of it that terminates the search. Which means the correct intervention is not a smaller victory. It is noticing that you are asking a superior to answer a question no superior can answer, and that the escalation will continue in whatever currency is to hand.
The Best Closer Is Not the Best Sales Manager — the organisational frame supplies the cost side that Greene's operator-level advice leaves out. The Wu Ch'i execution reads as tyranny until you notice what a general is actually optimising: not the sum of individual performances but the composability of them, and an officer who wins unpredictably degrades composability regardless of his win rate. That is the same structure as the closer/manager problem — the individual excellence and the system role are scored on different functions, and promoting or rewarding on the first destroys the second. What the pairing produces that neither does alone: it reframes measure your victories from a piece of courtier's caution into a statement about which function you are being evaluated on, and gives the subordinate a way to check. If your superior's job is to compose your output with other people's, unauthorised excellence is a genuine cost to them and the suspicion is not paranoia. If your superior's job is simply to hold the position above you, it is. Those two situations look identical from below and call for opposite responses — the first for tighter coordination, the second for the Pompadour exit into a different currency — and the chapter, which has only one theory of masters, cannot tell them apart.
The transferable claim is the audience rule: patterns are liabilities against opponents and assets with superiors, because the two parties extract different things from legibility. That is checkable and it resolves a real contradiction inside the chapter.
The open edge is the composability distinction from the second handshake — whether a superior's suspicion of an unauthorised win is a legitimate coordination cost or a defence of position. The two are indistinguishable from below and imply opposite responses, and nothing in Greene's material offers a test.