You can usually tell before anyone says it.
Meetings start getting scheduled that they're not in. Decisions get made and then mentioned to them. Someone two levels up starts asking you things directly.
Nobody has announced anything. But the room has quietly re-priced somebody, and the question Law 1 has been answering for four pages — how do I avoid outshining this person? — has stopped applying.
Greene's Reversal is about that moment, and it is the coldest paragraph in the law.
"You cannot worry about upsetting every person you come across, but you must be selectively cruel. If your superior is a falling star, there is nothing to fear from outshining him."1
The whole apparatus — the discreet flattery, the manufactured mistakes, the ascribed ideas — is a tax you pay for the master's capacity to hurt you. Remove the capacity and the tax stops.
That's coherent. It's also a much larger claim than it looks, because it means Law 1 was never an ethic. It was a threat assessment, and threat assessments expire.
Greene forks the situation by degree of weakness, and the branches don't share a logic.1
Weak: "discreetly hasten his downfall: Outdo, outcharm, outsmart him at key moments."
Very weak, ready to fall: "let nature take its course. Do not risk outshining a feeble superior—it might appear cruel or spiteful."
Firm, but less capable than you: "bide your time and be patient… Your master will fall someday, and if you play it right, you will outlive and someday outshine him."
The middle branch looks like restraint and isn't.
Read the reason Greene gives for sparing the nearly-fallen: "it might appear cruel or spiteful."1 Not that it would be cruel. That it would look it — to the audience whose regard you're managing.
So the same instrument runs all three branches: what does this do to how I am seen? Attack the weak because it works; spare the dying because attacking them costs you standing. Mercy never enters as a term.
That's worth naming plainly rather than smoothing, because the passage is easy to read as Greene showing a softer side. He isn't. He's noting that kicking a man who is already going down is bad optics.
The first branch carries the argument that licenses everything, and it's one sentence:
"Do not be merciful—your master had no such scruples in his own cold-blooded climb to the top."1
This is a tu quoque — the accusation that they did it too — and it has two problems Greene doesn't address.
It's asserted of everyone. No method is offered for finding out whether a particular superior climbed cold-bloodedly. The premise is applied to the class. Some people are promoted for being good at the work.
It licenses pre-emptively. The claim isn't that this master wronged you. It's that they presumably wronged someone, at some point, on the way up — so ordinary restraint toward them is suspended in advance, on the strength of a biography you haven't checked.
Set beside the Preface's own instruction to "judge all things by their effects" rather than by imputed intentions, the tension is direct: here the reader is invited to act on an imputed history nobody has observed.
Your boss is visibly sliding, and a project has come up that you could take.
The temptation is to read Greene and move. Resist that for one specific reason: your assessment of their strength is the input, and you have never tested it.
So test it cheaply before you act. Say something mildly independent in a meeting — a position of your own, not a challenge. Watch what happens over the following fortnight. Does it get absorbed, or does something cool?
Do that twice before concluding anything. What you are measuring is whether they still have the reach to make it cost you.
Then, if the read holds, choose the branch honestly. Note that two of Greene's three branches are wait — "let nature take its course" and "bide your time" — and only one is act. The paragraph reads as permission to move, and its actual arithmetic recommends patience in the majority of cases.
And if you do decide to move, be clear with yourself about which sentence you're relying on. "He'd have done it to me" is the only justification on offer here, and you almost certainly don't know that.
Law 1 is presented as near-universal — outshining the master is "perhaps the worst mistake of all."2 The Reversal makes it conditional on a variable Greene supplies no instrument for reading.
"Gauge his strength"1 is the entire method. That's not a method; it's the name of the problem. And it's the same shape as the gap at Law 1's other edge — Manfredi outshone Cesare Borgia involuntarily, and the only counsel offered was avoidance.
So the law's boundaries on both sides rest on assessments the reader is told to make and never taught to make.
The branches don't carry symmetric risk, and Greene never says so.
Misjudge a falling star as firm and you defer unnecessarily for a while. The cost is opportunity — some deference you didn't owe, some months of dimming yourself.
Misjudge a firm star as falling and you have visibly moved against someone who retains the power to end you, having announced your intention in the act. There is no recovery from that, because the move is not deniable: outdoing and outcharming someone at key moments is exactly the behaviour Law 1 spent four pages calling fatal.
So the two errors are not equivalent, and the conservative branch is right far more often than the paragraph's tone implies. A rule offered with this much confidence ought to carry that asymmetry on its face. This one doesn't.
Strongest evidence. The core insight is sound and generalises: deference is priced by the other party's capacity to retaliate, so it should be re-evaluated when that capacity changes. Most people under-adjust, continuing to defer to authority that has quietly evaporated.
Tension 1 — the tu quoque is asserted of all masters with no way to check it, and it does all the moral work.
Tension 2 — "gauge his strength" is a naming of the problem, not a solution to it.
Tension 3 — mercy is optics. The one restrained branch is justified by appearance, not by the harm avoided.
🚩 SINGLE SOURCE · 🚩 SECONDARY WITHOUT PRIMARY — no case, no citation; pure assertion. All claims [POPULAR SOURCE].
Open questions. Is there any observable signal that distinguishes a genuinely falling superior from one in a temporary trough — before the outcome? And if two of three branches counsel waiting, why does the passage read as licence?
The vault's Laws of Human Nature material on authority and its fickle subjects treats a leader's decline as something a group produces collectively — attention withdraws, the story about them changes, and the fall is largely the group's act rather than any individual's.
That reframes this Reversal usefully. Greene here casts the subordinate as the agent, deciding whether to push. The later book suggests the push rarely decides anything; what decides it is a shift already underway in the room.
Which makes the first branch mostly theatre. "Outdo, outcharm, outsmart him at key moments"1 is a way of being seen to be on the right side of a change that was happening regardless — with all the risk of having visibly moved first, if you misread the timing.
Handshake 1 — psychology: Envy as Shadow Signal
Greene assumes the subordinate can assess the master coolly. The shadow material says the one judgement you cannot make coolly is about someone whose position you want.
Envy doesn't present as envy; it arrives pre-converted into an assessment that feels neutral. He's lost it. People have noticed. He's not what he was. Every one of those is exactly what genuine decline looks like, and exactly what wanting someone's job produces.
The insight neither produces alone: "gauge his strength" asks for a reading that the reader's own motivation systematically corrupts, in the direction of the answer they want. The Reversal's single instrument is the one measurement most vulnerable to interested distortion — which is why the two-test discipline above matters more than any amount of confidence about the read.
Handshake 2 — history: Mao and the Long March
Greene asks the subordinate to gauge a superior's weakness. Here is the same measurement run by the stronger party, and got catastrophically wrong.
Chiang Kai-shek broke off the pursuit of Mao's forces once they were reduced to a remnant, on the assessment that what was left was no longer a threat. The judgement was reasonable on the visible evidence — the numbers were tiny and the survivors were exhausted. It cost him China.
Put beside the Reversal, the useful thing is the symmetry. Greene treats "gauge his strength" as a hard problem for the person below, solvable with attention. Chiang had vastly better information, a professional staff, and no motivated interest in underestimating an enemy he wanted destroyed — and still misread it.
What the two produce together: the unreliable element isn't the observer, it's the measurement. Residual strength in a declining position is genuinely not observable from outside, because most of what determines recovery — will, remaining alliances, what the other party is prepared to spend — is private. That reframes the Reversal's problem: it isn't that subordinates are biased (though the envy handshake says they are). It's that Greene has built the exception on a quantity nobody can read reliably in either direction, and then made it the single switch.
Sharpest implication. Law 1 is a threat assessment wearing the clothes of an ethic, and the Reversal admits it. Which means the honest question is never "should I defer?" but "what can this person still do to me?" — re-asked periodically, because the answer changes and most people never re-ask.
Generative questions.