You are playing badly and losing. And the person beating you stops, looks at you kindly, and offers to help.
Stephen Potter's The Complete Upmanship (1950) scripts the moment exactly, at a billiards table:
Gamesman: Look . . . may I say something? Layman: What? Gamesman: Take it easy. Layman: What do you mean? Gamesman: I mean—you know how to make the strokes, but you're stretching yourself on the rack all the time. Look. Walk up to the ball. Look at the line. And make your stroke. Comfortable. Easy. It's as simple as that.1
Nothing there is hostile. Nothing there is even wrong.
And then Potter names the mechanism, which is the reason the passage sits inside a chapter on one-upmanship:
In other words, the advice must be vague, to make certain it is not helpful.1
Potter's rule has a timing condition and a content condition, and both are precise.
Timing. There is exactly one correct moment:
there is only one correct time when the gamesman can give advice: and that is when the gamesman has achieved a useful though not necessarily a winning lead. Say three up and nine to play at golf, or, in billiards, sixty-five to his opponent's thirty.1
Not while losing, which reads as excuse. Not from an unassailable lead, which reads as gloating and produces sympathy. From ahead but not yet won — where the gap is real, the outcome is still open, and helping is the natural act of someone who is comfortable.
Content. Vague, and deliberately so. Take it easy. Comfortable. Easy. It's as simple as that. There is nothing here to act on, which is the design: actionable advice might work, and if it works you have improved your opponent.
Potter's verdict on the whole manoeuvre:
in general, if properly managed, the mere giving of advice is sufficient to place the gamesman in a practically invincible position.1
The mere giving. The content is irrelevant. The act is the entire content.
Unpack what is transmitted, none of it stated.
A status claim. Advice runs downhill. Offering it asserts that you know more, and the assertion is embedded rather than argued — which means it cannot be disputed without the other person sounding defensive.
A diagnosis. You're stretching yourself on the rack all the time tells your opponent that their difficulty is internal. Not the table, not the run of play, not the lie — them. And having been told that, they now have something to monitor while playing.
A reframing of the score. The lead becomes evidence of a difference in kind rather than a difference in the last few shots. You are not ahead; you are the sort of person who is ahead.
And an obligation. They must thank you.
Greene's parent chapter supplies the reason all of this bites, and it is the mask model: personality forms over weaknesses, and touch it the wrong way and it hurts. See Strategy 28: The One-Upmanship Strategy. Advice touches the wound in a way that criticism cannot, because it arrives as kindness and cannot be resisted without a cost.
The result is exactly what the parent chapter is designed to produce: someone who feels obscurely worse, is now self-conscious about their technique in the middle of performing it, and has no legitimate grievance to raise.
Potter's Glaciation material sits nearby in the same chapter, and it works on the same principle from the other direction — withholding the expected response rather than supplying an unexpected one.
If someone else tells a funny story, do not, whatever happens, tell your own funny story in reply, but listen intently and not only refrain from laughing or smiling, but make no response, change of expression or movement whatever. The teller of the funny story, whatever the nature of his joke, will then suddenly feel that what he has said is in bad taste.1
And the escalation, if there is genuine warmth in the room:
(a) join in the laughter at first. Next (b) gradually become silent. Finally (c) at some pause in the conversation be overheard whispering, "Oh for some real talk."1
Overheard. Not said. The move is constructed so that the target catches it obliquely and cannot address it.
Both gambits exploit the same asymmetry: a social act that would be rude to name. You cannot object to advice, and you cannot object to someone not laughing. In each case the only available response is to absorb it or to complain and thereby confirm the diagnosis.
Potter was writing satire about English club life, and the material has aged into something less funny than he intended, because the two structural conditions he identified are now standard features of professional life.
Advice runs downhill, and organisations run on it. Feedback, mentorship, coaching, the development conversation. Each is a channel with a built-in status gradient, and each is offered as help — which makes every one of them a valid delivery mechanism for the gambit, with institutional cover.
And vagueness is the norm rather than the exception. Be more strategic. Have more presence. Take it easy. Potter's insight was that unactionable advice is not failed advice — it is a different act that resembles advice, and it does something specific: it tells the recipient the problem is in them and gives them nothing to do about it.
Which yields the only real test, and it costs nothing to apply: could you act on it tomorrow? If not, whatever you were just given was not help. That does not make the giver malicious — most vagueness is laziness — but the effect on the recipient is identical either way, and that is the part Potter saw clearly.
You have been given advice and you feel worse, and you cannot say why without sounding ungrateful.
Run the two conditions first. Was the person ahead but not finished? That is the window in which advice is a move rather than a kindness — someone winning comfortably has no reason to help, and someone who has already won has nothing to gain from it.
Then apply the actionability test on the content, with no reference to the giver's motives. Could I do something differently tomorrow because of this? If yes, it is advice and you should take it, whatever it cost you to hear. If no, it is a status transaction that resembled advice, and the correct response is to thank them and discard it — not to argue, because arguing supplies the reaction the move is designed to produce.
Then notice whether you have started monitoring yourself mid-task. That is the actual damage and it is mechanical rather than emotional. Potter's billiards player was told he was tense; he now has a second job while making his shot. Self-consciousness during performance is the injury, and it is the one thing you can put down deliberately once you have named it.
And on the other side: if you are giving advice, the test is the same and it is worth applying to yourself. Could they act on it tomorrow? If not, you may be doing something other than helping, and — this is the uncomfortable part — you will not be able to tell from the inside, because the whole gambit is indistinguishable from generosity to the person performing it.
Stephen Potter's The Complete Upmanship (1950) is a real and influential comic work, and "gamesmanship" and "one-upmanship" entered English from it. Greene quotes it as a serious source throughout this chapter. [POPULAR SOURCE]
The source is satire being read straight, and Greene does not flag it. Potter was writing comedy about the absurdity of English competitive social life — the gambits are funny precisely because they are grotesque and because performing them would make you ridiculous. Greene extracts them as technique. The extraction mostly works, which is a real point in the material's favour, but the register shift is unmarked and a reader taking the Glaciation gambit at face value is being handed a joke as an instruction. 🚩 The chapter treats a humorist as a strategist without saying so.
And there is no reliability test at all. The two conditions identify when advice could be a manoeuvre. Nothing in the material distinguishes a competitor deploying vagueness deliberately from a genuinely well-meaning person who is bad at articulating what they know — and the second is far more common. Expert advice is frequently vague for an honest reason: much of what experts know is tacit and resists specification.
Which produces the risk this page has to name. The framework is unfalsifiable in the direction that damages you. If all advice can be read as a status move, and the tell is that it made you feel worse, then a person applying this consistently becomes uncoachable — armoured against exactly the input most likely to improve them. The actionability test above is the only defence that does not have that failure mode, because it evaluates the content rather than the giver, and it is the reason this page uses it instead of Potter's.
Against Strategy 28: The One-Upmanship Strategy this is the chapter's smallest and most repeatable instrument. The parent chapter's cases are set-piece manoeuvres requiring intelligence, timing and a target with a specific flaw. This one needs nothing but a scoreboard, which makes it the version most likely to be encountered — and to be used unconsciously.
Against The Bishop of Noyon and the Parody the mechanism is identical and inverted. The abbé praised the bishop in his own style so that praise carried a sting; the gamesman helps his opponent so that help carries one. Both weaponise a social form whose defining feature is that it cannot be refused — you cannot decline a compliment or resent an offer of help without becoming the problem.
Against Planted Information there is a real contrast worth keeping. That page holds discovered conclusions binding harder than delivered ones — the mechanism relies on the target doing the work. Here the target is given a conclusion about themselves, and it binds anyway. The difference is that a diagnosis of your own competence arrives as data about you rather than about the world, and it is checked against a self-image rather than against evidence.
Psychology → Attention paid to execution degrades execution. Potter's billiards player is told he is tense and must now watch himself not being tense while making a shot. This is a well-documented effect: directing conscious attention onto a practised skill disrupts it, which is why the advice does damage without containing any content.
The insight the pairing produces, and it is the reason this gambit is more than rudeness: the vagueness is not a failure to be helpful, it is what makes the advice unresolvable — and unresolvable instructions are the ones that stay live during performance. A specific correction can be applied and then dropped; take it easy cannot be completed, so it runs in the background indefinitely. Which gives a genuinely practical rule for anyone who has to perform under observation: an instruction you cannot finish executing should be set down before you start, not because it is unkind but because the only way to hold it is to keep monitoring yourself, and monitoring is the thing that costs.
Business → The help channel is a status channel, and organisations built one deliberately. Every workplace has formal structures whose stated purpose is advice flowing from senior to junior — feedback cycles, mentoring, coaching, review. Set that beside Grant and McClernand, where the whole technique consists of acting strictly within legitimate authority so that nothing can be appealed.
Together they produce something uncomfortable that neither gives alone: the institutionalisation of advice creates a sanctioned, unappealable, permanently available one-up channel, and its legitimacy is exactly what makes it unanswerable. You cannot object to feedback; objecting is itself evidence of the problem. This does not mean feedback structures are bad — it means they carry a load they were not designed for, and the single cheapest correction is the actionability standard: make specificity the norm, so that vagueness becomes conspicuous rather than routine. A culture where "be more strategic" is an acceptable output is one where the gambit is undetectable, whether or not anyone is running it deliberately.
Unactionable advice is not failed advice — it is a different act wearing advice's clothes, and its effect is to tell someone the problem is inside them while giving them nothing to do about it. The test is not the giver's motive, which you cannot see, but whether you could act on it tomorrow.