You are standing at the edge of a fast river you have to cross. There is no bridge. What there is, running out into the water in a haphazard line, is stones.
Greene's Image for this chapter is exactly that, and the instructions attached to it are the whole concept:
If you linger too long on one stone, you will lose your balance. If you go too fast or skip one, you will slip. Instead you must jump lightly from one stone to the next and never look back.1
Note what a stone is not. It is not a foundation, not a home, not something you build on. It holds your weight for as long as you are standing on it and its entire value is that there is another one within reach.
Think of your alliances as stepping-stones toward a goal. Over the course of your life, you will be constantly jumping from one stone to the next to suit your needs. When this particular river is crossed, you will leave them behind you.1
Greene names the practice: the Alliance Game.
The principles come from ancient China, and the historical conditions are the point:
ancient China . . . was composed of numerous states in continual flux—now weak, now powerful, now weak again.1
Two failure modes emerged from that instability, and they are still the two.
Invade and you are mistrusted. A state that attacked another would stir up a lot of mistrust among the others and would often find itself losing ground in the long run. Aggression is legible, and it re-sorts everyone else's alliances against you.
Stay loyal and you are dragged. A state too loyal to an ally might find itself pulled into a war from which it could not break free and would go down in the process.1
Which produced a striking result:
The formation of proper alliances was in some ways a more important art than that of warfare itself, and the statesmen adept at this art were more powerful than military leaders.1
The state of Chin expanded through the Warring States period (403–221 BC) on a rule simple enough to state in a line:
Chin would make alliances with distant states and attack nearby ones; the nearby state that Chin had invaded could not get help from its outlying neighbor because that neighbor was now allied to Chin.1
That is elegant. The alliance is not a source of support; it is a denial of support to the victim. You ally with the far country precisely because it has nothing to fear from you yet, which makes the alliance cheap, and because its neutrality is what leaves the near country alone.
And when a target already had a protector:
If Chin faced an enemy that had a key ally, it would work first to disrupt the alliance—sowing dissension, spreading rumors, courting one of the two sides with money—until the alliance fell apart. Then Chin would invade first one of the two states, then the other.1
Bit by bit, until in the late third century BC it unified China.
The distant-ally rule has a shelf life nobody mentions, and it is the strategy's own clock. Each conquest moves your border outward. Today's distant ally is tomorrow's neighbour. Chin's method necessarily consumed its own allies in sequence, and it worked because it finished — the game ended when there was nothing left to be far away.
To play the Alliance Game right, today as in ancient China, you must be realistic to the core, thinking far ahead and keeping the situation as fluid as possible. The ally of today may be the enemy of tomorrow. Sentiment has no place in the picture.1
And the payoff for the weak:
If you are weak but clever, you can slowly leapfrog into a position of strength by bouncing from one alliance to another.1
Greene then states the alternative fairly, which he does not always do:
The opposite approach is to make a key alliance and stick with it, valuing trust and an established relationship. This can work well in stable times, but in periods of flux, which are more common, it can prove to be your undoing: differences in interest will inevitably emerge, and at the same time it will become hard to disentangle yourself from a relationship in which so much emotion has been invested.1
The conclusion: bank on change, keep your options open, and base your alliances on need, not loyalty or shared values.1
Notice the actual argument, because it is narrower than the rhetoric. The claim is not that loyalty is stupid — it is that loyalty is a bet on stability, and Greene thinks flux is the more common condition. That is an empirical claim about your environment, and it is answerable. In a genuinely stable field, the long alliance wins.
The chapter's modern case is chosen for a structural reason: it comes from a system designed to make the Alliance Game impossible.
In the golden age of Hollywood, actresses had almost the least amount of power of anyone. Careers were short; even a great star would be replaced in a few years by someone younger. An actress would stay loyal to her studio, then watch helplessly as the roles dried up.1
Joan Crawford played it anyway, and the moves are worth listing because none of them is glamorous.
1933, Joseph Mankiewicz. A timid young man just starting out. She recognised the talent immediately and went out of her way to befriend him — much to his amazement. He wrote nine screenplays for her, greatly lengthening her career.1
Cameramen and photographers, who then worked overtime to light her well and make her look good.1 The people who decide what you look like on screen are not the people with titles.
Producers who controlled a script with a role she wanted.
Up-and-coming young talent, who valued a relationship with a star — and then she would gracefully break or forget the connection when it no longer served her needs.1
Nor would she stay loyal to the studio, or indeed to anyone—only to herself. Her unsentimental approach to her own shifting network of alliances allowed her to avoid the trap that most actresses found embedded in the system.1
The transferable part is the direction of the allying. Crawford went sideways and downward — a beginner writer, the crew, people on the way up — not toward power. Which is Greene's next rule.
The key to playing the game is to recognize who can best advance your interests at that moment. This need not be the most obviously powerful person on the scene . . . alliances that meet specific needs or answer particular deficiencies are often more useful.1
With a parenthesis carrying more weight than its punctuation suggests: (Grand alliances between two great powers are generally the least effective.)1
The historical proof is Louis XI, who needed the Swiss precisely because he had a weak army — see Louis XI and the Spider King's Web. The modern one is LBJ:
As an ambitious young congressional assistant in Washington, Lyndon Johnson realized he lacked all kinds of powers and talents to get him to the top. He became a clever user of other people's talents. Realizing the importance of information in Congress, he made a point of befriending and allying himself with those at key positions—whether high or low—in the information chain.1
Whether high or low. Johnson allied with whoever sat on the flow, at any grade. And he was particularly good with older men who enjoyed the company of a lively young man and the role of the father figure giving advice1 — which is a real observation about what senior people are often short of.
The chapter's compressed image for the whole thing:
It is a common strategy in bicycle races not to go out in front but to stay right behind the leader, a position that cuts down wind resistance—the leader faces the wind for you and saves you energy. At the last minute, you sprint ahead.1
Greene's verdict: letting other people cut resistance for you and waste their energy on your behalf is the height of economy and strategy.1
It is a good image and it quietly concedes something. Drafting is legal, universal, and everyone in the race knows it is happening. It works because the leader has their own reasons to be in front. The image describes cooperation with divergent exit conditions, not exploitation — which is a milder and more defensible thing than the surrounding prose claims, and probably the version most readers can actually use.
You are junior, or peripheral, or in a field that chews people up on a schedule, and the obvious route to advancement is closed.
Start by finding the flow rather than the ranks. Johnson's insight was that in Congress the currency was information, so he allied with whoever touched it at any level — clerks as readily as members. In any organisation, ask what the actual scarce good is (information, access, a specific technical judgement, the ability to make something look good) and then find who touches it. That list will not match the org chart, and the mismatch is the opportunity.
Then go sideways and early. Crawford's single best alliance was a nervous young screenwriter nobody wanted yet, and it bought her nine films. The person who will be powerful in five years is available now, at no cost, and will remember that you were there when it was cheap. The cost of an alliance is inversely proportional to how obviously valuable the person is.
Then be honest with yourself about the exit before you need one. Greene's advice is to break connections gracefully when they stop serving. You may not want to do that — but you should at least know, in advance, what this alliance is for, because the alternative is not loyalty, it is drift, and drift is what leaves people carrying relationships that stopped paying years ago.
Then check which environment you are actually in. This is the step the chapter skips and it changes the answer. Fluid field: the stepping-stone approach is correct. Stable field with long memories and repeated dealings — a craft, a small profession, a specialist market — and the person who bounced between alliances has a reputation that arrives before they do. Greene concedes the exception in one clause and then ignores it.
And take the draft when it is offered. Someone senior is going out in front for their own reasons. Sitting in their slipstream costs them nothing and saves you a great deal, and there is no need to dress this up as either strategy or betrayal.
The Warring States alliance dynamics, Chin's expansion and unification in 221 BC, Crawford's working relationship with Mankiewicz, and Johnson's early congressional networking are all well attested. [POPULAR SOURCE]
The Crawford case is being read as pure strategy where survival is the more parsimonious reading. Greene's own framing supplies it: Hollywood actresses had less power than almost anyone, careers were short by design, and loyalty to a studio was rewarded with the roles drying up. Crawford's behaviour in a system engineered to discard her at thirty-five is at least as well described as rational self-defence against a rigged arrangement as it is by a chapter about using people. The chapter needs her as an exemplar of unsentimental play, and files what she was playing against as background.
The chapter's core claim is empirical and never tested. Everything rests on periods of flux, which are more common. If that is right the strategy follows; if a reader's actual field has long time horizons and repeat dealings, the conclusion reverses. Greene offers no way to tell which you are in, and it is the only question that matters for applying any of this.
And the reputational cost is acknowledged in one line and then dismissed. The chapter concedes that your reputation will eventually keep people from making alliances with you, then answers that self-interest rules the world. That answer only works in a large market with poor memory. In small fields — which is where most people actually work — the reputational cost is not a distant risk, it is the mechanism by which the strategy stops functioning.
Against Strategy 27: The Alliance Strategy this page holds the systematised method and the parent page holds the selection criteria — but the two do not sit flush. The parent's most useful contribution is the false-alliance test, which asks whether an arrangement leaves you able to make your own decisions. The Alliance Game as described here would fail its own parent's test if applied to you, since a stepping-stone relationship is precisely one in which the other party cannot make plans.
Against Three Death-Ground Tactics there is a third Crawford treatment in this build, and reading them together produces something none of them says. Strategy 2 has her changing style across a career; Strategy 4 has her leaving MGM in 1943 to enter new waters; this chapter has her network. Greene presents each as an illustration of a different principle. They are one continuous survival strategy by a woman in a system with a built-in expiry date, and the book's chapter structure is what prevents that from being visible.
Against Strategy 7: Morale Strategies and LBJ and the Devoted Army the tension is sharp and personified. This page has Johnson building alliances of pure convenience with the information chain; Strategy 7 has him building a staff of eighteen-hour devotees through cruelty and unexpected kindness. Same man, same years, cultivating disposable alliances sideways and permanent loyalty downward — which is the concrete form of the book's unstated rule that you receive loyalty and do not give it.
Business → Weak ties carry the opportunities. Crawford's network was junior writers, cameramen and people on the way up; Johnson's was whoever sat on information whether high or low. Set that beside John Boyd and the Porcupine, where the operative structure inside a bureaucracy is a set of small cells recruited from the young and dissatisfied, deliberately outside the formal channels.
What the pairing produces that neither gives alone: the value of a connection tracks its distance from your existing cluster, not the other person's rank — because someone embedded in the same network as you knows what you already know, and their power is exercised through channels you can already see. The junior person in an adjacent function has non-redundant information, and the cost of reaching them is near zero because nobody is competing for them. Which inverts the ordinary instinct about networking upward: the senior figure everyone wants is expensive and low-yield precisely because everyone wants them, and their attention is the most rivalrous good in the room. Boyd could not have bought a three-star general and did not need one.
History → The distant-ally rule expires by succeeding. Chin allied with the far and attacked the near, and every conquest moved the border so that a former distant ally became a neighbour. Set it against Strategy 29: The Fait Accompli Strategy's piecemeal accumulation, which runs on the same tempo.
Together they yield a claim about the class of strategy rather than the tactic: incremental-expansion methods consume the conditions that make them work, so they are only correct if you can finish. Chin could — it ran out of China. A firm rolling up its adjacent partners, or a person leapfrogging alliances inside one industry, generally cannot, and the terminal state of an unfinished version is an actor with no distant allies left and a ring of neighbours who all watched what happened to the last one. The question to ask before starting is not whether the method works but whether the board is finite — and most boards are not.
The strategy's central claim is empirical, not moral: it says your field is unstable, and everything follows from that. In genuine flux, stepping-stones beat foundations. In a small field with long memories, the reputation arrives before you do — and Greene concedes this in a single clause before moving on.