Spring 1863. You are John Pemberton, holding the Confederate fortress at Vicksburg, and your position is genuinely excellent.
The fort sits on top of a two-hundred-foot bluff above the Mississippi; anything that tries to steam past is under your guns. West is the river and the cliffs. North, where Grant is camped, is swamp that is effectively impassable. East, not far, is Jackson — a railroad hub, firmly Confederate, feeding you supplies and reinforcements down a corridor you control.1
Grant has already tried several angles from the north and failed at all of them. The Northern papers are calling him an incompetent drunk. Lincoln's enemies are calling the whole campaign a waste of money and men, and the pressure on him to give up and pull back to Memphis is enormous.1
Then Grant crosses the river south of you. And you are still fine, because you know the rule:
Even if Grant got thousands of men across the river, what could he do once there? . . . Defeat in this corridor would be a disaster, for Grant would have no line of retreat. He had committed himself to a foolhardy venture. Pemberton waited patiently for his next move.1
Greene's phrase for Pemberton's state of mind is exactly right. It was a game of chess that Pemberton could not lose.1
The rule is worth stating properly, because the whole case depends on how solid it was.
An army eats. An army needs ammunition. Therefore an army needs a supply line back to a base, and a base needs protecting, and everything the army does is constrained by the need to keep that thread intact. This is not military superstition; it is arithmetic, and it had governed campaigning for centuries. It is also why a formation has a front and a rear — the rear is where the thread runs.
Greene's account of why such rules calcify is the sharpest general claim on the page:
war is such risky business that generals are often tempted to take the road well traveled. When so much is necessarily unsafe, what has proven safe in the past has amplified appeal.1
Risk does not make people experimental. It makes them conservative, because the downside of an unproven move in a lethal environment is unbounded. So the more dangerous the domain, the more rigidly it runs on precedent — which is precisely what makes a dangerous domain exploitable by anyone willing to break precedent.
And Greene notes the timing: Napoleon had already loosened these principles decades earlier, and forty years after his death officers like Pemberton still could not imagine an army behaving according to any other plan.1
The part of this campaign that matters for the chapter is the part that looks like nothing.
On the moonless night of 16 April, a Union flotilla ran the Vicksburg batteries with lights off and got past with minimal damage. More runs followed. Union forces on the western bank were seen moving south.1 Pemberton reads this correctly — Grant will cross downriver and come at the fortress from the south.
Grant crosses, moves northeast toward the Vicksburg–Jackson rail line, and establishes a supply base at Grand Gulf on the eastern bank. He forms a front and a rear and marches like an army.1
Greene's Interpretation names what that base was for:
This strategy was hidden from Pemberton's view because Grant kept up ordinary appearances by establishing a base at Grand Gulf and forming front and rear to march toward the rail line.1
The base was not a mistake, and it was not really a base. It was the banal, expected, entirely legible object that fixed Pemberton's model of what was happening — the ordinary manoeuvre that the extraordinary one would unfold out of.
Pemberton does the correct thing. He sends forces south to threaten Grand Gulf, because threatening a supply base forces a retreat. That is what threatening a supply base does.
To Pemberton's utter dismay, Grant ignored him.1
Not countered. Not defended. Ignored.
Grant pushed on to Jackson and took it on 14 May. Instead of relying on supply lines to feed his army, he plundered the area's rich farmlands. And he moved so quickly and changed direction so fluidly that Pemberton could not identify which part of the army was front, rear or flank.1
Greene's summary sentence:
Rather than struggle to defend lines of communication or supply, Grant kept none. No one had ever seen an army behave in such a manner, breaking every rule in the military playbook.1
There is nothing left to threaten. The lever Pemberton had spent the whole campaign preparing to pull was attached to something Grant had detached and walked away from.
Jackson taken, Grant wheeled on Vicksburg. Pemberton rushed his men back from Grand Gulf too late, lost at Champion Hill, and was pushed into the fortress and besieged. He surrendered on 4 July — a blow, Greene writes, the South never recovered from.1
Greene is clear that this was not a clever idea somebody had; it was a decision taken against every available authority, including friendly ones.
It took great courage for Grant to disobey these conventions and cut himself loose from any base, living instead off the rich lands of the Mississippi Basin. It took great courage for him to move his army without forming a front. (Even his own generals, including William Tecumseh Sherman, thought he had lost his mind.)1
That parenthesis is the most useful thing in the passage. The person telling Grant this was insane was not a fool or a rival — it was Sherman, who within eighteen months would march an army from Atlanta to the sea living off the country, doing precisely this. See Sherman and the Horns of a Dilemma.
Which means the objection was not ignorance. It was that the move had not been done yet, and Sherman's own experience did not contain it — the same mechanism the chapter identifies in the French in 1940, running here inside Grant's own command.
Greene's closing note is aimed at when to reach for this:
This is particularly important when we are on the defensive or in a weakened state. Our natural tendency at such times is to be conservative, which only makes it easier for our enemies to anticipate our moves and crush us with their superior strength; we play into their hands.1
Grant did this while being publicly called a drunk and a waste of the Treasury's money. The weakness is not incidental to the story — it is the condition that made the conventional path a guaranteed loss and therefore made the unconventional one cheap by comparison.
Someone has you in a position they consider closed. Not merely strong — closed, in the sense that they have identified the thing you cannot do without and positioned themselves on it. That is the specific shape of this case.
Start by finding what they think your supply line is. It is whatever they keep threatening. Watch where their pressure actually goes: the client relationship they assume you cannot lose, the approval they assume you must have, the fallback they assume you are keeping open. Their leverage is a hypothesis about your dependencies, and they have shown you the hypothesis by acting on it.
Then ask the question almost nobody asks: what happens if I simply don't have that? Not "how do I defend it" — that is the move Pemberton has already priced in. Whether the dependency is real, or whether it is real only because you have never operated without it. Grant's army did eat. It just ate something other than what came up the river.
If you can live without it, keep the base anyway for a while. This is the part that separates the manoeuvre from mere recklessness. Grant built Grand Gulf, formed a front, and marched conventionally for days — because the value of cutting loose comes from the other side not knowing you have. An abandonment they can see is just a retreat with extra steps.
Then move faster than their model updates. Grant's speed was not for covering ground; it was so that Pemberton could not tell which part of the army was front, rear or flank. The advantage lasts exactly as long as their picture is stale, and it ends the moment they redraw it.
And be honest about the register of the decision. This is not a low-risk option that looks risky. Grant genuinely had no line of retreat, and Sherman genuinely thought he was mad. What made it correct was that the safe path was already losing.
The Vicksburg campaign is one of the most thoroughly documented operations of the Civil War, and the essential facts — the April run past the batteries, the crossing south, Grand Gulf, the capture of Jackson on 14 May, Champion Hill, the siege, and the 4 July surrender — are uncontroversial. [POPULAR SOURCE]
"He plundered the area's rich farmlands" is the whole moral content of the campaign, compressed into six words and used as a logistics note. Living off the country means taking food from the people who live there, at scale, in wartime, without payment. Greene reports it as the enabling mechanism, calls the decision courage, and does not return to it. This is the book's running SIGNAL at its most economical: the strategy is stated as clean instrumental logic and the human cost enters the sentence as a supporting clause. It also sits directly alongside the same book's treatment of Sherman's March, where the identical mechanism is praised for not wasting men in battle.
The campaign is read backward from 4 July. Grant's earlier attempts had failed repeatedly, and Greene's own framing — he tried every kind of maneuver, with nothing working1 — describes a commander running out of options under political pressure at least as well as it describes a strategist selecting an unconventional line. The chapter resolves that toward design. What the record supports is narrower and more interesting: the unconventional move became available because the conventional ones had already been exhausted, which makes desperation part of the causal story rather than mere context.
The rule Grant broke was not stupid, and the chapter needs it to be. Armies that lose their supply lines usually die. Grant could live off the Mississippi basin in May because that specific country, in that specific season, was rich enough to feed him. The chapter draws the general lesson — cut yourself loose from the base — from a case whose success depended on a local fact it never mentions. The convention was not obsolete; it was inapplicable in one place at one time, and Grant appears to have known the difference.
Against Grant and Self-Reliance this is the same commander and the same campaign, told for a different purpose eleven chapters earlier. That page holds Strategy 3's psychological reading — the drift into dependency on subordinates, the drinking, and the deliberate rebuild of self-reliance by riding the terrain and reading intelligence himself. This page holds the operational manoeuvre.
The two accounts fit together better than Greene notices, and he never connects them. An army with no supply line and no front cannot be run from a map by a commander relying on subordinates' reports — the fluidity that confused Pemberton is only available to someone who has already rebuilt the habit of seeing the ground himself. Strategy 3's psychological fix is a precondition of Strategy 24's operational move, and the book presents them as two unrelated illustrations of two unrelated principles.
Against Strategy 8: The Perfect Economy Strategy there is a real tension. Strategy 8 teaches knowing your limits, counting hidden costs, and Greene's own counter-case there is the Dardanelles — a campaign abandoned within days of success by over-economising. Cutting loose from all supply while surrounded by enemy forces is the least economical decision available, and it is correct here. The two chapters give opposite instructions and neither supplies the rule for choosing between them; the nearest thing to a bridge is Strategy 24's own line about weakness making conservatism fatal, which Strategy 8 does not acknowledge.
Against Genghis Khan and the Shah of Khwarizm the structural echo is exact and six hundred years apart: mares' milk and remounts doing for the Mongols what the Mississippi farms did for Grant, and in both cases the enemy's model of what an army needs is the thing being attacked.
Business → The dependency you have never tested. Pemberton's leverage was a belief about what Grant could not do without, and the belief was inherited rather than checked. Organisations run on the same kind of belief constantly: the channel that supposedly cannot be left, the customer whose loss is treated as fatal, the platform everything is built on.
The insight the pairing produces that neither side gives alone: an untested dependency is simultaneously your largest risk and your opponent's largest assumption, and the same experiment resolves both. Whoever runs it first gains twice — the one who discovers they can operate without the thread removes a threat, and the one who keeps assuming the other side needs it has built their entire position on a hypothesis they have never falsified. Which suggests a defensive practice with no obvious equivalent in the strategy literature: periodically operate for a fixed period as though the critical dependency were gone, not as a resilience exercise but as intelligence about what your competitors currently believe they have on you.
Psychology → Precedent hardens fastest where the stakes are highest. Greene's explanation for why generals take the road well travelled — that in an unsafe domain, what has proven safe acquires amplified appeal — is a claim about how risk shapes cognition, and it predicts the opposite of what most people assume. High stakes are usually imagined to sharpen thinking. Here they narrow it.
What the pair produces: the domains most in need of innovation are structurally the least able to produce it, and the effect is strongest in exactly the people with the most experience. Pemberton was not a bad officer; he had more precedent available than a junior one, and every additional precedent raised the cost of departing from the set. This is why the outsider advantage in high-stakes fields is real and also why it is so often mistaken for recklessness — the outsider is not braver, they simply have a smaller library of things that have previously worked, and therefore less to override. Sherman thought Grant had lost his mind because Sherman knew more, not less.
Your opponent's leverage is a hypothesis about what you cannot survive losing, and they have already told you what it is by showing you where they push. The move that ends the game is not defending it — it is finding out whether the dependency was ever real, while continuing to behave as though it were.