Three duels, three opponents, three completely different weapons and postures and moods — and the only thing Miyamoto Musashi actually repeated across all of them was the refusal to repeat himself.
Everyone he fought had studied him, or thought they had. Every one of them was wrong about what he'd do, because what he'd done last time told them almost nothing about what he'd do this time.
This is Robert Greene's second illustration of Strategy 2, the Guerrilla-War-of-the-Mind: the discipline of never fighting the last war, run at the tightest possible scale. Not a career, not an army. One fight at a time, with your life on the table.1
In 1605, Musashi — already famous at twenty-one — was challenged by Matashichiro of the Yoshioka clan, whose father and uncle Musashi had already killed in earlier duels. This time the challenge was a trap. Matashichiro arrived with a small army hidden in the grass, counting on Musashi's well-known habit of showing up hours late to unnerve his opponents.
"He'll arrive way behind schedule as usual," one of the ambushers reportedly said, "but that trick won't work with us anymore."1
Musashi came early instead, and hid in the trees. When Matashichiro's men had settled into their ambush positions, Musashi leaped out, killed Matashichiro in one stroke, then took up a position at an angle to the remaining men rather than letting them surround him.
Caught off guard and startled by the reversal, the ambushers stood in a broken line instead of encircling him. Musashi ran down the line, killing them one after another in seconds.1
It's worth being precise about what Matashichiro's men got wrong, because it wasn't a failure of numbers or preparation — they had both. What they had instead of a real plan was a prediction, built from a genuine pattern in Musashi's past behavior, treated as though a pattern observed twice was a law that would hold a third time. The ambush wasn't badly designed. It was designed for an opponent who had already stopped existing the moment Musashi decided to break his own habit.
Later, Musashi sought out a warrior named Baiken, undefeated with an unusual weapon: a sickle and a long chain with a steel ball on the end. Baiken's method was consistent and had never failed — swing the ball to build momentum, force the opponent backward with a relentless charge, then hurl the ball at the face; while the opponent's sword arm was occupied fending it off, slash with the sickle across the neck.1
Musashi showed up with two swords — one long, one short — a combination Baiken had never faced. Instead of letting Baiken charge first, as every previous opponent had, Musashi charged, pushing Baiken back on his heels.
Baiken hesitated to throw the ball, because Musashi could parry it with one sword and strike with the other — the entire logic of Baiken's proven method depended on an opponent with only one weapon to divide between defense and attack. Musashi knocked him off balance with the short sword, then killed him with the long one.1
There's a specific irony worth naming: Baiken's perfect record is exactly what made him beatable. A method that has never failed has also never been stress-tested against its own blind spot, because no opponent has yet forced it to reveal one. Musashi didn't need to be a better fighter than everyone Baiken had already beaten. He only needed to be the first person to bring a configuration — two weapons instead of one — that Baiken's flawless technique had simply never had to account for.
The final duel, against the great swordsman Sasaki Ganryu, is the strangest and the one Greene lingers on longest. Ganryu fought with an unusually long, beautiful sword. The duel would take place on a small island, and word had spread — a crowd gathered to watch.
Ganryu arrived on time. Musashi didn't. An hour passed, then two, Ganryu furious. Finally a boat appeared with a passenger lying half-asleep, whittling at a long wooden oar, staring at the clouds. It was Musashi — arriving not with a sword but with an oar cut down into a makeshift weapon, longer than Ganryu's famous blade, and a dirty towel tied around his head in place of proper battle dress.1
Ganryu, furious at the lateness, drew his sword and threw the sheath into the sand. Musashi smiled: "Sasaki, you have just sealed your doom." Confused, Ganryu demanded to know how. "What victor on earth," Musashi replied, "would abandon his sheath to the sea?"
The line meant nothing tactically. It was pure psychological provocation, timed to land right before the fight, on a man already destabilized by two hours of waiting and an opponent who wasn't taking any of it seriously.1
Musashi charged, aiming the oar at Ganryu's eyes. Ganryu struck at Musashi's head and missed — his first miss ever, by reputation — cutting only the towel in two. In the same instant, Musashi's wooden weapon knocked Ganryu off his feet, and killed him with a blow to the head as he tried to rise.
Musashi bowed politely to the officials and left as calmly as he'd arrived.1
Nothing about technique connects them. What connects them is that Musashi diagnosed each opponent's specific expectation and broke exactly that one.
Matashichiro's men expected lateness — he came early. Baiken's method assumed a single-weapon opponent — Musashi carried two. Ganryu expected a serious swordsman meeting him with equal seriousness — Musashi met him with an oar, a towel, and a taunt engineered purely to unbalance his state of mind before the first blow landed.
Musashi's own writing, in The Book of Five Rings, names the underlying discipline directly: "Whenever you and your opponent become stagnant, you must immediately employ a different method of dealing with him in order to overcome him."1
And the deeper warning, from a contemporary of his — the Zen monk Takuan — goes further than technique: "It is a disease to be obsessed by the thought of winning... to be obsessed by the thought of using everything you have learned."1 Even fixation on your own skill, applied by reflex rather than freshly assessed each time, is its own form of fighting the last war.
Cross-reference added 2026-08-08 from the Strategy 28 build. Greene returns to Musashi in the one-upmanship chapter with a duel this page does not carry — Genzaemon of the Yoshioka family, Kyoto 1605 — and it is built as its own page rather than folded here, since this one is already at capacity and its title is bounded at three. Full treatment: The Duel Before the Duel.
The short version: Musashi arrives to challenge the head of Kyoto's most renowned sword family dressed like a beggar and so haughtily that Genzaemon assumed he thought himself the more famous man, then fails to turn up for the duel for three hours while Genzaemon and his students wait in a field at dawn — and wears a scarlet headband where Genzaemon wore the traditional white. Both men cut each other's forehead. Only one headband showed it.
What it adds to the pattern above is the axis. Matashichiro, Baiken and Ganryu are all defeated by an expectation about method being broken. Genzaemon is defeated by an expectation about conduct — dress, punctuality, the correct colour of a headband — before any technique is involved at all.
Which sharpens what connects the duels. It is not that Musashi never repeated a technique. It is that he never fought on the axis his opponent had prepared, and the Yoshioka duel demonstrates that the axis need not be martial.
Enrichment added 2026-08-08. Greene returns to Musashi fifteen chapters later, in Strategy 17's Keys to Warfare, for a different kind of fight — and the technique there is the one that generalises furthest.2
Musashi on several occasions faced bands of warriors determined to kill him. The Yoshioka ambush above is one instance; Greene treats the recurring situation as its own problem with its own solution.
The sight of such a group produces two reactions and both are fatal. Most people hesitate — a fatal flaw in a samurai. Or they lash out violently, trying to kill as many as possible at once, and lose control of the situation.2
Musashi did neither, and Greene's account of what he did instead is almost boringly procedural:
He would place himself so that the men would have to come at him in a line or at an angle. Then he would focus on killing the first man and move swiftly down the line.2
Read what the placement accomplishes before any blade moves. A band is not a number of opponents; it is a geometry. Standing in the open, one man against eight is one man against eight. Standing where the ground, a wall or an angle forces them to approach in sequence, it is one man against one man, seven times.
The rest follows from that:
Instead of being overwhelmed or trying too hard, he would break the band into parts. Then he just had to kill opponent number one, while leaving himself in position to deal with opponent number two and preventing his mind from being clouded and confused by the other attackers awaiting him.2
That last clause is where the technique stops being tactical. The line is not only a physical arrangement, it is an attention management device — it removes the six men he cannot act on from the problem he is currently solving.
And there is a compounding effect Greene records that reverses the psychology of the encounter:
as he proceeded down the line, they would become the ones who were intimidated and flustered.2
The men at the back watch the men at the front die one at a time, in order, with time to consider it.
How Greene generalises it. He transposes the whole thing to mental process, and it is the most portable instruction in the chapter:
Whether you are beset by many small problems or by one giant problem, make Musashi the model for your mental process. If you let the complexity of the situation confuse you and either hesitate or lash out without thought, you will lose mental control, which will only add momentum to the negative force coming at you.2
The procedure: divide the issue, place yourself in a central position, and proceed down the line. Plus a sequencing note that is not obvious and is easy to get backwards — begin with the smallest problem while keeping the most dangerous one at bay, because solving the small one creates physical and psychological momentum that helps overwhelm the rest.2
How this relates to the three duels above. It doesn't, quite, and the difference is worth naming. The three duels are each a bespoke solution to one opponent's specific expectation — Musashi as diagnostician, never repeating himself. This is the opposite: a standard answer, applied identically every time the same situation arises. Greene draws the anti-formula lesson from the duels and the formula from the bands, in two different chapters, and never observes that his exemplar of never fighting the same way twice had one move he used over and over.
The reconciliation available: the duels are about defeating a person, where the variable is their mind and no two are alike. The bands are about defeating a situation, where the variable is geometry and the geometry repeats. Fresh assessment for people; a fixed procedure for structures. See Strategy 17: The Divide-and-Conquer Strategy for the fuller context. [POPULAR SOURCE]
You're heading into a difficult conversation with someone who's watched you handle three previous conflicts the same way — calm, measured, give them room to talk first, then respond carefully. It's become a known quantity to them. They've prepared for it, the way Matashichiro's men prepared for your usual lateness.
Before you walk in, run the Musashi diagnostic: what does this person expect from you, specifically, based on the pattern you've already shown them? Not "what's the right approach in general" — what have you personally trained this opponent to anticipate?
If the honest answer is "measured patience," consider what Musashi considered: not becoming more aggressive (that's just fighting harder with the old weapon), but changing the actual shape of the engagement — naming the real issue first, before they've built their opening argument, the way Musashi struck first against an ambush built around his lateness.
The goal isn't to be unpredictable for its own sake. It's to notice, specifically, what the other side is counting on you to repeat, and to decline to supply it.
There's a real trap on the other side of this advice, worth naming before it gets applied carelessly: breaking pattern every single time is itself a pattern, and a predictable one — "this person always does the unexpected thing" is exactly as exploitable as "this person always does the expected thing." Musashi didn't have a rule that said always surprise. He had a habit of correctly reading what a specific opponent had come to expect, which sometimes meant breaking his own pattern and sometimes — against opponents who expected him to break it — might have meant not breaking it at all. The discipline is diagnosis, not a reflex toward novelty.
These are samurai legends, transmitted through centuries of retelling before reaching print — closer to folklore with a documented historical figure at the center than to verified biography. [POPULAR SOURCE], and a step further: even within the popular-source category, duel narratives like Ganryu's are the kind of story that accrues embellishment with each retelling precisely because they're satisfying to tell.
The core outcomes (Musashi's undefeated record, his reputation as Japan's greatest swordsman) are well-attested. The specific dialogue and staging details (the oar-whittling, the exact taunt) are almost certainly reconstructed or invented in transmission.
Musashi's own The Book of Five Rings, written near the end of his life, is a firmer source than the duel legends themselves — but it's also retrospective, a master's late attempt to systematize a lifetime of instinct into teachable principle. The gap between "what Musashi actually did in the moment at twenty-one" and "what the sixty-year-old Musashi later said the twenty-one-year-old had been doing" is exactly the kind of gap every practitioner-turned-theorist account carries, and it's worth holding lightly rather than treating the book's retrospective clarity as proof the young swordsman was consciously executing a named doctrine in real time.
Open tension: the chapter frames each duel as proof that adaptability beats fixed method. But all three of Musashi's opponents also had fixed methods that had worked many times before this specific fight — Baiken was genuinely undefeated. The story as told risks implying that any fixed method is inherently inferior to adaptability, when the more precise claim is narrower: a fixed method becomes exploitable specifically once an opponent has had the chance to study and predict it, which is a description of Musashi's opponents' situation, not a universal law about consistency itself.
Set this beside Hohenlohe and the Fossilized Army, the chapter's other central case.
Hohenlohe commanded an institution with decades of doctrine behind it, unable to adapt because the whole apparatus of Prussian military culture was built to enforce the old method. Musashi is the opposite limiting case: a single mind, no institution, free to discard a winning technique the instant after using it, accountable to no one's expectations but his own.
The pairing sharpens the real difficulty in applying Greene's strategy: an individual's adaptability is nearly frictionless — a decision made in a moment. An institution's adaptability requires first dismantling machinery built specifically to prevent exactly that kind of change, which is why Jena took an afternoon to become a disaster and Musashi's reversals took seconds to become victories.
Creative Practice → Breaking Your Own Signature Style on Purpose. An artist who has developed a recognizable, successful style faces a version of Musashi's problem the moment critics or audiences start correctly predicting the next move — the very success of the style becomes the vulnerability. See the general pattern in Constraint-Driven Coherence's treatment of deliberate formal disruption.
The insight the pairing produces: Musashi's willingness to show up with an oar instead of a sword is not chaos for its own sake. It's the same disciplined move as a novelist deliberately breaking their own established voice at exactly the moment it's become predictable enough for a reader to anticipate.
Psychology → Cognitive Load and Opponent Modeling. Game-theoretic and psychological research on adversarial interactions shows that predictability isn't just a tactical weakness — it measurably reduces an opponent's cognitive load, freeing their attention for offense rather than defense, because they no longer have to actively model what you might do.
The insight: Musashi's opponents weren't just wrong about his specific move; they were operating with more spare mental capacity than they should have had, because they'd stopped actively computing the possibility space and started running a lookup table. The oar didn't just surprise Ganryu tactically — it forced him back into expensive, real-time computation at the exact moment he could least afford the distraction.
Both handshakes point at the same underlying fact from different directions. The creative-practice reading says predictability is a vulnerability an audience or opponent can eventually exploit. The cognitive-load reading explains the actual mechanism by which that vulnerability operates — not that the opponent "figures you out" in some vague sense, but that they literally stop spending mental effort modeling you, freeing that effort for their own attack. Break the pattern, and you're not just surprising them. You're forcing their attention back onto a computation they'd stopped running.
Musashi's opponents didn't lose because they were worse swordsmen. Baiken, by all accounts, wasn't. They lost because they had correctly identified a pattern in an opponent who had already decided, before the fight started, not to be a pattern anymore.
The lesson isn't "be unpredictable." It's narrower and harder: know exactly what the other side thinks they know about you, and treat that specific knowledge as the thing to break.