You are losing a game of chess. The other player is better than you — sharper, more practiced, reading three moves ahead. So you stop playing chess. Mid-move, you sweep the pieces aside and deal a hand of poker. His chess skill is now worth nothing. He is worse than behind — he is in a game he didn't agree to enter, scrambling to figure out the new rules while you are already playing.
That is sankaikawari, the "mountain-and-sea change." When your attacks keep bouncing off a strong defense, you don't push harder. You change so completely that the gap between your old approach and your new one is as wide as the difference between a mountain and the sea.1 The size of the change is the weapon. A small shift, your opponent absorbs without breaking stride. A total shift breaks his footing.
Sankaikawari is the maximal version of a broader principle Lovret calls kawari — changing.1 The general rule is simple: switch your strategy at the exact moment your opponent settles into reading you.1 In a typical fight, he spends the opening seconds studying you — throwing feints, watching how you react — and only commits once he thinks he has your pattern.1 The instant you feel him getting confident and aggressive, you change. He has to throw out everything he just learned and start the reading process over, from the back foot.1
Sankaikawari takes that and pushes it to the limit. You don't make a small adjustment — you make the largest possible one. Shift from attack to defense. Switch from high cuts at the head to low thrusts at the body. Trade a long-range, mobile jabbing style for a close-in, heavy-hooking one.1 The change must be instant — flip it like a light switch, light to dark in one motion — because a gradual shift just lets him adapt as you go.1 And it must be genuinely large. Lovret is exact about this: a shift between short runs and long passes works, but a swap between screen passes and sweeping runs doesn't — those two are too alike to exploit anyone's inertia.1 [PARAPHRASED] [POPULAR SOURCE]
Things have inertia. So do ideas.1 A body moving fast in one direction cannot instantly reverse. A mind locked into a winning rhythm cannot instantly abandon it. Sankaikawari attacks both kinds of inertia at once.
The internal logic is brutally linear: the greater the change, the greater the confusion.1 When you make a small adjustment, your opponent's existing model still mostly fits — he tweaks it and keeps going. When you change everything, his model is worthless, and the cost of building a new one is the whole point. While he is building it, he cannot attack. He has to curb his aggression and start over: moving carefully, feinting, testing how this new version of you reacts.1 If you keep doing this, he is always on defense and you are always attacking.1
There is a deliberate setup behind it. To build the strongest mental inertia, you let a strategy work first.1 The longer his approach succeeds, the more committed he becomes to it, and the harder it is to drop. A change you spring on an opponent who is locked into a pattern of success lands far harder than the same change after he has already failed a few times.1 You are not just surprising him — you are pulling the rug at the moment he is standing most confidently on it.
Every strategy you run emits information. The opponent's whole opening game is information-gathering — feints, probes, watching your reactions — until he has assembled a working read.1 Sankaikawari treats that accumulated read as a liability you can detonate. The moment his model of you is most complete is the moment a total change costs him the most.
This connects to the vault's broader map of variation and timing. The principle of switching before he finishes reading you is the same engine that drives Magiri — Constant Variation, where a fighter never repeats a pattern long enough to be modeled. Sankaikawari is what magiri escalates into when small variation stops being enough — when the defense is too strong to be beaten by tweaks, and only a regime change will do. It also leans on Katsuri — Changing Speed, because the change is useless if it is slow: tempo is part of the disorientation. And it presupposes the difference between a single move and an overall plan — see Technique vs Strategy Distinction — because sankaikawari is not a new technique, it is a wholesale swap of the set of techniques you are drawing from.1
Lovret's clearest example is a football offense.1 The team opens with a series of long passes, deliberately spreading the defense thin across the field. The defenders adapt — they drop back, spread out, commit to covering deep. That adaptation is exactly what gets exploited. The offense suddenly switches to a running attack into the now-undermanned middle, and it works.1
Reverse it and the logic holds. Start with a string of running plays. The defense tightens up, packs the box, braces for the run. Then you throw the long pass over a defense that has pulled itself in.1 In both cases the success comes from the size of the swing — run to deep pass — not from any single brilliant play. Lovret's counter-example seals the point: screen passes and sweeping runs are too similar; switching between them buys you nothing, because the defense's inertia is the target and there isn't enough difference to catch it.1
The army version is the same shape. March your armor integrated with your infantry, taking strong points methodically, fighting set battles. The enemy adapts by concentrating his defense along your line of advance. The moment he does, you free your armor to operate alone — striking deep, bypassing the strong points entirely — and while his commanders scramble to adjust, you hit with the infantry.1 His concentration, which was the right answer to your old style, is now exactly the wrong shape for your new one.
You are a karate student preparing for matches, and you start long before any opponent shows up. You sort every technique you know into sets, grouped by similarity.1 One set is kicks, thrown from an erect, mobile posture. A second is heavy punches, delivered from a low, planted stance. A third is fast, snappy techniques off a light, moving base. You drill each set on its own until fighting in that single mode feels completely natural — until each one is a whole person you can become.1
Now the match. Your opponent opens by studying you — a few feints, watching how you answer.1 You give him one style and only that style. The kicks. You hold the line: no matter how tempting a punch from another set looks, you do not reach for it. You let him work. You let him start to win the read.
Then you feel it — he gets aggressive, the small confident lean of a man who thinks he has solved you. That is the trigger. You change, instantly and totally. You drop into the low stance and start landing heavy punches. Even your personality shifts; it should feel to him like a new opponent walked onto the floor.1 His aggression dies. He has to back off, move carefully, feint again to learn this stranger. While he is rebuilding, you are scoring. You ride the effective style as long as it lasts — and the instant he starts defending it well, you change again.1
You can tell sankaikawari is failing by how often you have to do it. Done right, the change is so large and so fast that one application should be enough; if you find yourself swapping strategies again and again inside a single fight, the usual cause is that you are changing too slowly, giving your opponent time to adapt to each new style as it arrives.1 A change he can adapt to is not a mountain-and-sea change at all — it is a small change wearing the name.
The second failure mode is changing into something too similar to what you left. Screen passes to sweeping runs. Economy ads to slightly-better-economy ads. The gap is too small to catch anyone's inertia, and the opponent simply slides his existing model over to fit.1 The third is timing: spring the change after he has already failed, instead of while he is winning, and you forfeit the strongest version of his commitment — there is no locked-in pattern to break.1
The evidence here is a practitioner's pattern-library, not a controlled study — fencing, karate, boxing, football, advertising, and armored warfare, all asserted to share one structure.1 That breadth is the claim's strength and its risk. The strength: the same shape recurs across domains that don't borrow from each other. The risk: a single author selecting his own confirming examples can make almost any maxim look universal. [PARAPHRASED] [POPULAR SOURCE]
A real tension sits inside the doctrine. Lovret's general rule is to pick a strategy your enemy doesn't know, precisely because using a familiar one reveals your hand.1 But the counter to sankaikawari violates that rule on purpose — you answer a big change with your own big change, fighting fire with fire, even though it shows familiarity.1 He resolves it by exception rather than by reconciling the two principles, which leaves an open question: when does the value of surprise outrank the value of an unknown strategy, and is "your opponent just used sankaikawari" really the only case? An open practical question follows: how many genuinely distinct technique-sets can one fighter actually maintain at full readiness before the sets blur and stop being mountains and seas to each other?
Lovret is the sole source here, so the tension is internal to his own system rather than a clash between authors. The interesting friction is between his default counsel and his counter-doctrine. Across the kawari material he treats the unknown strategy as the prize — keep your enemy from modeling you, switch before he finishes reading, never hand him familiarity he can use.1 Then, at the one moment an opponent deploys a mountain-and-sea change against you, he reverses: don't try to adapt to the new style, and don't worry that fighting fire with fire reveals what you know — take a third path instead.1
That "third path" instruction is the sharpest thing in the chapter and the place where Lovret converges with himself at a higher level. Adapting to your opponent's change means letting him set the terms of the new game — he chose the shift, so he is ahead in it. Picking a third option refuses the binary entirely. When the competitor swings from stressing economy to stressing performance, you don't chase performance; you jump to quality or style.1 It is the same move as the master metaphor: don't get better at his new game, change the game again. The convergence is that sankaikawari and its counter are one idea — refuse to play on the board your opponent prepared — applied first as offense and then as defense.
The mountain-and-sea change is a strategy of inertia exploitation, and that engine drives the same effect in two domains far from the dueling floor.
In behavioral-mechanics, sankaikawari is the structural twin of the pattern-interrupt — the disorientation tactic where a practitioner deliberately breaks an expected sequence to create a brief window of suggestibility. A handshake (the social ritual) interrupted mid-grip, a sudden non-sequitur in a sales pitch, an unexpected tonal swerve in an interrogation: each one strands the other person's running model of "what is happening here." The structural parallel is exact — both weaponize the cost of rebuilding a mental model under time pressure, and both depend on the target having first settled into a confident read. Lovret's instruction to let the opponent succeed first maps directly onto why pattern-interrupts work best on people who are cruising on autopilot rather than people already alert and uncertain. The insight that neither domain generates alone: disorientation is not produced by novelty as such but by the destruction of an investment. The interrupt only bites in proportion to how much the target had committed to the broken pattern — which means the tactician's real job is not the surprise itself but the patient setup that makes the surprise expensive. A pattern-interrupt sprung on someone who hadn't yet committed to a pattern is just noise, exactly as a "mountain-sea change" between two similar styles is just motion.
In business, sankaikawari is the structure of a category-redefining pivot that strands a competitor's expertise. Lovret's own example is advertising: shift from competing on economy to competing on performance, and the rival is helpless until it builds a new campaign — and when it copies you to performance, you jump again to quality or style.1 At full strength this is the chess-to-poker move at company scale: a firm losing a feature war redefines the category so that its competitor's accumulated advantage — the deep expertise that made it the better chess player — suddenly applies to a game no one is playing anymore. The structural parallel: the competitor's strength is built on a model of the market, and a large enough redefinition makes that model the wrong shape, the same way the football defense's correct adaptation to deep passes becomes exactly the wrong posture for the run. The insight neither domain produces alone: the deeper a competitor's expertise, the more exposed it is to a category change, because expertise is inertia with a résumé. The very investment that makes a rival formidable inside the current game is the investment that cannot be redeployed when the game changes — which inverts the usual intuition that the more skilled player is the safer one. The "third path" counter is the defensive half: when a rival redefines the category against you, don't chase their new axis (you'll arrive late and behind), redefine it a third way yourself.
The Sharpest Implication. The weapon is not your new strategy — it is your opponent's investment in his old read of you. Sankaikawari converts his competence into a vulnerability: the better he has modeled you, the more it costs him when the model dies. This flips the usual logic of skill. You don't beat a stronger opponent by getting stronger inside his game; you beat him by changing games so totally that his strength has nowhere to land. The discipline this demands is counterintuitive — you must let him win first, hold a single style while he builds confidence, and resist the tempting better move, because the size of his eventual confusion is built entirely from the size of his prior commitment.
Generative Questions.