There's a pin on a map at Marengo. Napoleon is on the floor among the charts and he puts a pin down beside a village, and later that's where the battle happens, and the pin is the proof that the endpoint was chosen rather than reached.
The pin comes from Bourrienne's memoirs, written after everyone knew how it turned out. And the campaign it's meant to prove was one in which Napoleon dispersed his own forces, was outnumbered two to one, got broken through at three in the afternoon, and was rescued substantially by Desaix acting on his own initiative.
Once you've seen that, you see it everywhere in the book. Alexander I's retreat before Napoleon is presented as a design executed to perfection — but the retreat was substantially compelled, it was politically toxic, Barclay de Tolly was replaced because retreating was unpopular, and Kutuzov then fought Borodino, which is precisely the decisive engagement the strategy supposedly existed to prevent. De Gaulle arrives in London in June 1940 with a fully formed five-stage plan, in an account where every stage depended on something nobody could have forecast. Louis XI's ten-year courtship of the Swiss was a device for destroying Burgundy from the first day. Dessalines's entire period of French service was a ploy. FDR's 1940 draft was premeditated from the start, in a record where his private intentions are genuinely disputed.
Six cases. Same move each time. Wherever the record is ambiguous about intention, the ambiguity resolves toward design.
[RESONANCE] Because it isn't a lie and that's what makes it interesting.
Not one of those readings is refuted by the evidence. Each is a possible reading of a record that also supports a duller one — that Napoleon improvised brilliantly, that Alexander was pushed, that de Gaulle took what was in front of him, that Louis cultivated options patiently and used the one that ripened. The duller reading is usually at least as well supported. It just doesn't produce a transferable technique, because "he responded well to things he didn't choose" is not something a reader can go and do on Monday.
So the bias isn't toward falsehood. It's toward the reading that yields an instruction. Which is a much harder thing to catch, and a much harder thing to argue about, because you can't point at a wrong fact.
The vault now holds this pattern documented case by case, and the interesting part is what it costs the reader rather than what it costs the history.
If designs are what produced these outcomes, then studying the design is the right thing to do and the book is correctly built. If most of them were skilled improvisation inside conditions the actor didn't set, then the transferable skill isn't planning at all — it's reading a situation fast and committing, which is a different competence, trained differently, and this book doesn't teach it because it can't see it.
Marengo is the sharpest instance because the chapter's own title is about branches — plans with room to move — and the case underneath it is a man improvising his way out of a mess he made. The chapter has the better lesson inside it and reaches past it for the pin.
Two connections the build already surfaced and this spark ties together. First, this is the same operation as the Tet problem seen from a different side: there, one event supplies three causal roles; here, one ambiguous record supplies whichever intentionality the chapter needs. Second, it is the specific mechanism behind Greene's own admiring account of how Machiavelli wrote — historical anecdotes stitched in as "cleverly altered versions." Resolving ambiguity toward design is the alteration, and it's the most common one in the volume.
Against the vault's own history corpus, which mostly does the opposite — Tet Offensive Media Coverage and the Miracle of the House of Brandenburg both preserve the contingency rather than converting it. The vault and the book disagree methodologically, not factually, and that's a cleaner disagreement to develop than any single case.
Promotion criterion: find a case anywhere in the 33 chapters where an ambiguous record is resolved away from design — where Greene says a figure got lucky or improvised. The build found exactly one candidate, Frederick and the czarina's death, and Greene converts even that into evidence of method, against contemporaries who named it a miracle precisely because they understood it as chance. If a genuine counter-instance exists, this is a tendency. If none exists across a book this long, it's a rule.
Search the built corpus for the phrase-level tells — "from the beginning," "his plan was always," "he had foreseen" — and count. Then check whether the same tells appear in the Laws of Human Nature and Mastery corpora already in the vault, which would make this Greene's habit rather than this book's.