This is the first of ten reports working through Shivaji from angles that don't usually get put next to Maratha history — psychology, craft theory, organizational behavior, and more, each one built on the report before it. This one is the foundation. Before the arc can ask what Shivaji's tactics meant, or what his mother installed in him, or what his fort-building revealed about his instincts, it has to settle a more basic question: where did the tactics actually come from? The story usually told is that Shivaji invented a new way of fighting — the hit-and-run, terrain-eating, never-stand-still doctrine historians call ganimi kava. The vault's own sources say otherwise. A generation before Shivaji was born, an Ethiopian former slave named Malik Ambar was already running the same system, under a different name, against the same empire. This report treats that claim the way you'd treat any handoff of specialized skill from one operator to the next — what exactly moved, how strong is the paper trail, and does the label "genius" survive contact with the evidence.
Picture the problem first. You command ten thousand horsemen. Your enemy commands a hundred thousand, with siege artillery and a treasury that never runs dry. You cannot win a straight fight, ever, not once. So you stop offering one.
That's bargi-giri — Malik Ambar's answer to the Mughal empire, named after the bargir, the light raiding horseman who carries no baggage and lives off the land. The mechanics are simple to state and brutal to survive. Never let the enemy's army stand still long enough to use its size. Hit the column's edges — the foragers, the baggage train, the stragglers — and vanish before the heavy formation can wheel and answer. Target the enemy's revenue, not his army, because an empire that can't collect taxes stops being an empire faster than one that loses a battle; Ambar's 1610 sack of Surat, the Mughals' richest port, was a raid on the treasury, not a battle for territory.1 Use terrain the enemy's heavy formations can't cross — Ambar chose the ground at Bhatvadi in 1624 specifically because flooding it turned Mughal artillery into dead weight stuck in mud.2 And run a second front purely on nerves: night raids with no goal but to keep the enemy's army awake, so that "terror seized the hearts of the enemies' troops" who "spent their nights without sleep and their days without repose."2 Put together, none of these moves is decisive on its own. Stacked, they add up to something a Mughal historian described in one line that is also the whole doctrine: "Malik Ambar was sometimes defeated, and sometimes victorious, but did not cease to oppose."3 The win condition isn't a battle. It's refusing to disappear.
Feel what that means from inside the heavier army. You're riding with the Mughal vanguard, deep in the Deccan, dust from your own column thick enough to taste. The baggage train stretches for miles behind you — grain carts, the siege guns you'll need if the enemy ever stands to fight, the elephants. He never stands. At dawn a band of light riders comes out of the scrub, hits the supply carts, and is gone before the alarm is fully up. By noon the wells ahead have been fouled. By dusk it's the rearguard's turn. No battle has been offered anywhere, which means there's nothing for your artillery — the empire's real advantage — to say anything to. Every day costs you men, grain, and nerve, and the enemy grows no weaker while you grow hungrier. That's the mechanism working as designed: not a fight you lose, a fight that's never offered, which is worse.1
The system has exactly one way to fail, and Ambar demonstrated it himself in 1616, which is part of why the doctrine is legible at all — you can see its edges precisely where it broke. At Paithan, on the banks of the Godavari, Ambar resolved to attack with forty thousand cavalry on a fixed day, abandoning the raid-and-vanish logic for a single set-piece assault. The Mughals, forewarned, dug a trench and planted their artillery behind it. Ambar's vanguard rode straight in and "could neither advance nor retreat, being huddled together as if they had been chained," and was "mown down by the Mughal artillery, like leaves of a tree under a destroying wind."17 The mobility that was the bargir's entire armor was gone the moment he agreed to mass. Without it, a light horseman is just an underequipped man standing in front of a cannon. The lesson embedded in that one afternoon — never let the enemy convert your raiding force into a target — is the exact tripwire this report will need again once it asks what specifically transferred to Shivaji's own doctrine, and what he built in response to avoid ever tripping it himself.
Here's the part that turns a resemblance into a documented handoff. Shivaji's grandfather was, in the vault's sourcing, "an associate of Ambar's."4 Shivaji's father, Shahji, spent his career fighting on for the Nizam Shahi cause — Ambar's state — after Ambar's death, and only entered Bijapur's service once that cause collapsed.5 The scholar Richard Eaton, quoted in the vault's source material, calls the Nizam Shahi state under Ambar "the nursery in which Maratha power could grow, creating the political preconditions for the eventual emergence of an independent Maratha state."4 That's not loose metaphor. It's a claim about where an entire generation of Maratha military families — including Shivaji's own — did their apprenticeship: inside a state whose entire operating doctrine was Ambar's raiding system, run by an officer corps that was, in the vault's phrase, "a joint Habshi-Maratha enterprise."6
The clearest single data point is the word itself. In 1616, a Bijapuri court chronicler described Ambar's cavalry with the line "every day Ambar's nobles used to come and fight like bargis [i.e. guerrillas]."7 Bargi wasn't Shivaji's coinage. It was already the working term for this style of war, in use by Ambar's contemporaries, a generation before Shivaji fought his first campaign. And the specific operations echo directly: Ambar's 1610 sack of Surat is credited in the vault's sourcing with giving "inspiration to the famous raid by Shivaji on that city five decades later" — same port, same empire, same logic, one human lifetime apart.8
The apprenticeship wasn't abstract, either. By 1610 Ambar commanded ten thousand cavalry "of his own caste" — fellow habshis — riding alongside forty thousand Dakhnis, the local Deccani troops, in a force the vault's sourcing calls "a joint Habshi-Maratha enterprise."6 That officer corps is where Shivaji's grandfather stood as an associate, and where Shahji kept fighting for years after Ambar's death. It's worth sitting with how unglamorous that transmission channel actually is: not a training montage, not a named mentor handing down technique, but two generations of a single family serving inside a state whose entire military identity was the raiding doctrine, absorbing its logic the way any officer corps absorbs its own institutional habits — by doing the job, watching it work, and doing it again. The same institutional proximity cuts both ways, though, and the vault's own sourcing doesn't let the "joint enterprise" framing go uncomplicated: on the eve of the 1616 disaster at Paithan, a number of Ambar's own Maratha comrades deserted to the Mughals, and Marathas sacked his capital at Khirki "almost to the ground."18 The same men who apprenticed inside Ambar's system could turn on it. Whatever transferred to Shivaji's household wasn't uncomplicated loyalty to a shared cause — it was proximity to a working method, which is a different and more durable kind of inheritance than loyalty would have been.
Lay the two systems side by side and the tactical grammar is close to identical: intelligence before movement, surprise over force, terrain as a weapon, speed standing in for size, and a refusal to ever offer the enemy a fixed target. The vault's own page on Maratha guerrilla doctrine names five operating principles — and every one of them is already present, fully formed, in Ambar's method decades earlier.9 That's not resemblance. That's the same doctrine wearing two names.
What changed is everything built around the doctrine. Ambar never had a fort network — his power sat inside someone else's sultanate, and when his son proved unequal to holding it, the whole apparatus collapsed within a decade.10 Shivaji built and held more than sixty forts as a distributed, redundant sovereignty network, with internal checks — the havaldar, sabnis, and sarnobat triad, mixed-caste appointments designed against treachery — that outlived any single defeat.11 Ambar had no navy; Shivaji did. Ambar ran scorched-earth denial as pure military expedient; Shivaji paired it with an administrative capacity to compensate the farmers whose stores he destroyed, which meant the tactic didn't burn through his own population's loyalty the way it would have burned through anyone else's.12 And Ambar was always a regent, propping up someone else's throne. Shivaji built a throne of his own. The raiding grammar transferred whole. The state that could turn a raiding grammar into a permanent kingdom did not exist yet when Ambar died, and had to be built from nothing.
The loyalty problem is the sharpest place to watch the difference. Ambar governed the same structural weakness his method exploited in others — a joint, multi-ethnic force with no single overriding identity holding it together beyond his own personal authority — and he survived four separate assassination attempts from within his own command before dying of old age.19 Shivaji inherited the same raw material (a mixed force, competing loyalties, a state built on conquest rather than settled tradition) and answered it institutionally rather than personally: the havaldar/sabnis/sarnobat triad at every fort split power three ways on purpose, with the keys to the gate always held separately from the two other offices, and appointments deliberately mixed across caste lines specifically to make coordinated treachery harder to organize.11 Ambar held his state together by being irreplaceable. Shivaji built a state that didn't need him to be.
Some of this chain is as solid as popular-history sourcing gets. The word bargi, independently attested for Ambar's men in 1616, is a real anchor — a name in circulation before Shivaji's birth for exactly the fighting style later credited to him alone. The "nursery" framing comes from a named scholar (Eaton), not just narrative color. The Surat-to-Surat echo, fifty years and one dynasty apart, is close enough to be more than coincidence.
But the sharpest claim in the chain — that Shivaji personally inherited the method through his grandfather's and father's proximity to Ambar's state — rests on a single popular source (Pillai) with no primary-document corroboration, no named training event, no account of anyone actually teaching Shivaji anything.13 It's an institutional-proximity argument, not a transmission record: the grandfather served in Ambar's orbit, the father fought for Ambar's successor state, therefore the son absorbed the doctrine. That's plausible. It is not the same thing as proven. And the vault's own sourcing undercuts the strongest version of the claim before it can even be made: Ambar himself may not have invented bargi-giri from nothing — he may have systematized a raiding style "already native to Deccan light cavalry," in which case both Ambar and Shivaji are heirs to a regional tradition with no single author, and "Shivaji inherited it from Ambar" collapses into "Shivaji inherited it from the Deccan," a much weaker and much less citable claim.14
This is worth running through the vault's own tradecraft vocabulary, because the question it's actually asking is a skill-transfer question, and the behavioral-mechanics corpus has already built the diagnostic for it. Chase Hughes's tradecraft mastery rubric distinguishes a technique you can recite from a technique that survives contact with a changed situation — Level 1 is recall in a safe room; Level 4 is the ability to "creatively modify and edit the [...] skill in real time to counteract [...] large changes to environment or culture."15 Shivaji didn't run Ambar's script. He ran Ambar's method against different terrain, different politics, and a different scale of ambition, and adapted it into forts, a navy, and an administration Ambar's playbook never had to include. That adaptability is itself evidence — not proof of person-to-person teaching, but evidence that whatever moved from Ambar's world to Shivaji's was a transferable method and not a memorized set of steps. Hughes's own technique-creation material makes the same distinction from the builder's side: a real tradecraft transfer hands someone the underlying models — read the enemy's logistics, deny the fixed engagement, hit revenue, never stop — not a fixed script to recite unchanged.16 Judged by that standard, what the vault can actually prove is a method-level inheritance riding on institutional proximity, dressed by later legend as personal genius.
Put the pieces together and the honest verdict has two parts, and they cut in opposite directions. Calling ganimi kava "Shivaji's invention" is not defensible — the tactical grammar is on the record a generation earlier, under a different name, inside a state where Shivaji's own family served their apprenticeship. But calling Shivaji's achievement "nothing new" would be just as wrong, and for the same reason the mastery rubric flags: he didn't just repeat what he inherited, he built the institutional scaffolding — forts, navy, administration, an ideological frame Ambar's regency never got to construct — that turned a raiding tactic into a durable state. Ambar could not have built that scaffolding even if he'd wanted to, because he was never sovereign; he was always managing someone else's collapse. Shivaji's real originality may not live in the tactic at all. It may live entirely in the institution-building the tactic made possible once someone finally had a throne of his own to put it under.
Notice, too, that this is exactly what the Paithan disaster predicts and Shivaji's institutions quietly correct for. Ambar's one recorded failure came from converting a raiding force into a fixed target and from governing a state that depended entirely on his own continued presence to hold together. Shivaji's fort network — the triad of offices at every gate, the mixed-caste appointments, the sheer number of independently defensible points — is the institutional answer to both failure modes at once: no single engagement can ever be forced to matter as much as Paithan mattered, and no single death can unravel the whole structure the way Ambar's own succession did within a decade. That's not visible anywhere in the tactical grammar itself. It's visible only in what got built around it, which is where this report's verdict actually lives.
That distinction is the hinge this whole arc turns on. The next report follows the same operational tradition into a different register entirely — not the raid, but the stage. The instincts that built a fort network out of a raiding doctrine also built Pratapgad: a fort engineered, before its first enemy ever arrived, to host a killing. If Shivaji inherited a tactical grammar from Ambar's world, the question worth asking next is what else that same inheritance shaped — including, it turns out, a taste for architecture as narrative device.
Threads: → Malik Ambar's Playbook as a Skill-Transfer Case — Thread Map