A hundred armed men needed to walk into a city that had been ordered to stop exactly them. Shaista Khan had decreed that no Maratha enter Pune without a pass, armed or unarmed. And on the night of 5 April 1663, Shivaji's band walked in anyway. Nobody fought their way through the gate. They talked their way through it. The disguised entry into Pune is the part of the Lal Mahal raid that is pure cover-story — the moment where a sword would have failed and a plausible lie succeeded. It is also a small, sharp lesson in how history is made, because two reliable chroniclers both watched this entry and tell it in ways that cannot both be true.
Shaista Khan's order was explicit. Khafi Khan, whose father served in the Khan's army, records it: "no person, especially no Mahratta should be allowed to enter the city or the lines of the army without a pass, whether armed or unarmed, excepting persons in the imperial service," and to be safer still, "no Mahratta horseman was taken into service."1 Whether the rule could really be enforced is doubtful — the Mughal army held a sizable number of Maratha soldiers and was recruiting more — but the intent was a sealed town.1
The approach was as careful as the entry. Shivaji left Rajgad in early April with four hundred trusted men and wound through the forests to Sinhagad, close to Pune.2 Along the roughly forty-kilometre stretch from Sinhagad to Pune he dropped small detachments at intervals — for reconnaissance and to cover the way back if the return turned into a fight.2 Two larger groups, under Netaji Palkar and Moropant Pingle, took positions a mile from the camp's edge.2 On the evening of the 5th, Shivaji and a chosen band dismounted and left the horses with a detachment under Sarjerao Jedhe on the banks of the Mula river, just across from Pune.2 From there they went on foot. At the gates, in the night, they were stopped and questioned — who are you, whose men, where have you been — and at that tensest moment, the roughly hundred-strong force somehow slipped inside.3 How they did it is exactly where the record forks.
A cover story works on one principle: give the guard a category he already expects, so his own assumptions do the work of waving you through. Both surviving accounts run on that principle; they just disagree about which expected category Shivaji used.
In Sabhasad's version, the category is "us." Babaji and Chimnaji Bapuji of Khed — described as among Shivaji's "favourites" — walked at the head of the column, Shivaji behind them carrying a shield and a sword, the men behind him.4 When the vast Muslim army questioned them at various points in the camp, the two lieutenants answered as they walked: "We belong to the army and had gone on sentry duty."4 The meaning was that they were Deccani soldiers of the imperial army returning to their posts. The disguise was no disguise at all — just confident men claiming to be the night watch coming back.4
In Khafi Khan's version, the category is spectacle, doubled. He describes twin acts. One group of Marathas, serving as foot-soldiers, went to the kotwal and got a pass for two hundred Marathas accompanying a marriage party; a boy dressed as a bridegroom, his face hidden under chaplets of champak blossom, entered with drums and music.5 The same day a second party came in on the report that prisoners had been taken at an outpost — they posed as imperial captors marching their captives in, bare-headed and pinioned, abusing them aloud as they went.5 Both flashed the right papers. Both gave the guards a scene they could file under "ordinary." A wedding and a prisoner column are loud, public, expected things; nobody hunts for an assassin inside a brass band.
This page hands the vault two things. The first is a catalogued instance of deception-as-entry — the tactic of breaching a hardened perimeter not by force but by impersonating a category the defender already trusts. It feeds directly into The Night Raid on the Lal Mahal, which is the deed this entry made possible; this page is the threshold, that page is the room. The second gift is epistemic. Because two trustworthy chroniclers give incompatible accounts of the same fifteen-minute window, this page is a working example of how a vault should hold contradiction — naming both versions, refusing to merge them, and treating the disagreement itself as data about how the event was remembered.
Set the two stories side by side and the disagreement is not a detail — it is the entire mode of entry. Sabhasad's Marathas walk in quietly as returning sentries. Khafi Khan's Marathas walk in loudly as a wedding and a prison detail. One is camouflage by blending; the other is camouflage by being too conspicuous to suspect. Both cannot have been the method on the same night for the same hundred men.
Notice who is telling which. Sabhasad is the Maratha court chronicler; his version centers Shivaji's own lieutenants, the named Bapujis, and keeps Shivaji armed and close behind them — a portrait of disciplined infiltration by trusted men.4 Khafi Khan is the Mughal historian with a family connection to Shaista Khan's army; his version is richer, more cinematic, and more flattering to Maratha cunning in a backhanded way — the enemy fooled the Khan's own kotwal into signing the pass that doomed him.5 A hostile source has reason to dramatize the deception, because a clever trick explains the disgrace better than plain negligence does. A friendly source has reason to keep the entry tight and credit it to named heroes.
The honest reading is that we do not know which happened, and the source's tradition shapes the shape of the story. Purandare names this directly: "there are two theories of how they slipped in undetected," and he lays both out without choosing.3 That refusal is the right move. The two accounts may even both contain truth — a hundred men might have entered in more than one group, by more than one ruse, which would let both chroniclers be partly right and neither be wholly so. But the text does not license that reconciliation, and so it stays an open fork. What the case study teaches is that an event can be solidly attested and still genuinely unknown in its mechanism, because the witnesses who agree it happened disagree on how.
You are holding two accounts of the same night, from two people you have reason to trust, and they describe different events. The instinct is to pick the better-written one, or to split the difference into a tidy blend. Resist both. Lay them on the table side by side and read them as two reports, each shaped by who filed it. Ask of each: what does this teller gain by remembering it this way? The Maratha chronicler gains disciplined heroes; the Mughal historian gains a cunning enemy who explains away a humiliating failure. Hold the gap open. Write down both, mark which rests on a single chronicler, and let the contradiction stand as part of the record rather than a problem to be solved. The disagreement is telling you something true — that the entry was confused, fast, and seen differently by friend and foe — even though it withholds the one fact you wanted.
The deception-as-entry tactic and the reading of it both have a signature failure, and it is the same impulse: smoothing. In the field, the cover story fails when it stops matching what the guard expects — a wedding party at the wrong hour, a prisoner column with no warrant, a "sentry" who cannot name his unit. The lie has to ride the defender's own assumptions; the moment it asks him to believe something unusual, it dies at the gate.
In the record, the failure is the historian's version of the same sin: forcing the two accounts into one clean narrative because a single story reads better. You can spot the smoothing when a retelling silently picks the bridegroom over the sentry, or stitches them together without flagging that the sources never agreed. The diagnostic sign is a too-confident account of a moment the witnesses themselves were unsure about. Where you see seamless certainty over the Pune gate that night, someone has quietly resolved a contradiction the sources left open.
Everything here comes through Purandare's popular biography, which quotes the older chroniclers directly. [POPULAR SOURCE] The central tension is structural and unresolved: the two entry methods. [SINGLE SOURCE] Sabhasad alone gives the returning-sentry account; [SINGLE SOURCE] Khafi Khan alone gives the twin wedding-and-captives ruse. The two cannot be true in the same form for the same band, and Purandare deliberately declines to choose, presenting them as "two theories."3 The approach details — four hundred men, the forest route to Sinhagad, the dropped detachments, Sarjerao Jedhe holding the horses at the Mula river — come through Purandare and read as the settled Maratha tradition rather than independently verified logistics.2 The foreknowledge of how entry could be gained is credited to Shivaji's spies, which the source asserts but does not document in primary detail.2 The number at the gate (about a hundred) and the questioning are reported in common across the framing before the fork.3
The two voices Purandare sets against each other, Sabhasad and Khafi Khan, converge on the frame and split on the mechanism — and the split is the whole interest of the page. They agree that a roughly hundred-strong Maratha band entered sealed Pune by deception on the night of 5 April 1663, that they were questioned, and that they got through without fighting.3 That a Maratha court chronicler and a Mughal historian with family ties to the victim's army agree on this frame is strong evidence the frame is real. Then they part completely on how. Sabhasad's quiet "we had gone on sentry duty" and Khafi Khan's loud bridegroom-and-captives spectacle are not two angles on one scene; they are two different scenes.45 The Maasir-i-Alamgiri, the official Mughal record, is silent on the entry mechanism altogether, consistent with its broader embarrassment at the episode. What the disagreement reveals about deception-as-entry is that successful infiltration tends to leave a confused record by design — the very chaos and speed that let the band through also blurred what witnesses could later report, so the cleanest operations may be the ones history can least reconstruct. The contradiction is not a flaw in the sources. It is a fingerprint of the tactic working.
The disguised entry connects outward through the craft of the cover story and the structure of a perimeter that a cover story can defeat.
Sun Tzu — Intelligence and the Five Spies. Sun Tzu builds his intelligence system on agents who operate under false identity — men who pass for something they are not, inside the enemy's own space. The Pune entry is that doctrine compressed into a single night: the raiders pass for the night watch, or for a wedding party, or for prison guards, each a false identity sized exactly to what the gate expects. The structural parallel is the impersonation of a trusted category. The tension is scale and duration — Sun Tzu's spies live their cover for months; Shivaji's men wear theirs for minutes. The insight neither produces alone: a cover identity does not need to be deep, only deep enough to survive the single moment of inspection it must pass — a wedding band needs no biography, just a pass and a drum. Deception scales down to the duration of the threat, and a one-night lie can be as effective as a one-year legend if it is matched to the checkpoint it must clear.
Deception and Misdirection Tactics. The behavioral-mechanics treatment of misdirection turns on directing the target's attention toward a plausible expectation so the real action passes unwatched. Both Pune cover stories do exactly this. The wedding party and the prisoner column are loud, attention-grabbing, and entirely ordinary; the guard's eye goes to the bridegroom's blossoms and the abused captives, and not one man in that crowd reads as an assassin walking toward the general's bed. The sentry story works the same way at lower volume — it directs attention to the mundane category "returning watch" so nothing about the band invites a second look. The structural parallel is attention-routing: the deception does not hide the men, it hides their purpose by giving the observer something else to file them under. The tension is that misdirection in the manual is usually a sustained social performance, while here it is a single-pass gate trick. The insight neither domain produces alone: a checkpoint is the most concentrated misdirection problem there is — one observer, one moment, one decision — and the optimal cover is therefore not the most elaborate disguise but the most boring expected category, because boredom is what makes the guard stop looking. The raid teaches the tactician that the strongest cover story is the one that bores the gate, and the manual teaches the historian why a band of armed men could walk past sentries who had been ordered to stop exactly them.
Religious Tolerance as Political Philosophy. The link here is sideways but real: the cover stories worked because Shivaji's men could pass as part of a religiously and ethnically mixed Mughal army — Deccani soldiers in imperial service, a category porous enough that Marathas moved inside it. The same heterogeneity that made the Mughal force formidable made its perimeter unsealable; an army that recruits everyone cannot reliably tell its own from the enemy at the gate. The political-philosophy page treats pluralism as a source of strength and legitimacy; the raid exposes its security cost. The insight neither domain gives alone: tolerance and porousness are the same structural property read from two sides — what makes a polity inclusive and broad is exactly what makes its boundary permeable to a determined impostor. Shaista Khan's order banning Maratha entry was an attempt to seal a boundary that his own army's composition had already dissolved.
The Sharpest Implication. The page says that the best-executed deceptions leave the muddiest historical record — and that this is not a bug but a consequence. We cannot say how Shivaji entered Pune precisely because the entry worked: it was fast, confused, and seen differently by everyone. The destabilizing thought for anyone who trusts the historical record is that clarity in the sources may correlate with failure in the event, and that the operations we can reconstruct in clean detail may be the ones that went wrong loudly enough to be documented.
Generative Questions. When two trustworthy sources give incompatible accounts of a deception, is the right inference that one is wrong, or that the deception itself produced both impressions — and how would we tell those cases apart? And does the porousness that let Shivaji's men pass as imperial soldiers point to a general law that the more inclusive an army, the more vulnerable its perimeter to impersonation?