The Map That Wasn’t

How Mechanistic Doctrine Entered an Adaptive War—and What It Could Not See

Somewhere in Tehran, a man named Sepehr leaves his front door unlocked. It’s not due to carelessness but to calculation—an act of cautious adaptation to a city that has become unpredictable in ways you can feel through the soles of your feet. When the explosions start, his family will sprint for the underground parking garage. He needs every second. “The war might last weeks,” he told a journalist from Al Jazeera in early March 2026, the toxic smoke from burning oil facilities still hanging over Tehran’s ten million people. “My family and I will only leave if it gets too bad. For now, life goes on.”

Seventeen days into Operation Epic Fury. Life goes on.

That phrase—its exhausted, almost defiant ordinariness—captures something that the architects of this campaign did not fully understand. Not because they underestimated Iran’s military resilience, though the record now suggests they did. Not because the operational planning was sloppy, because it clearly was not. But because they misjudged the nature of the system they were entering. They stepped into a complex adaptive system with the mindset of engineers approaching a complicated machine. Their own machine—the US Military—worked nearly perfectly. But the map they used was wrong.

What follows is not an argument against using military force as a tool of statecraft. It’s a more uncomfortable argument: that the doctrine guiding how the United States plans and executes adaptive wars is fundamentally unable to see what matters most about them. And the consequences of that failure are now visible, in real time, across nine countries, a closed strait, and a city where a man keeps his door unlocked because he has learned to live inside somebody else’s plan.

The Plan

The logic behind Operation Epic Fury was not irrational. After years of diplomatic failures, a nuclear program that had outpaced every effort to constrain it, and a brutal crackdown that killed thousands of protesters, the United States and Israel concluded that the window for military action had narrowed to a single moment—now. Iran was weakened—economically, militarily, and diplomatically. Its proxy network had been heavily damaged by two years of Israeli operations. The June 2025 twelve-day war had gutted its air defenses and wounded its nuclear infrastructure. Its currency had collapsed, and its streets were filled with the largest protests since the Islamic Revolution.

The strategic calculus, such as it was, relied on a set of assumptions that, simply put, boiled down to this: Iran was a complicated problem that had reached a manageable moment. Strike the leadership, take out the missile arsenal, annihilate the navy, sever the proxy networks, and the internal pressure would do the rest. Regime change from within would follow the external shock. The four stated military goals—no nuclear weapons, no missile industry, no functional navy, no active proxy network—were the technical guidelines for a desired end state.

The operational preparation was truly impressive. Months of joint US-Israeli strategic deception manipulated satellite imagery, masked aircraft readiness, and ensured that on the night of the strike, Israeli leadership arrived at the Ministry of Defense in private vehicles, making sure satellite images would show nothing unusual. On the night of February 28, 2026, nearly 900 strikes fell across Iran within the first twelve hours—a series of precise hits targeting missile sites, air defense systems, command infrastructure, and the compound where Supreme Leader Ali Khamenei was meeting with senior officials. Khamenei was killed in the opening volley, along with senior officials whose deaths were confirmed in the first hours of what the White House confidently called Epic Fury.

The military machine performed exactly as designed. Every kinetic metric that could be measured was tracked and tabulated. Missile launch rates. Sortie counts. Targets destroyed. Interceptor success percentages. The operational reporting was meticulous, the press conferences sharp, and the objectives numbered and reiterated with the precision of a branding campaign.

What the plan couldn’t anticipate was the system it entered, because the system wasn’t a machine. It was alive, and it was already adapting before the first Tomahawk left its tube.

The System Pushes Back

Eight days after Ali Khamenei was killed, his son Mojtaba was chosen as Iran’s new supreme leader. The IRGC pledged allegiance within hours. Senior officials and religious scholars aligned. The proxy networks—Hezbollah and Iraqi militias—confirmed their loyalty without delay. Even the Houthis, conspicuously absent from the kinetic fight, declared solidarity and resumed Red Sea disruptions, calculating, for now, that political statements cost less than strikes. The succession the campaign’s planners might have imagined would trigger a vacuum, instead produced, with unsettling speed, the Iranian regime’s continuity.

This was not luck. It was architecture. The Islamic Republic spent four decades designing precisely this capability, not to win conventional wars against superpowers, but to survive them. The succession protocol is encoded in the constitution. The distributed proxy network is designed to operate on standing instructions without active command. The IRGC’s institutional identity is organized around a core narrative of external enemies that must be resisted at any cost. Killing the supreme leader did not erase that narrative. It intensified it and passed it on to his son with a fresh grievance attached.

Meanwhile, Iran’s response revealed something that the campaign’s target lists couldn’t fully capture: The adversary had co-evolved alongside the plan. The twelve-day war in June 2025 was its inflection point. After that conflict concluded with a US-declared ceasefire, Iran learned the strategic lesson more carefully than its opponents apparently believed. Its succession architecture, already encoded in the constitution, absorbed the blow faster than anyone anticipated. It developed drone capabilities tailored specifically to the Gulf’s geographic realities—encrypted, frequency-hopping unmanned boats disguised as fishing vessels that threaded the world’s most critical shipping lane. That lane is decisive terrain, not just key terrain, which aids whoever holds it. Decisive terrain must be held at all costs to secure victory. The Strait of Hormuz is decisive terrain. And we don’t hold it. Iran does.

By March 5, 2026, five days into the campaign, Iran had fired roughly 585 ballistic missiles and 1,522 drones. Its launch rates were decreasing due to ongoing suppression, but the system was not degrading at the same pace. It was evolving. In the opening hours, about forty percent of Iranian missile launches targeted Israel. By March 4, that number had fallen slightly above twenty percent. The system was adjusting its targeting priorities in real time, identifying new pressure points among Gulf Arab states that the US couldn’t fully defend, whose airspace was now filled with ballistic and drone threats traveling at velocities that compressed defender reaction times to seconds.

The Strait of Hormuz, just twenty-one miles wide at its narrowest point, was closed off. Explosive drone boats patrolled shipping lanes, thwarting electronic countermeasures. Brent crude oil surpassed a hundred dollars a barrel. The IMF announced that every ten percent rise in energy prices would add nearly half a percentage point to global inflation. Universities closed early in Bangladesh. Pakistan and the Philippines adopted four-day workweeks. Twenty-one miles. Twenty percent of the world’s daily oil supply. The disproportion between those two numbers is not a coincidence of geography. It’s the signature of a nonlinear system, where the cost of disrupting a node scales not with the node’s physical size but with everything that flows through it. Iran didn’t need to defeat the United States Navy. It needed to close a strait.

At Prince Sultan Air Base in Saudi Arabia, five US Air Force refueling aircraft were damaged on the ground—this followed the loss or damage of two others in a mid-air collision a few days earlier. The air tasking order, which relies on aerial refueling for long-range strike missions, was compressed. Pro-Iranian militias operating across Iraq claimed 291 attacks on US targets over twelve days, conducting operations from locations that sustained bombing campaigns could not easily reach.

And in the rubble of a school in the port city of Minab in southern Iran, investigators eventually concluded that US targeting databases contained outdated coordinates. The school had once been part of a naval complex, and the records had not been updated to reflect the young children now inside. Here is how the mistake is understandable: The military significance of the strike was zero. The political consequences are catastrophic, scaled in ways no linear model, and no linear military planner could have predicted. That asymmetry is not accidental; it is what nonlinear systems do. China offered $200,000 to the families—a small input with relational dividends that will compound for years. The US offered excuses. Also a small input. Also compounding.

What the Map Couldn’t Show

A complex adaptive system requires diagnosis across multiple simultaneous dimensions—and the gaps in this campaign align precisely with the dimensions the doctrine doesn’t account for.

The United States military operates under Army Doctrine Publication 5-0—the governing framework for how commanders understand their environment and design operations. Its structural flaw, which Operation Epic Fury has made visible at a cost still being tallied, is this: It treats the operational environment as a problem to be understood rather than a system to be diagnosed. A problem has a solution. A complex adaptive system has states, and interventions produce emergent outcomes—emergent states—shaped by the system’s own adaptation. The adversary is not a passive target set. He reads your actions, updates his model, and responds in ways your plan did not specify. This is not an argument for better intelligence, faster targeting, or more precise munitions. It is an argument for different questions, asked before the decision, not after the consequences.

What a genuinely diagnostic framework would have revealed, applied honestly before the first strike, runs across several dimensions.

The pattern of the Iranian system was never “collapse under pressure.” The Islamic Republic has survived the Iran-Iraq war, the Green Revolution, the 2022 Women, Life, Freedom movement, repeated rounds of crippling sanctions, the 2026 protests, and the deaths of thousands of its own citizens at the hands of its security forces. Its attractor basin—the configuration it consistently returns to after disruption—is not fragility. It’s contraction, consolidation, and continuation. The assumption that a sufficiently decisive strike would break that pattern was not supported by the pattern’s history. It was contradicted by it.

The adversary had already evolved past the intelligence model used to plan the campaign. A preliminary but disputed report from the Defense Intelligence Agency assessed that Iran had moved much of its enriched uranium stockpile before the June 2025 strikes, and that those strikes set back its nuclear timeline by months, not years as some claimed. Iran read that assessment too. By February 2026, it had adapted its succession architecture, hardened its proxy network’s operating autonomy, and developed the drone and naval technologies that are now making the Strait of Hormuz one of the most dangerous passages on earth. The campaign plan was based on a 2025 version of a 2026 adversary.

The network was never a hub-and-spoke model. Killing the hub—even the supreme hub—does not disable a distributed network that operates on standing instructions. The proxy architecture across Iraq, Yemen, Lebanon, and the broader region was designed from the beginning to function without direct command from Tehran. Those 291 attacks over twelve days did not require Khamenei’s personal authorization. They only needed the standing orders, encoded across years of relationship-building and pre-positioned munitions, to remain intact. And they did.

The chokepoint has always been the leverage point. The Strait of Hormuz isn’t a military target in the traditional sense. It is the world’s most consequential energy node—the passage through which about one-fifth of the world’s oil supply flows daily. A campaign that doesn’t carefully plan, in advance and with precision, the ripple effects of Iranian actions against that decisive node isn’t truly modeling the entire system. It’s only modeling a part of it—the part that resembles a target list.

The perceptual pipeline was compromised before the first decision was made. The day before the strikes began, Oman’s foreign minister described a diplomatic breakthrough: Iran had reportedly agreed to significant nuclear concessions, and peace was, in his words, “within reach.” The talks were expected to resume within days. The strikes went ahead anyway. The framing surrounding the decision—that diplomacy had been exhausted, that Iran was irredeemably bad-faith, and that military action was the only remaining tool—was not an unbiased intelligence assessment. It was an advocacy stance, amplified by parties with strong interests in the outcome. The confirmation bias toward military action was never challenged by a diagnostic instrument capable of surfacing it. By the time the Oman foreign minister’s statement became public, the bombs had already fallen.

Every system—organizational, institutional, human—operates through a predictive model: the internalized map through which it interprets incoming information. The dangerous property of predictive models is that they cannot detect what they were not built to see. A model calibrated for the issues of Iranian nuclear proliferation, Iranian proxy aggression, and Iranian regime brutality will focus on signals that confirm those perspectives and filter out signals that complicate them. A diplomatic breakthrough the day before the strikes is not a signal that affirms the framing. It is a signal that challenges it. The model did not pay attention to it. It arrived—and it changed nothing.

The question that a genuinely diagnostic framework forces, before the decision is made, is not: “Can we destroy their missiles?”

The question is: “What kind of system is this, and what will it become after we do?”

Answering that question requires viewing Iran not as just a target but as a living system—one with a fifty-year memory, a distributed network that can operate without a single leader, a succession structure built for this exact scenario, a chokepoint it has always known how to exploit, and a perceptual model that has been calibrated over four decades to interpret external military actions as validation of its own founding story.

It requires, in other words, a doctrine that can diagnose complexity, not just plan against it.

THE INSTITUTION THAT WROTE THE MAP

There is a harder question underneath the doctrine critique, and the article demands it be asked: What does the doctrine reveal about the institution that produced it?

The answer is precise, and it’s not flattering.

The US military promotes individuals who master its doctrine. It builds schools, curricula, and promotion boards centered around the model’s categories. By the time a group of senior officers sits down to write ADP 5-0, the pool of writers has already been filtered over decades by the very assumptions the document will formalize. Those who might have viewed things differently—who held mental models shaped by complexity science, ecology, or anthropology—were not the ones who rose to be general officers. The model influenced the group of perceivers before the first word was even written. This is not conspiracy theorizing. It is simply the normal operation of a self-preserving institution: Predictive models don’t just shape perception, they recruit it and turn it to their ends.

But the deeper problem is not the model. It is the identity.

There is a crucial difference between a predictive claim and an identity claim. A predictive claim says this is how the world works. It can be wrong, updated, and survived. An identity claim says this is who we are. It cannot be updated in the same way because evidence that challenges it is experienced not as mere data but as an existential threat. The institution’s response is fundamentally different—and more intense.

We are the world’s most capable warfighting institution, and this is how warfighting works. This isn’t a hypothesis the institution holds; it is the load-bearing wall of everything the institution understands itself to be—its strategic narrative, its internal culture, its relationship with Congress, and its view of history. Therefore, arguing, seriously and with evidence, that the doctrine is fundamentally blind to adaptive warfare is not just an analytical challenge—it threatens the institution’s coherence. The typical response won’t be to examine the evidence; instead, it will be what the framework predicts when identity is threatened: discrediting messengers and redefining evidence. This is the immune system firing. Not because the people involved are dishonest but because that is what systems do when you challenge not what they believe but what they are.

The doctrine cannot be falsified from inside the institution that treats it as sacred. The metrics were created by the model, measuring what the model predicted should matter. The people who would challenge it were filtered out or worn out long before the current conflict began. And each piece of evidence that challenges the identity—a closed strait, a successor elected in just eight days, a war that expanded to nine countries—is reinterpreted within the doctrine’s own framework: we are accomplishing our military objectives.

Again, this is not an indictment of the people in those rooms. They are, by every institutional measure, among the most capable professionals their system has ever produced. That is precisely the problem. The system made them exactly that way. And what they are is optimized for a kind of warfare the battlefield no longer demands, operating within an institution designed to keep that fact invisible.

The results don’t suggest they’re bad at their jobs. They confirm they’re exceptionally good at a job the world has stopped assigning.

The Longer Reckoning

Seventeen days in, as this is written, the military metrics remain favorable in a narrow sense. Iranian drone launch rates have declined by up to eighty-five percent from their opening-day peak. The Iranian Navy is being systematically degraded. Missile production infrastructure is taking significant hits. American airpower remains, as it has been in every conflict for the past thirty years, genuinely unmatched.

None of this is wrong. All of it is insufficient.

The new supreme leader has issued his first statement, vowing to avenge those killed. The Strait of Hormuz remains closed, with Iran permitting ships from China, Russia, Pakistan, India, and Turkey to pass while blocking Western traffic—a wedge driven between the United States and its key great-power competitors. The war has spread across nine countries. The first 100 hours cost approximately $3.7 billion, most of it unbudgeted. As of now, there is no clear way out. The decapitation strike intended to create conditions for regime change from within has instead resulted in a regime with a new supreme leader, a new grievance, a dispersed proxy network conducting hundreds of operations per week, and a narrative of existential external threat that has been the Islamic Republic’s most enduring source of internal unity for nearly fifty years.

A man named Sepehr keeps his door unlocked. He has adapted to conditions his door’s designers did not foresee. In this, he is doing something the Islamic Republic of Iran has always done, and something the doctrine guiding American operations in adaptive wars has never learned—at least not before the bombs fall.

WHAT MUST CHANGE

The lessons here are not technical. They are doctrinal and cognitive—and they are not new. They are simply unlearned. Each of these failures corresponds to a diagnostic instrument the doctrine does not possess.

The first: Iran is not a complicated problem that reached a solvable point in February 2026. It is a complex adaptive system with deep historical grooves, a distributed architecture, and a perceptual model built for exactly the kind of external pressure now being applied to it. Treating it as a machine, with input levers and predictable output states, does not produce defeat. It produces the war that is currently unfolding across nine countries and one very narrow strait. A doctrine that cannot distinguish between complicated and complex will keep producing this outcome, in this region and in others, until the distinction is forced on it by consequences too large to reinterpret.

The second: The planning process that governs American military operations contains no structural mechanism for identifying when the commander’s model of the world has departed from the actual world. It produces assessments, mission analyses, and courses of action. It does not produce diagnostic challenges to the assumptions those products rest on. A planning process that cannot interrogate its own foundations is not adequate to adaptive warfare. It is adequate to the warfare it was designed for—and that warfare is increasingly rare.

The third and fourth together: The adversary you are fighting in month six is not the adversary you mapped in month one. And the question of what comes after each major action—after the supreme leader is killed, after the strait closes, after the proxy network activates at scale—is not a political question to defer until after operations begin. It is the system question. It determines whether the intervention produces a new attractor state or simply a more energized version of the existing one. Both failures share a root: The doctrine treats the operational environment as a fixed object to be understood once, rather than a living system to be diagnosed continuously. The enemy reads the plan. The system co-evolves. The map expires.

The fifth: The day before the bombs fell, a diplomatic signal arrived that the planning process was not built to hear. A diagnostic framework capable of asking, “What would it mean if this were real, and what are we choosing to foreclose?” is not a constraint on action. It is the difference between a decision and a reflex. The model was not built to ask that question. Until it is, the question will keep arriving the day after, year after year, war after war.

A diagnostic framework asks, before the first strike, what kind of system this is, what its attractor basin looked like under pressure, whether the diplomatic signal from Oman represented a bifurcation point worth examining, and what cascading consequences flowed from a closed Strait of Hormuz. It would have modeled succession, not assumed its failure. It would have mapped the proxy network as distributed, not hub-and-spoke. None of those questions guarantees a different decision. But they ensure a better-informed one—and a decision-maker who understands what they were choosing to foreclose.

Epilogue

Somewhere beneath a Tehran apartment building, Sepehr’s family waits for the explosions to stop. He has done what complex adaptive systems do when the environment shifts: He adapted to the conditions that emerged, not the conditions that were planned, maximizing his family’s chances of survival within a landscape someone else designed. He leaves the door unlocked because he understands something that doctrine has not yet fully internalized—that in an adaptive environment, the most important preparation is not the one made for the scenario you envisioned, but the one you make for the scenario that emerges.

The machine is still running. The strikes continue. The missile rates are declining. The Iranian Navy is being degraded. The metrics are mostly favorable, in the way that metrics always are when they measure what they were built to measure.

The system, meanwhile, is doing what systems do. It is finding new attractors—new adaptive configurations. It’s shifting targeting priorities. It’s closing straits and electing supreme leaders and running 291 proxy operations in twelve days from nodes that kinetic campaigns cannot reach. It is producing outcomes that no one in the planning chain fully specified, because no one in the planning chain was equipped to see the system as a whole.

That is not an indictment of the people in those rooms. They are operating with the doctrine they were given. It is an indictment of the doctrine—and a demand, purchased at considerable cost, for something better.

The machine is performing brilliantly. The map is wrong.

---