Survival of the Agreeable

Incompetence arrives with the system that learned to reward comfort, and stupidity has nothing to do with it.

Norman Dixon wrote about this in 1976, and the book still hasn’t been forgiven. A British psychologist, a sapper before he turned to science, and so a man who knew firsthand the price of someone else’s miscalculation. He took one hundred and twenty years of mostly British disasters and asked the question that still makes historians flinch. What if these failures aren’t glitches? What if they’re the predictable product of how armies select, promote, and shape people?

Then he did what nobody expected. Instead of hunting for guilty generals, he drew their collective portrait, and the portrait turned out not to be about stupidity. Dixon’s failed commander is almost always conscientious, disciplined, often brave. Something else sinks him. Knowledge is dangerous for this type: it might reveal he was wrong, and his ego can’t bear that, so new information gets rejected before it even arrives. Hence the anti-intellectualism that armies spent decades mistaking for reliability. Hence the cult of what the British call “bull”: polished boots, perfect drill, form valued over substance. When fear lives inside, shine on the outside soothes. Beneath the rigid discipline, Dixon saw a deep fear of failure, armored in certainty, and an almost physical discomfort with ambiguity. Such a commander would sooner give a hopeless order than admit he doesn’t know the answer.

And this is where the book stops being about the past. These traits aren’t accidental. They’re cultivated. Private schools and military academies spent centuries selecting for obedience, endurance of humiliation, and contempt for “clever types.” The output: a person perfectly adapted to peacetime. And lethally dangerous in a crisis. Because the virtues that lifted him up turn into liabilities in combat. Obedience, predictability, unquestioning compliance. The very things that save you on the parade ground, in battle make a commander deaf to the news that the plan is falling apart.

Balaclava, 1854. The Light Brigade was hurled straight at the batteries. An order everyone saw as madness, but questioning it upward was more terrifying than carrying it out. Six hundred horsemen rode headlong into the guns. They saw the guns. Because the system had already taught them: an uncomfortable question costs more than a life. Reason here serves as the lawyer for a decision already made.

I read it and recognized it.

Military organization has a rare talent. It selects for the top those who know how to survive inside themselves, bypassing the ones who win battles. These are different skills. Sometimes opposites.

Let’s not confuse this with the market, though. In the market, the difficult ones actually win: Judge and his colleagues showed in 2012 that agreeable people earn less, because the tough ones negotiate better for themselves. Hierarchy works in reverse. It has no love for those who negotiate, argue, drag inconvenient numbers into the meeting. It loves those it never has to figure out each time. Kindness has nothing to do with it; it’s about predictability. The boss sleeps easier with someone who’ll never surprise him, and that’s exactly who he pulls up behind him, floor by floor, until the top fills up with an entire story of people whose core competence is not causing the chief discomfort.

But in war, the environment shifts. And those who spent their lives learning to read the mood of the office suddenly find themselves facing a reality indifferent to seniority. It doesn’t file reports. It doesn’t wait to be approved.

The Soviet tradition had a precise word for the first stage of the disease: tufta. A report written to match what the commander wants to hear. Inflated tonnages, nonexistent kilometers, divisions destroyed on paper. At first glance, petty fraud. In truth, tufta kills, and history left the minutes.

Mao’s Great Leap Forward ran on it literally. Communes competed to report the biggest harvest, numbers were inflated two- or threefold above reality, and the state dutifully collected its grain tax based on paper figures. The peasants were left with what didn’t exist on paper. Thirty to forty-five million dead of starvation amid reports of record harvests. No one ordered people starved. It’s just that for years only good news traveled upward, and at a certain altitude the leadership was already looking not at the country but at its own fantasy of it.

1986 showed the same thing above the reactor. In the first days after Chernobyl, official radiation levels were understated by a factor of a hundred, and the local hierarchy spent several days not daring to call the accident an accident. In a system where bad news gets you fired, no one wants to be the one who brings it. And the worst of it began before the explosion: operators who had doubts about the dangerous test had nowhere to put those doubts. The channel for “stop, this is risky” simply didn’t exist.

Tufta breeds wherever the punishment for bad news exceeds the punishment for a bad decision. Then the slow slide begins. The information environment detaches from reality report by report, a slow slide until the map on the staff wall describes the leadership’s mood and the terrain has fallen out of it.

2022 showed this without embellishment. Russian command entered Ukraine with a map of its own illusions. Intelligence was the least of its shortages; there was an ocean of it. But for decades a system had been built where the loyalty filter stood higher than the accuracy filter. The plan, later obtained by Britain’s RUSI, was known only to a narrow circle; it allotted two to three days for the capital, the same number named independently by both the CIA and the American Joint Chiefs, because everyone was reading the same Russian confidence. FSB assessments were packaged to the Kremlin’s taste, sources had a financial interest in confirming the desired narrative, and the promised “flowers from the population” lived only in reports. The analyst who wrote the truth risked his career: the head of the relevant FSB directorate was the first arrested after the failure. The analyst who confirmed the ready-made conclusion went up for promotion. Researchers called this the autocrat’s intelligence paradox. The tighter the power, the worse it sees, because everyone beneath it knows that the way up is a mirror of the boss’s beliefs. Enough years of this selection, and the system simply stops seeing.

It’s easy to think the whole disease comes down to fear of the boss. Here the theory takes an unpleasant turn. The loudest fiascos are often made by cohesive teams of smart people who fear no one; cowed subordinates of a despot have nothing to do with it. In 1972, Yale psychologist Irving Janis dissected several American catastrophes, among them Pearl Harbor, the Bay of Pigs, and the escalation in Vietnam, and found a common mechanism. He called it groupthink. When a team is tight-knit and loyal, the drive to preserve agreement quietly defeats the desire to see the truth. An illusion of unanimity appears: everyone is silent, and each reads that silence as consent. Self-censorship appears: the person with a doubt decides not to spoil the mood, thinking “I must be missing something.” What Janis called mindguards appear: well-meaning souls who filter out inconvenient information before it reaches the chief, so as not to trouble him. No one is lying. No one is afraid. And the group marches toward catastrophe, unanimous and uplifted. That’s why the most dangerous meeting is the one where everyone agrees immediately. Unanimity is rarely a strength. More often, it’s a symptom.

Israel in 2006 added a third shade after fear and herd instinct: hubris. For over a decade, from 1994, the Israeli army cultivated an elegant doctrine, the Systemic Operational Design of General Shimon Naveh and his OTRI institute. It was meant to revolutionize command through systems theory. Instead, it paralyzed it. Battalion commanders received orders sprinkled with postmodernist terms: rhizome, cognitive mediation, deconstruction of the enemy system. Naveh later admitted those texts weren’t meant for ordinary mortals. While staffs admired the depth of their own thought, the infantry stood and waited for a translation. After the war, the institute was disbanded.

Three failures. One nature. The first detached from reality out of fear, the second out of arrogance, the third just to avoid spoiling the shared mood. The bill is paid the same for all of them, and always by those at the bottom.

This has a name, and a confusion of authorship worth clearing up. The psychology of the individual commander was described by Dixon. And how an entire failure machine gathers around such a leader was described thirty years later by Padilla, Hogan, and Kaiser. The toxic triangle. Three vertices, and none sufficient on its own. A destructive leader with a recognizable set: charisma, personal lust for power, narcissism, a dark life history, an ideology that divides the world into us and enemies. Two kinds of followers: conformers obey to belong and keep their place; colluders actively support because they share in the leader’s spoils. And an environment without counterbalance, where oversight is weak, safeguards are absent, there is no one and no way to say “no.” Remove any one vertex, and the triangle won’t assemble; you’re left with just an unpleasant boss the system could survive. The point isn’t that destructive leaders are born somewhere. The point is how much ready soil always gathers around them.

And this soil is silted up by years of small decisions. Every time someone stayed silent because it was dangerous. Every time an inconvenient person was passed over for promotion. Every time a report was smoothed down before being sent upward. None of these decisions is a crime on its own; each seems reasonable, prudent, appropriate right now. That’s how it accumulates, in quiet concessions, none of them quite rising to betrayal.

Incompetence settles.

I’ve written about this separately, from another angle: a selection system inherits the maturity of those who own it and outlives its creators, carrying forward what was loaded into it, whether a standard or a poison. Tufta is that poison, poured into the structure drop by drop. Inertia gains mass imperceptibly, until it starts to look like the nature of things, like “that’s just how we do things here.”

The most dangerous moment arrives without thunder. The commander is still making decisions, still certain he sees the field, and the channel through which truth used to reach him is already dead. The system around him was long ago trained to say what’s expected of it. One bad order on a fresh head can be corrected by the next. A dead channel can’t be corrected by anything: from then on everything moves blind, and the darkness has already been named clarity.

And here arises the only question that carries any weight at all. Can this be cured? It can, but not with what people usually try. Not with a new regulation, not with a reform, not with a slogan about initiative.

In 1999, Amy Edmondson, a Harvard researcher, was studying errors in hospital teams. She expected something simple: better teams make fewer errors. The data gave her the opposite. Better teams reported more errors. They reported more errors because it was safe for them to talk about mistakes. In bad teams, a nurse who admitted an error risked her reputation, so errors were buried, and the system didn’t see them until someone died. In good ones, admitting was safe, so a small slip was hauled into the light while it was still cheap. Edmondson called this condition psychological safety: the shared sense that you won’t be punished for inconvenient truth. Softness and comfort have nothing to do with it. It’s the only known way to make a system see itself.

From this an uncomfortable yardstick follows. An organization is healthy not because its reports contain no bad news. The quietest reports often belong to the sickest. The healthy one is where bad news reaches the top fast and costs nothing to the one who brought it.

And it is achieved by one thing. Structure and motivational posters are powerless here. One thing: how the commander reacts when he is told something he did not want to hear. Whether the person who brought the bad news leaves the office feeling he did what was necessary, or already regretting he opened his mouth. From this single gesture, subordinates read the real rule of the game faster than from any order, and within half a year, either a living nervous system grows around the commander, or a closed loop of tufta.

The fifth year of a major war makes the arithmetic literal. On the other side of the front, the system that spent decades castrating its own ability to see is now on display for the whole world. Laughing at someone else’s tufta is easy. It is harder to spot its embryo in your own report, right where the impulse to round up in the boss’s favor kicks in. The price of our rounded figure is paid in lives, and a reprimand is the cheap coin we weigh it against.

Incompetence settles in silence. Those who stay alive are the ones who managed not to let that silence take hold.