Unit combat learning curve
A Socratic walk-through of the unit combat learning curve — reasoned out one step at a time, not lectured.
The question we started with
THE QUESTION #Why is a unit's first battle far deadlier for it than its fifth?
Two battalions go into the same sector, against the same enemy, with the same equipment and the same doctrine. One has been fighting for six months; the other arrived last week. The new one will very likely take heavier losses — often strikingly heavier — for its first few engagements.
The obvious explanation is training: the veterans are simply better trained. But that is the part that does not hold up. Peacetime armies train hard, and the newly arrived battalion has usually done more recent, more systematic training than the tired unit it is relieving. So what is the veteran unit carrying that no training schedule managed to install?
Reasoning it through
REASONING #Start with something small and concrete. A shell lands somewhere nearby. What does a soldier need to know?
Not the doctrine — everyone was taught the doctrine. What is needed is a judgement made in under a second: was that outgoing or incoming, near or far, ranging or on target, and therefore do I go flat now or keep moving? Ask yourself where that judgement could possibly come from. It is a calibration of the senses against a specific world of sounds and sights, and no classroom supplies it. The new soldier either goes flat too often, and stalls, or too rarely, and is hit.
Multiply that by everything else: which piece of cover is actually cover here, how long the enemy's mortar crews take to adjust, what a treeline looks like when it is occupied, how much noise a movement can make before it draws fire. Notice that all of these are local and specific. They are not general skills. That is the first mechanism, and it explains why training cannot fully substitute — the knowledge is about this enemy, this ground, this season.
Now step up from the individual to the unit, because the question was about a unit. What does a company know that no individual in it knows?
Consider a platoon that must move while another suppresses. Who fires first, how long they keep firing, when the moving element signals that it is set — in a veteran unit these are settled without conversation. The sergeant knows which of his people freeze and which get reckless. The commander knows what he can ask of each subordinate and, more importantly, what he cannot. Radio traffic drops, because the routine no longer needs to be spoken. This is the second mechanism: a unit's real procedures are a shared, mostly unwritten model of one another, and it can only be built by acting together under real stress.
There is a third, less comfortable one. Some of what looks like learning is selection. The soldiers and the officers most likely to make a fatal error are disproportionately the ones removed by that error, so a unit's average competence rises partly by subtraction rather than by improvement. Here the honest answer is that the two are very hard to separate in the historical data, and analysts still argue about the mix. The pattern itself is well attested — in air combat, for instance, loss rates were markedly highest in a crew's or pilot's earliest sorties and fell as sorties accumulated — but a raw curve of losses against experience cannot by itself tell you how much is learning and how much is survivorship.
One more observation that helps separate them. When armies fed individual replacements into veteran units — as the United States largely did in the Second World War, and was criticised for afterwards — those replacements suffered badly in their own first days, inside units that were already experienced. What does that tell us? That the unit's collective knowledge does not simply transfer to a newcomer by proximity; a good deal of the curve is individual and has to be paid for again by each person.
And does the curve keep falling forever? No — it flattens, and it can reverse. Fatigue, cumulative psychiatric casualties, and the slow replacement of the original members all erode it, which is why long-committed units eventually decline even as their nominal experience grows.
The analogy
THE ANALOGY #Think of a hospital's night shift in a building that has just opened. Everyone is qualified; everyone has passed the same examinations. But nobody yet knows which lift is slow, which alarm is usually spurious, whose judgement to trust at three in the morning, or how long the pharmacy really takes. Their first weeks are error-prone in ways that have nothing to do with medicine. Six months later they are visibly better, and yet not one of them has learned any new medicine — what changed was their knowledge of this building and of each other.
the night shift learns by making mistakes it usually survives, whereas a unit's first battle removes many of the people who would have carried the lesson forward — so the military curve is bought at a price that has no counterpart in the hospital, and part of its apparent slope is that subtraction rather than any lesson learned.
Clarifying the model
THE MODEL #Two things worth correcting. First, "experience" here is not a quality of persons but largely a property of a particular group in a particular setting: move a veteran battalion to a different theatre and part of the curve resets, because the specific calibrations no longer apply.
Second, the curve is not a reason to regard training as useless. Training determines where the curve starts and how steeply it can fall — realistic force-on-force exercises, and rotating whole units rather than individuals, both measurably shift it. What training cannot do is deliver the parts that are specific to an enemy who has not yet been met.
A picture of it
THE PICTURE #How to readtrace the states in order: a unit begins trained but untested, pays its heaviest bill on first contact, then improves as local knowledge and unspoken coordination settle. The back edge is the important part — fatigue and the arrival of untested replacements return part of the unit to the costly state, so the curve is not a one-way trip.
What became clearer
WHAT CLEARED #The first battle is deadlier not because the unit knows less about war but because it knows nothing about this war — the sounds, the ground, the enemy's habits, and the reflexes of the person next to you. That knowledge is local, largely unwritten, and held collectively, so it cannot be issued in advance; it can only be bought, and part of the price is that some of it is bought by subtraction rather than by learning.
Where to go next
ONWARD #- How much of the observed curve is learning and how much is survivor selection, and what data could separate them.
- Unit rotation against individual replacement, and what each does to the curve.
- Why the curve flattens and then reverses in long-committed formations.
Key terms
TERMS #| Term | What it means |
|---|---|
| Tacit coordination | the unspoken, learned expectations that let a group act together without instruction. |
| Survivor selection | the rise in a group's average competence caused by the removal of its least effective members rather than by anyone improving. |
| Individual replacement system | feeding single soldiers into existing units rather than rotating whole formations; each newcomer then pays the early part of the curve alone. |
Every term the collection defines is gathered in the glossary.