Fading advantage of a new weapon

Why does a weapon that wins decisively on its first use rarely win that way twice?

A new weapon appears and something that had been impossible becomes routine. Then, within a season or two, the same weapon is producing ordinary results against the same enemy — and nothing about the weapon has changed. If it worked because of what it is, that should not happen. So perhaps it did not work because of what it is.

Consider what the first use is measured against. The defender on that day has arranged everything — his positions, his formations, his training, his procurement — around the threat he already knew. The new weapon's effect is the size of the gap between his arrangements and this new thing. That is a statement about the mismatch, not about the weapon. Which means the advantage was never a property of the device; it was a property of the pairing, and the enemy owns one half of the pair.

Now notice a second, sharper problem. To have any effect at all, the weapon must interact with the enemy — and interaction is disclosure. He gets the wreckage, the fragments, the recording of whatever signal it used, the account of what it did and from where. A weapon in a shed teaches nobody anything and achieves nothing; a weapon employed achieves something and teaches. You cannot buy the effect without paying the tuition, and the two are the same transaction.

Ask what the enemy does with what he has learned, and here is the asymmetry that sets the decay rate. Almost any weapon works by completing a chain — it must be delivered, it must find the target, it must track it, it must arrive with its effect intact. Breaking one link is enough to spoil the whole, and a link can usually be broken far more cheaply than the chain was built. A decoy, a smoke or chaff cloud, a jammer, a different route, a change of spacing: each is small compared with the programme that produced the weapon.

And the cheapest counter is not equipment at all. It is behaviour — disperse, move at night, keep off the roads, stop transmitting, dig deeper. That response costs nothing to procure and propagates across an entire force by instruction in days, while the innovator's advantage arrives in small numbers and grows only as fast as a factory. So the two sides adapt at wildly different rates, and the faster adaptation is the free one. Much of the decline in a weapon's effect is usually spent before any counter-device exists.

That in turn explains a decision that looks strange from outside: the argument, before every such introduction, about whether to hold the weapon back until enough exist to be decisive, or to use the few available now. Both options are bad and the choice is genuinely contested at the time, because the first use is the only cheap one and it is spent whenever you spend it. The same reasoning drives the effort to make a use look like something else, so the enemy attributes the loss to bad luck or an old cause rather than a new capability.

If all this is right, the decay should not be uniform. It should be fastest where a use reveals the mechanism plainly and the counter is cheap, and slowest where the advantage rests on something the enemy cannot cheaply replicate or spoil: an industrial base, a geographic position, a trained cadre, an accumulated body of practice. That is the test. What would refute it: weapons whose principle was fully public from the start decaying just as fast as secret ones — which would mean disclosure was not doing the work — or counters that routinely cost more than the weapons they defeated and were adopted anyway.

One honesty flag. Attributing a change in results to a particular countermeasure is retrospective, and the parties are interested: the service that owned the weapon has reasons to say the enemy adapted, and the one claiming the counter has reasons to say it worked. Contemporary records rarely settle it as cleanly as later accounts imply.

The claim is not that new weapons do not matter. They plainly do, and some reset a whole style of fighting. The claim is narrower: the decisive portion of the effect — the part that comes from meeting an opponent arranged for something else — is a one-off, and it is consumed by the same act that produces it.

Two refinements connect the steps. First, "the counter" is rarely a single thing; it is a package of small, boring changes, which is why the decline looks gradual and is hard to attribute. Second, what remains after the decay is the weapon's real value — the margin the enemy cannot reach with cheap measures. Judging a system by its first employment therefore overstates it, and judging it by the eventual settled margin is the honest test, though that number is only available years later, long after the procurement decision has been made.

• Why the choice between early use and holding back for mass has no good answer, and how it has been argued.
• How a side conceals that a loss came from a new capability, so the adversary looks for the wrong cause.

Every term the collection defines is gathered in the glossary.