Airborne drop dispersion

Why did badly scattered parachute drops sometimes confuse the defender more than tidy ones?

A scattered drop is a disaster by every measure its planners care about. Sticks land miles from their zones, companies arrive as knots of strangers, the containers holding the radios and mortars come down elsewhere, and objectives go unattacked because nobody meant to attack them is there.

And yet the claim keeps recurring in accounts of these operations: the mess helped, because it left the defender unable to work out what was happening. That is a suspicious sort of claim — exactly the shape of a consolation invented afterwards by people whose plan failed. So the honest question is not whether scatter helped, but whether there is a real mechanism by which it could, and what would have to be true for it to bite.

Shift the point of view. Stop asking what the paratroopers can do, and ask what problem the defender is solving that night.

He is not trying to kill paratroopers. He is trying to answer one question in time to matter: where is the main effort, and when do I release my reserve against it? His reserve is finite and slow, and committing it is close to irreversible.

Now ask what he has to answer it with. Reports — telephoned from scattered posts, each saying some version of "parachutists, near here, unknown strength, some time ago". His picture is assembled from those, and reports are what a drop actually delivers to him.

So compare the two drops as signals rather than as tactics.

A tight drop produces few reports, mutually consistent, all pointing at one small area. That is a clean signal. He localises it quickly, concludes that is the effort, and commits — which is precisely what the attacker was hoping to avoid.

A scattered drop produces many reports from widely separated places, each individually weak. And here is the inferential trap: he cannot distinguish "one force in the middle plus a great deal of noise" from "several separate landings across a wide frontage". Both hypotheses fit his data. The estimated frontage inflates with the spread, and the estimated strength inflates with the number of reports, because a report is a report whether it came from a battalion or from four lost men.

That leaves him only bad options. Commit the reserve to one place and he is probably wrong. Split it and he is weak everywhere. Or wait for a clearer picture, which is free of the first two errors and costs the one thing the attacker most wants to take from him: hours.

There is a second mechanism, mechanical rather than psychological, which does not depend on anyone being fooled. Small parties spread across a wide area cut telephone lines, ambush couriers and shoot up staff cars over that whole area. The defender's reporting system — the very thing the first mechanism attacks the interpretation of — is being physically degraded at the same time. He is not merely confused; he is going deaf.

Is there evidence the mechanism is real rather than retrospective? One piece is unusually clean: dummy parachutists, dropped deliberately away from the real zones, together with small parties whose only task was noise. Nobody spends aircraft on decoys unless the effect being bought — inflated apparent frontage, delayed reserve release — is one they believe in beforehand. It is the deliberate version of what scatter produces by accident.

Here is the test. If the mechanism is inflated estimated frontage forcing a reserve decision under ambiguity, then scatter should help only where the defender's reserves are centrally held, where he has too little force to cover everything, and where his doctrine is to commit on the density of reports. Against a defender thick enough to picket the whole area, or one whose rule is to wait for confirmation and then strike the concentration, it should buy nothing and cost everything. The refuting observation would be a defender who held back, identified the real objective anyway, and destroyed the concentrated element — the mechanism running in reverse.

The correction to make, firmly, is that none of this makes scatter good. No one has ever planned a dispersed drop, every airborne force spends enormous effort on navigation, pathfinders and drop-zone marking, and the operations remembered as costly failures were costly largely because of dispersion. The claim available here is much narrower: given that the drop went wrong, the defender's problem got harder in a specific and describable way, and that partly offset a loss that was real.

A refinement that joins the steps: the two mechanisms have different time signatures. Inferential confusion is worth hours — it delays a reserve. The physical degradation of the defender's communications lasts as long as the parties survive. An operation needing a bridge held for a day gains from both; one needing an objective taken in the first hour gains from neither, because the men who were to take it are ten miles away.

The honest limits are the important part of this piece. The evidence is almost entirely narrative, and both sides supplying it are interested parties: the accounts come from official histories written by the services whose operations are being judged, and from the defeated side's post-war interrogations and memoirs, where "we were confused by an unprecedented dispersion" is a more comfortable sentence than several alternatives. There is also a plain selection effect — scattered drops that preceded successful campaigns are recounted at length, and the argument is rarely tested against the ones that preceded failures. So treat the mechanism as plausible and structurally sound, and the claim that it decided any particular night as contested.

• How the same inference problem appears wherever a defender must commit a scarce reserve on ambiguous reports, as in Strategic deception.

Every term the collection defines is gathered in the glossary.