Shoot-and-scoot artillery

Why do modern gun crews displace after a handful of rounds when older batteries stayed put for hours?

For most of the history of gunpowder, a battery dug in and stayed. It surveyed its position, registered its targets, built up ammunition, and fired from that spot for as long as the fight lasted — sometimes weeks. Today a crew fires a few rounds and runs, and everything about the gun is designed around running.

The gun did not get more fragile. Its range and accuracy improved. So what changed, if it was not the gun?

Begin with an uncomfortable property of artillery: firing is self-revealing. A gun announces itself with sound, muzzle flash, smoke and disturbed dust — signatures armies have hunted for a century. But one signature cannot be suppressed at all, because it is the weapon: the projectile.

Think about what a shell in flight is, physically. Ignoring air resistance, it follows a ballistic arc determined by where it started, how fast it left, and at what angle — few enough parameters that observing a short segment of the arc is sufficient to reconstruct the rest, including the part you did not see, running backwards to the muzzle. Drag and wind make the real computation messier without changing the point: the trajectory contains the firing position, and anything that can watch a piece of it can recover that.

So the question becomes how quickly, and how accurately, the other side can read it. And here is where the history sits.

The older techniques all worked, but indirectly. Flash-spotting had observers on separated posts cross bearings on a muzzle flash. Sound-ranging used microphones along a known baseline and the differences in arrival time of the report. Aircraft looked; crater analysis worked backwards from the impact. Each needed several separate observations, brought together by people, passed by voice or wire, and turned into a grid reference — and each carried error a commander had to allow for. The loop from first round to something worth shooting at was slow.

Radar collapsed that loop, doing in one place and automatically, while the shell is still climbing, what a chain of observers and telephones used to do. I will not quote detection ranges or reaction times: published values are estimates and vary enormously by system, terrain and clutter. What matters is the structure, not the constants.

Now write the survival condition, because once you see it the tactic is forced. The battery lives if it is gone before the answering rounds arrive. Two clocks run against each other:

• the battery's clock — rounds fired, then the time to hook up, move, and be ready again elsewhere;
• the opponent's clock — locate, pass the firing point as a target, lay the guns, and the shell's own flight time.

Ask yourself what happens as the second clock shortens. At first, nothing dramatic: you fire a little less at each position. But displacement takes a roughly fixed amount of time — you cannot pack up faster than the vehicles and the crew allow. So there comes a point where the opponent's clock is shorter than yours no matter how few rounds you fire, and the position stops being defensible in principle rather than in degree. That is a threshold, not a slope, and it is why the change in behaviour was abrupt rather than gradual.

It also explains what modern gun design has spent its effort on, which is not firepower. Self-propulsion, so displacement does not need a tractor. Onboard navigation and automatic laying, so a gun can arrive at an unsurveyed patch of ground and fire without the survey and registration that once took an occupation's worth of time. Burst-fire mechanisms, so the few rounds you can afford arrive in a useful concentration. Every one is an attack on the battery's own clock.

Is this testable? Sharply. If the driver is the opponent's find-to-strike latency, displacement discipline should track the enemy's sensors rather than the gun's own technology: the same modern battery should be willing to stay put against an opponent with no counter-battery radar, and armies fighting such opponents should drift back to long occupations. The refuting observation would be crews displacing just as fast against an enemy demonstrably unable to locate them, or displacement rates following gun generations rather than enemy capability — which would mean doctrine or habit is doing the work.

A comparison is useful, because this looks superficially like the story of walls that stopped working when gunpowder arrived. It is not the same shape. There, the defence's own virtues — tall and thin — were turned against it by a new kind of load. Here nothing about the gun or its position changed. What changed was the latency of somebody else's loop: a position survivable at one latency is untenable at another. The vulnerability was always present; only the speed of exploiting it moved.

Two honest qualifications. First, shoot-and-scoot is a trade, not an improvement: fewer rounds per occupation means a lower sustained rate of fire, far more movement and wear, and guns scattered rather than concentrated — a real loss, since artillery's effect has always come from mass. Digital fire control partly buys that back, letting physically dispersed guns converge their rounds, so dispersion no longer means dilution.

Second, the radar is not free either. It must radiate to see, and radiating is itself a signature that can be located and attacked, so counter-battery radars are used in short bursts and moved — the same logic one level up. Each side's sensor lives under the constraint it imposes on the other.

• Why counter-battery radars are switched on so sparingly, and what that costs the side that owns them.
• How the same find-to-strike latency argument now applies to headquarters, logistics dumps and radio transmitters.

Every term the collection defines is gathered in the glossary.