From a post to ALT.BOOKS.DAVID-WEBER dated November 2, 1998:

Counter-missile canisters

    First, there are such things as CM canisters. They are, however, seen primarily as a means of replacing standard CM launchers which have been lost to enemy fire.

    The problem is tracking and guidance. A counter missile is a smallish missile with a hellacious wedge. There are two reasons for this (as I think I have said before): first, to get it out to interception range as quickly as possible, second, because its wedge is its primary missile-killing system, a "sweeper" that has to make direct contact with the wedge of its target. This second point requires the most "reach" or width that the designers can figure out how to engineer into the system

    When these factors are combined, you get a few problems.

    Most importantly, perhaps, the sheer size of the drives required to produce the needed performance preclude packing the bird with lots and lots of really sensitive and smart homing systems. Obviously, the CM has at least some homing capability, but it is inherently short-ranged. Think of it as myopic, if you will. In order to get any really useful PK numbers, each CM requires external guidance from something which has much better sensors and much better ability to penetrate defensive ECM--i.e., the fire control systems of the ship which launched it. This limits the number of intercepts to the number the launching vessel has "channels" to control. Once the CM reaches a range at which its shorter-ranged, more simple-minded onboard systems have secured a high confidence lock on its target, it hands itself off to onboard control, releasing the fire control link to the main ship (and, incidentally, giving missile defense a much better handle on threat assessment, since the CM won't hand itself off unless the PK has reached certain parameters stored in its onboard computers, and in the RMN, those parameters basically specify about an 80% PK). Until and unless a CM disengages itself, the launching ship either controls it clear to target or decides when it will disengage, leaving the CM to its own devices, in order to pick up and control a fresh CM.

    Additionally, the width and power of the wedges themselves present some fairly severe fire control problems, since each of them creates a "hole" in the sensor coverage of anyone behind them. The same thing is true of both incoming and ouotgoing shipkillers, as well, and I suppose I should have made a bigger point out of the fact that there are, in many ways, "blinding clouds of gunsmoke" (i.e., impeller wedges) cluttering hell out of the sensor environment once the missiles begin to fly. One function of a wall of battle (and its escorts) is to spread the available sensor platforms as widely as possible in order to see around the holes, and a major function of CIC is to take input from every available platform and combine it to create/extrapolate the plots from which the tac officers and flag officers work. But in addition to the "blinding" effect of the wedges, they also cut holes in the available communication linkages. If a missile wedge wanders between a warship and the missiles (SK or CM) it is currently controlling, it cuts the link (the equivalent of "cutting the wire" on a wire-guided wet-navy torpedo), with what is usually a very severe degradation in the accuracy of the missiles involved.

    Because of this, the number of CMs a ship can control at any given moment is limited by both the number of control channels available to it and the clear communication paths between it and the missiles and the quality of sensor data available to it from all allied sensor platforms. The result is a system which gives very high probabilities of kills on low numbers of missiles, does fairly well against "normal" missile-heavy environments, but is subject to swamping by enormous salvos... which is what just happens to make the pod format so effective with the new light-weight launchers and (especially) the new longer-ranged missiles.

    The function of a CM canister is not to throw out hordes of CMs (which couldn't be effectively guided, anyway), but rather to allow a ship to put out all the CMs it can handle at any given moment, regardless of what has happened to its dedicated CM launchers. If, for example, you've taken a hit that destroyed three CM launchers but the guidance channels are still intact, you simply seed your next salvo of shipkillers with a CM canister, only three of whose missiles (I am assuming a 5-missile canister from a larger unit) actually spin up and go after incoming threats.

    Unless something changes in a major way where missile defense is concerned, things will probably stay pretty much this way with simple incremental increases in CM effectiveness rather than wholesale enhancements.