Miller:
Well, this question of why
you needed so many weapons is a good one. And I certainly think that the Navy people, and a lot
of people were concerned about it. It wasn't just the Navy alone. The Army tactical air command
people, a lot of people within the Strategic Air Command itself were always somewhat concerned
about the numbers of weapons. The first SIOP had about 3,500 weapons in it, which we thought was a
lot of weapons. But if you use the process that was developed, and it was a SAC process and I
think it was a good process. It was a mathematical process. If you used that process in
computing the probability of doing the damage, you could arrive at the number of weapons that
were required and you could build it up. The, as I indicated earlier, the thing that determines
really the numbers of weapons to do the job, is the guidance, what kind of targets do you want
to attack. And what level of damage do you want done. Then comes into play, the factors of
capabilities of the weapons and the delivery systems. Now, maybe I can lay it out for you here,
simply on a piece of paper. See if I can describe this. And this was a SAC process, which was a
good one. They came up with a simple mathematic term called "damage expectancy", the damage that
we expected to do. The damage that we expected to do with a particular weapon. D.E. We always
used to talk about the D.E. What is the D.E. of this particular weapon? Alright, to compute, the
damage expectancy is a function of two things. First, it would be the probability of the weapon
arriving at the target, the probability of it arriving at the target. And the second would be
the probability of the weapon doing damage after it got to the target. Now the probability of
arrival is a function way back. Let's say, the target is over here. The probability of arrival
starts way back here at the launch platform. What is the probability of the launch platform
surviving if it's under attack? If there's, you've got a lot of strategic warning. If you
preempt, for example, the survivability is very good. It can be 1.0. But if you had to
sustain this under, had to sustain a bit attack, it probability might be much lower. Now this
where we started to get different factors. For example, if you were an open target like SNARC, or
something like that. A cruise missile in Presque Isle in Maine, and you had to undergo an attack on that, or it
was an Atlas missile out in Vandenberg Air Force Base sitting out in the sunshine of southern
California, the probability under an attack of it's surviving would be pretty low, so it's
probability of arrival would be rather low and therefore, it's damage expectancy would be very
low. If it was a weapons in a submarine, submerged below the ocean, it's probability of survival
was pretty high, because anti-submarine warfare capability at that stage was pretty low and it
still is to this day. The survivability factor for that submarine that's carrying those weapons
is still 1.0. It's 100 percent. So, you start off with a pretty high factor
right there. Now then, the next factor that's involved is the probability of the weapon leaving
the launcher when you push the button. They don't always work. And this was one of the great
things that McNamara did when he became Secretary of Defense. Statistically oriented. He had Alain
Enthoven as OK. Had one of his guys. And so he went for more testing to get us better
reliability factors. But the probability of the thing leaving the launch pad could be 95
percent or something less. Some of them it was down as low as 30 percent. And so you have that
in there. Then there's the probability of it as it goes to the target of it malfunctioning.
Whether it was a bomber aircraft or anything else. A certain number, every so many miles, a
certain percentage of them are going to fail. They're not going to go. And then there is, when
you got near the target area, where there was the penetration capability. If you got to go by a
big bomber base here, a big fighter base rather, and this is a bomber, the fighters are going to
come up, the probability of knocking it down might be pretty high. So, the first thing you've
got to do is take out this base. You take out a military target in order to get through to your
urban industrial target. Then finally, you arrive at the target itself, you've gotten it there,
you're probability of arrival now say is 75 percent or something of that nature, cleared out all
the way. Now, the weapon goes off and it's probability of doing the damage is a function of how
close does it land to the target point, the aim point. And will it go off. About five
percent of them probably wouldn't go off. So right away you've lost that. And then if the weapon
was not very accurate, its probability of doing the damage would be degraded by its lack of
accuracy. Now then, if you're given a high, a high, the guidance says you'll get a 90
percent confidence level that you're going to do 95 percent damage to this target. And
you start off over here, with a weapon that's only got a 50 percent probability of surviving
under attack, and by the time it gets over here it's going to be less than that, you're a long
way from meeting this factor. And so you start adding weapons. And you keep adding weapons and
adding weapons to get to this damage level that's set up here, and you keep adding them on and
adding them and pretty soon you get asymptotic on the curve. And if it's a very hard target, like an
ICBM, and you don't have great accuracy, you can keep adding weapons and adding weapons and
you're never going to get there. And this is why Schlesinger and all of us had said for years,
you can't get a counterforce capability unless you have great accuracy. And if you cannot attack
the submarines and a lot of the weapons are in submarines, if you can't even target them, how
are you going to get a counterforce capability? But this is how you got to so many weapons. You
had a high, a damage expectancy that was dictated to be very high. That meant a lot of weapons
had to go on a single aim point in order to guarantee the probabilities that were specified in
the guidance. You want to reduce the arsenal? You really want to do an arms reduction problem?
Change the guidance. Say just, all the guidance says is put a weapon on a target, and the number
of weapons would go down.