It is a philosophical stance, because things can be relative without giving units, and often relative is much easier to understand. "How far is it to John's house?" "About half way to Seattle." Okay, so maybe the units are "distance to seattle" and the units to John's house are 0.5. Also, for anyone who's lived in Western Washington, we tell distance using time. "How far is it to Seattle?" I have no idea how many miles, but I can tell you it's 1-1.5 hours depending on traffic. Here, we are not arbitrarily assigning the .45 a designation. We are comparing it to the 9. If we had the number of units, we'd still compare it to the 9 to see which of the two is more effective. But we can make a comparison without units.
Predictions mean nothing when we are considering independent events. No one shooting is has anything to with another. You can guess what the weather is based upon simple observations of atmospheric moisture in the evening (Example: High thin cirrus clouds indicate minimal moisture in the atmosphere and a lower chance for rain in the next 24 hours) because atmospheric systems have a set of characteristics that preceed/accompany them.
There are no such precursors for shootings.
You are right in that weather is a continuum, and not a series of events. However, most of the factors associated with shootings are independent of caliber selection. Those which are based on caliber selection, you can see some of the factors in a lab setting or based on prior events, which give you a good idea of what will happen.
You bring up shot placement, and like I said above - unless it is related to your inability to handle the recoil (on the first shot) or your inability to compensate for recoil on follow-up shots, placement is largely independent of caliber. This of course assumes the target isn't more than 30 feet away.
A judge of the capability of the round in the given scenario wouldn't be "if I shoot this bullet, what is the amount of damage it will cause." But rather, it would be "if I hit someone in X (be it lung, heart, etc) how much damage will it cause?"
In your example, the question to ask would be:
If I hit the target in the head with ____ caliber, what is the likelihood that the round will penetrate the skull? If it does penetrate the skull, what is the likely amount of brain damage we can assume?
Yes, there are a lot of possibilities, and a lot of factors for each possibility (i.e. if you hit the target in the chest, what is the chance of hitting a major blood vessel, how much organ damage can you expect, etc. etc.). However, this is what we are comparing.
EDIT:
Using a .45 ACP 230 grain FMJ @ 850 fps as an example how do you reconcile the difference in performance between that bullet passing through the heart of a person and one that passes through someone's buttocks? They both possess the same "stopping power" (about "61"; "61" of what is another matter altogether) and should by that measure be equally "effective" (there's that word again) against an armed assailant.
Because the measure of power assumes decent placement. Okay, it's hard to define placement. However, we're not comparing a .22 to the hand and a .45 to the chest. We're comparing a .22 to the chest to a .45 to the chest. Or we're comparing a .22 to the hand to a .45 to the hand. Are all shots going to be in the same spot? No. However, I do believe if I got into a gunfight with a .45, and then you rewound time and (with everything else equal) gave me a .22, I would hit the target in the same spot.