>> over-penetration would actually be pretty low on my list of concerns about a submachine gun.
That's what I said. I haven't argued with #1 at all, in fact I've specifically agreed with it. #2 would be my biggest concern in deciding when to deploy fully-automatic weapons to routine law-enforcement. It's a common (and increasingly so) thing in the US to do is to issue semi-automatic only rifles to police officers for their patrol vehicles. In London you'll see a number of police officers at various high-risk locations carrying MP5s, slung while on guard. If I'm not mistaken, they are select-fire.
So in deciding which firearm I would issue for which deployment, what I'm saying is that my list of concerns about a submachine gun or similar firearm, whether or not the situation and the training level of the officers is appropriate for the potential of fully-automatic fire is far higher on my list of concerns than over-penetration. If you miss the target, everything is over-penetration.
I don't see how that's "absolutely absurd". I don't think data is available on how well officers manipulate safety switches under stress, but the rate at which even moderately trained personnel blow through all their rounds without hitting anything, military or SWAT, is astounding. So yes, don't use automatic fire if you don't need it. The reality is not so simple, IMO.
I'm perfectly calm. Please don't be condescending. I'm being insistant because you seem intent on missing the point, which is the following:
1) Single, controlled shots from pistols are less accurate than their equivalents fired from SMGs with stocks
2) Yes, automatic fire decreases accuracy. Don't use automatic fire where accuracy is a concern.
To claim otherwise is, again, absurd.