I certainly found your posts interesting. My immediate thought was "If the .40 then why not the 9 since the .40 bullet is held by a bigger circumference area than the 9?" My next thought was that the pressure is limited by the tension in the case wall pressing inwards on the bullet and that as the 9 has a smaller diameter then this will produce a greater inward pressure for the same tension. Greater pressure over less distance probably equals the same total grip.
There is no doubt that setback bullets will produce higher pressures. I doubt if a 0.001" will make much difference but by the time we get to 50 or 60 thou instead of one, that could be significant. What you say about LEO's is interseting. The problem was recognised by Fairbairn and Sykes back in Shanghai. The instituted a controlled cycling of their ammunition so that all duty rounds were shot in training within a three month period. They also had a system of preventing individual rounds being chambered too many times. I don't remember the figures of the scheme but it was all set out in tables.
As a round is chambered it follows a random path as it bounces from surface to surface. Part of this randomness would almost certianly exist even if every round was absolutely identical (see chaos theory) but in fact there will be a variation in charge and bullet weight, a variation in case wall thickness and consequent possible tension, and a variation in crimp. As you have shown with hand cycling rounds, the speed of the slide after recoil is not the same as with hand cycling. This is because the slide bounces off its stop after recoil and so has a running start before the recoil spring starts to accelerate it towards the breach. So a slighty heavier charge could produce a faster slide pickup and if that is followed by a bad bounce for the next cartridge it could cause exceptional setback in a cartridge with a weak grip on its bullet and that could do it! It takes a lot of fairly rare things coming together but KBs are very infrequent.
I am not trying to knock Glocks, I think that they are a brilliant design and no engineer can anticipate all possible problems, but there seems to be some evidence that this is more of a problem with Glocks than with other makes. The same goes for the .40S&W compared with other cartrridges, though all have the same problem at, perhaps, a lower frequency. One of the complaints about .40S&W Glocks, which I share, is that they were built on a barely converted 9mm platform. This was, of course why the .40S&W was invented in the first place! Many people complain about its snappy recoil and I think that is because the slide is not heavy enough. The momentum imparted to the pistol by the bullet is much the same as a .45 and the most significant thing is that the bullet has left the barrel before the breach unlocks. The felt recoil is almost entirely a matter of the way that the slide moves after the bullet has left the barrel. I believe that the snappiness of the .40 cal Glocks is because of the extra speed with which the slide hits its stop. If this is so then the speed of return of the slide will also be greater because it starts off with a bigger rebound. It all fits together reasonably well as an explanation. Congratulations!
It might even be possible to find forensic evidence in the form of a cartridge case showing evidence of the bullet being driven back before it is fired. Apart from that it should be possible to do lots and lots of painstaking test and measurement to produce a distribution curve of the ammount of setback so that an estimate could be made of the frequency of sufficient setback to cause a KB. Fire a round. Take the chambered round out and measure it. Repace it as carefully as possible from another magazine. Replace the original magazine. Fire the previously measured round. Take out the chambered round etc. A few thousand round should be enough - of each of several manufacturers of course. I would not like to do it! In any case we can't assume a nice distribution curve or that our batch of ammo is representative.
The phenomenon that your hypothesis does not seem to explain is how some KBs combine with a partial breach opening. (I repeat that simply firing out of battery will not produce excess pressures and cannot seize the case in the chamber but would just blow the case without damage to the barrel.) Somehow the pistol has to be in battery before it is fired and then excess pressure still has to exist as the pistol comes out of battery.
Maybe I will think about this another day.