Wars do that. The tolerance level goes down.
As it should. I can believe the percentages and they are probably about as good as you can get for accuracy based of human fallibility. Shame about the 20% in one sense but I bet if you dig, you'll find out most were the result of a series of small events that culminated in selection of one trifling event to provide a legally sufficient 'cause.' IOW, probably a lot more were truly deserved than is at first readily apparent.
ADDED: I don't know the circumstances on the 1SG cited but if a Co sized unit is having a spate of negligent discharges; somebody's wrong and needs to go. Hopefully, they also nailed in some fashion (not necessarily an Art 15) each individual and his immediate leader for each of said discharges. If not, that may be why the 1SG got zapped...
We don't relieve enough for cause and the tendency to just move people who don't perform is a bad peacetime habit that needs to be stomped out. It's partly induced by DOPMA and partly caused by political correctness and the insane requirements imposed on anyone who just wants to rid the Army of the excess deadwood that accrues at the >16 years of service mark. We should tolerate honest mistake and errors in training -- even some errors in judgment -- but clamp down viciously on malfeasance, obfuscation, evasion of responsibility and leadership, tactical and technical incompetence or failure (especially that last cluster).
The Peter principle is valid, the Army can at times prove that with a vengeance -- and in peacetime, it's ignored. That's bad. In wartime, it is not ignored, the longer the war and more people exposed to combat, the more rigorous they are in culling. That's good.