So as explained the AI servo is predictive. It will adjust focus based on what it tracked previously. The subject may not be on any AF point when the shutter fires. The AF points are used to complete the math for when the subject is not on an AF point. It's not just tracking what's on an AF point.
Most people don't understand what's going on in the background when they select one case or another and tnhen blaim the camera is not working right.
I haven't been working with it long enough to blame the camera for anything yet
The gist of my initial inquiry is: why doesn't the camera SHOW you what it is tracking because it evidently IS doing it anyway? The shots and metadata show it.
The answer seems to be: it will, but only in selected modes... opening a different can of worms: why?
Case #1 didn't handle the same scenario very well (though I've since learned quite a bit). Case #6 is the next I tried because it "fit" the description of the shooting scenario the best. The results confirm that. A very high in-focus hit rate. It remains to be seen if zone-expansion was a factor in that success.
Since then, I've tweaked the Case#6 settings a little AND set the camera to FOCUS priority. Next shoot, I'll try those settings an then various combinations and permutations on them. Considering how little time I've spent behind the viewfinder, it's taken remarkably little time to start producing decent shots. Mostly because I'm still at the steep end of the learning curve, NOT because of any inherent limitation in the camera. In fact, unless I see something impossible in an image, my natural assumption is "I'm doing it wrong."
I'm trying to learn it step by step, section by section. The lenses are all AFMA'd now (FoCal). Now my head is starting to wrap itself around the AF. I don't think I'll be moving from Av/Tv/AutoISO to M any time soon though