I have decided that the idea of a 16 subtype model is ok with me.
For one, it prevents you from sticking everyone in a too-small box. Now I understand that, by expanding the theory to include 16 subtypes, or an additional type, I'm really just making the box bigger, so to speak. But I think that it's an appropriate box-expansion. Let me explain.
Personally when I think of what a "subtype" means, it's really just a way of isolating a person's general focus in the world from his or her actual personality. For example, I am probably EIE, but I have a lot of interests in the realm of Ti: symbolic logic, philosophy, political theory, religion, psychology, etc. I think of FeNi as "the bigger picture," the lense through which I view the world, and the things I focus that lens on and become accutely interested in tend to be Ti in nature.
To me, the flaw of talking about "subtypes" in the way that they are currently discussed leads to fallacies, such as assuming that an Ni subtype EIE is more like an LIE in terms of having stronger Te and weaker Si than an Fe subtype EIE. Now this may have some basis to it, in terms of relative functional strengths, but the idea that an Ni-EIE is "more like" an LIE leads to ideas like "an Ni EIE values Te more than an Fe-EIE," which may be true in some cases, but in reality, the use of an "emphasis" on an ego function in order to explain behavioral differences implies connections and transitions within the Quadra value system that may or may not be true; there very well could be an EIE who seems more withdrawn or phlegmatic in temperament, and therefore more like an Ni subtype, but in actuality places greater emphasis cognitively on Fe than, say, your typical cheerleader-esque female EIE who is more likely to be seen as Fe subtype, but who very well may have more pronounced practical interests.
Basically this is all about navigating the grey areas of the theory, and is incredibly imprecise, but it is ringing true with me and I'm going to run with it.