So if you care about small differences, why don't you consider the huge differences that no one even addresses of the large gap differences between impulse response and the actual cab. You have guitar cabinets, it shouldn't be difficult make an impulse response and see that the current way of doing impulses isn't accurate at all by any means.
Done that for a long time, man, also to show that certain IRs are not really that "perfect" It's just not common for people to do comparisons of that sort. That was the point. Kemper popularized such comparisons. It may eventually become popular with IRs as well.
The reason why I like using IRs is not because I think they are by definition more accurate than kemper cab.
There can be an IR that sounds better for you than a more accurate one due to all kinds of "coloration". That is part of why I have at times profiled amps just to suit certain IRs, monitoring direct amp signals live to begin with. Kemper has advantages too here, as partly explained in your post.
If that cab changes tone when converted to KPA format, and it's easy to test whether that happens or not, apart from personal opinions, that is an issue. It may not be something many care about. Ok, but if you use IR and load into kemper, that can be a shortcoming. That's all, really.