Seems that in all cases you can always measure maximum overall distance, regardless of how the shots are dispersed, outside-to-outside, then subtract 1 bore size. That gives CTC equivalent.
While it's hard enough to imagine what the software is modeling to predict a dispersion at all (if indeed that's what it's doing), it's even harder for me to imagine what basis it would have for modeling dispersion in two axes. For example, one could assume a muzzle velocity distribution and from that get a vertical trajectory distribution (one axis). But you'd have to be modeling drift to predict a horizontal axis trajectory variance. (Drift is the change in bullet trajectory due to aerodynamic lift acting on the bullet due in part to its angle of attack in flight).
Anyway, providing a single MOA value could be handy because that value allows you to calculate extreme spread at
any range--although that calculation is going to become less valid as range increases.
A single MOA value is somewhat like the 'circular error probable' (CEP) values calculated for mass consumption Powerpoints in the world of orbital mechanics, and re-entry vehicles in particular. So, take a nuke re-entry vehicle as an example. These are designed to have extremely predictable trajectories, but variables impact those trajectories just like they do a bullet.
The best models of reality predict the re-entry vehicle will impact Earth somewhere inside a very long, narrow ellipse. But the Powerpoint audience doesn't want to hear about ellipses or how long they are (the equivalent of the max group dimension). So, the models calculate the area of ellipse, then calculate the radius of a circle having that same area. That radius is then the figure cited as the 'circular error probable'. It's utterly meaningless, but easy to understand.
That sort of measurement would actually be more representative of probable
hits in shooting than is the overall max size that we use, in some ways. It's neither the worst case nor the best case, but a balance between the two. We seldom are asking "What is the most likely distance my next shot will be from my aim point?", which is the question better answered by a value somewhere between the min and max group size. More often, we're asking "How big will my group be?", and that's best answered by the value gotten from measuring the biggest dimension of previous groups.
I thought one of the main purposes of shooting was to get away from the pain and agony of computers and software. 
