As for the point about matching real life as much as possible, I'm curious, OR, as to your feelings (I just had this discussion with billyg) on this: how can you have realistic results for things like two teams going uptempo fastbreak/press? answer you can't for any one game - just like if you had a rex ryan / kurt shotenheimer run passing offense, you couldn't have realistic results versus the league average for NFL passing, other than it is real, they exist on the edges - if half the teams ran unrealistic stuff this may be an issue, but my sense is the average team in hd runs the game somewhat average, heck, near half the teams are sims, they run totally average pace and zero defense.
A couple of related things, some half theorized that the average defense in the game is about -1.5, which should cause 3pt shooting to be slightly higher, off rebounding to be slightly less, and 2pt fg% to be slightly less.
I would assume a pair of test teams could be rigged up that would represent absolute average, simmed vs each other 10,000 times, and a true fg% could be determined, sounds like a fun experiment. But, in leu of such a test, this is strictly opinion, if you look at naismith just completed d1, and look at fg% # 1 vs FG% last, and lay out the remaining 320 teams or so between them in a graph, I would guess a bell curve exists with a mean and a mode. The same can be done for real life NCAA. How do they compare? Same can be done for other stats. Same can be done for all std deviations.
Overall, I think this game would beat the pants off competing products in such a test. The purpose of my post is to urge the coaching population to try and keep it that way.