I ran some numbers for the top 25 scoring teams in two worlds (Argento vs Bryant) at about the point in their seasons (end of playoffs 19 vs 17 games) to give a justifiable comparisons. Because I can't compare all teams I must assume these comparisons would be maintianed for the rest of the teams in the world. Here are some comparisons and comments:
D1AA - Argento has only 4 human coaches so the comparison will be pretty much Sim coaches vs the combo of human/sims in Bryant.
So with a Sim vs human/sim mix - I see some differences which could be explained by the fact that the team ratings difference in 2.0 should lend itself to some increased values (total points, total yards, perhaps rush yards) Sacks for the best 2.0 teams (probably human coached and recruited) are lower than for the mostly sim 3.0 version. Glaring differences are for interceptions and forced fumbles. The 1.9 for fumbles is the average per team for the season. Minor problem is the 20% difference in rush average.
More of an even mix of sim and human coaches in this comparison. Points and total yards are not too far off. But Pass yards vs running yards are way off as we human coaches figure out you can't run effectively in 3.0 and passing is hard to stop. Rushing yardage is 50% in 3.0 compared to 2.0. Sacks are higher due to the increased passing plays. INT and FF again are way off between the games.
Again a comparison of sim/human mixed worlds. This shows a better indication that the scheme of the engine is looking at ratings numbers themselves as a means of success vs ratings/player match-ups. As these world mature through 4 seasons, better teams should be able to put up higher values, but in 3.0 the numbers for total points, passing and rushing yards are all lower. Rush average is still low, sacks up due to increased pass attempts and INT and FF still very low compared to 2.0.
Comparing these worlds I would say: Rushing needs to ramped up to get a 4.5 average of so for the best teams. Interceptions and forced fumbles need to be increased - but using reasonable combinations of ratings to target those players who are prone to fumble. Sacks need to be looked at, but they may fix themselves if passing isn't as overwhelming.