Spos,

Interesting results. The one thing that is missing are the stats for the complete season in the test replays. However, while acknowledging the possible perils of a small sample size, your results suggest that the first half numbers (presumably before the PTL had much, if any, impact) were closer to "reality" than the second half numbers (presumably influenced by PTL limits) were. This suggests that the limits you set or the stat engine itself were too agressive and may require some further tweaking.

Do you have any sense about how individual stats, especially for those players on whom the limits were most strongly imposed, were impacted?