I'll need to get to my home PC to post the formulas, but here is the basic overview.
Defensive Aggression
Data Used: Turnovers Forced from 2006-08, Interceptions from 2006-08, Sacks from 2006-08
The formulas weigh 2007 data as the most important, followed by 2008 data. Then there is a formula to normalize the data on a 0-100 scale. My goal was to get the Median and Average of the range between 45 and 55 and still have values on the extreme ends. While this method won't exactly duplicate the results of the 2008 season, they will make teams play more like their real life counterparts. Teams like Boston College and Wake Forest have forced turnovers consistently over the years and have high Def Agr values using my formulas.
Defensive Run/Pass
Data Used: Total Plays Defended from 2007-08, Pass Attempts Against from 2007-08, Pass Yds Against Per Game from 2007-08, Run Yds Against Per Game from 2007-08
There were about 5 formulas and I can't remember enough to describe them fully right now, but they pretty much took a team's pass defense compared to the league average and divided that by a team's run defense compared to the league average. That number was multiplied by the percentage of plays against that were passing plays that the team saw each season and then multiplied again by the passing defense on a yds per game basis versus the league average.
I probably just confused the hell out of you there, but at the end of the day the data supports which teams usually stop the run and which teams usually stop the pass.
The only factor I'd still like to look at is a team's defensive playbook and the number of defensive run versus pass plays in it (similar to your offensive formulas). Since my results have been much better in the sim engine so far I haven't really felt the need to go into that level of detail yet.