Lamefest Posted January 1, 2017 Share Posted January 1, 2017 I'm taking some Grad classes at UW-Milwaukee. One class I took this semester is Analytic Models for Managers, where we got to use a pretty helpful statistical modeling program called JMP. For my Final Project, we had to find a problem and database and build a model to help answer the question. At the risk of the Professor having no clue what the hell I was talking about, I decided to do my project on a Tecmo related topic. I figured I'd see if I could find any sort of reliable formula for Defensive player attributes by building an exploratory model on the 1990 statistics and TSB Running Power attribute. I wanted to post the paper here just because no one outside of this community probably knows what the hell I'm talking about and some may find it helpful in their ROM editing. After finishing the paper, I played around with it a bit more and tightened it up a bit. The final model actually has an R-square value of 0.562 which is much larger than I'd ever anticipate. While there's still a fair amount of RP unexplained by the model, using it would give a starting point for pain-in-the-ass Defensive attributes for seasons before mid-2000s. I think this gives a pretty good idea at which statistics had how much influence on the programmers as they determined attributes. While there's still a fair amount of RP unexplained by the model, using it would give a starting point for pain-in-the-ass Defensive attributes for seasons before mid-2000s. The formula would look like this (Y/N are 1/0 variables): Defensive RP = 49.79 + 4.965*(1990 Pro Bowl Y/N) + 5.687*(1989 Pro Bowl Y/N) - 0.003923*(Team Pass Yards Allowed) - 0.002814*(Team Rush Yards Allowed) + 0.6427*(Sacks*Defensive Tackle Y/N) + 0.7969*(Sacks*Rush End Y/N) + 0.8829*(Sacks*Linebacker Y/N) + 1.421*(Interceptions*Conerback Y/N) + 1.093*(Interceptions*Safety Y/N) + 4.975*(TDs*Cornerback Y/N) I'm going to try my hand at other attributes before my temporary license on this program runs out. I'll post if I get through any more. Grunze_Project.docx Quote Link to comment Share on other sites More sharing options...
bruddog Posted January 1, 2017 Share Posted January 1, 2017 That looks fairly similar in nature to the approximate value formula use by pro-football reference Where you could use the following 18+ AV =44RS 56 RP 69 MS 17-15= 38 RS 50 RP 63 MS 14-12= 38 RS 44 RP 56 MS 11-8 = 31 RS 38 RP 50 MS <7 = 25 RS 31 RP 44MS Quote Link to comment Share on other sites More sharing options...
bruddog Posted January 1, 2017 Share Posted January 1, 2017 Almost all the speed attributes are tied together for defenders(rs, rp, ms) other than interceptions. Aside from the players that are obvious errors. Bob nelson with 69rp and lonnie young with 69rs. Quote Link to comment Share on other sites More sharing options...
Recommended Posts
Join the conversation
You can post now and register later. If you have an account, sign in now to post with your account.