Has anyone ran any regression analyses on current ratings and player outputs?  Currently in a masters level statistics course and see some correlation here...  just wondered if anyone has done anything in regards to this?

Yeah, it's been a while though.  You just have to be careful about which variables you choose and how you interpret your results.  For instance, I got a negative coefficient on some rating that obviously positively affects performance (contact, eye, something like that).

I think I used 3 seasons of data from 50 different players, all from neutral parks.  If nothing else you get a decent formula to rate players.

I think this may be my "slows days at work in December" project.  Sounds like you may be able to extract enough data out of here to do some decent predicitons.

I bet you could get some really good numbers on baserunning, steals and times getting caught....

Anyone have any thoughts on doing this type of stuff to predict final ratings?  Obviously, very few players ever get to their projections... but I wonder if it's feasible to compute  final ratings of player... ?

PBM chess? And you still claim you were never in prison?

I did one using all major league players with a certain number of ABs, and a certain number of IP, about 15-16 seasons ago.  Dependent variables were (1.4*OBP) + SLG and ERA, respectively.  I found some pretty startling stuff, most of which has been borne out as I have played...  well worth doing if you have the time and skills.