In October of 2008 I did some regressions on using the stats and ratings of all the players who meet a certain playing time threshold in one of my worlds. The linear equation I came up with for hitters was. The idea was to try and figure out a formula that converted ratings to OPS or ERA. This is what I got:
For Hitters
0.218+(Contact*0.001711)+(Power*0.002918)+(VsL*0.001002)+(VsR*0.001471)+(Eye*0.001298)
For Pitchers
12.422+(Control*-0.01791)+(VsL*-0.01557)+(VsR*-0.02664)+(1p*-0.01779)+(2p*-0.0279)+(3p*-0.00467)+(4p*-0.00382)
I
used these formulas as an important part of my player evaluation when I first started. I think they helped me get a better understanding of the relative importance of each. I do not think the actual formulas work as intended, as I do not believe player performance is linear and certain combinations of ratings tend to outperform what one would expect based on the formula. Additionally, the formula did not deal well with the fact that each pitcher had a different number of pitches.
Finally, I would like to point out that I believe that my use of formula for my first few seasons hurt my development as a player because it kept me from understanding the nuances of the game and in some cases lead to me completely misevaluating players particularly those who had good ratings in most places but were significantly deficient in one but why would "score out" as average.