Fielding ratings

[fearofmimes]

It does not appear that the fielding ratings included in CSV and SQL dumps are related directly to data in the database files. I am referring specifically to the following fields in the CSV dumps:

fielding_rating_pos1
fielding_rating_pos2
fielding_rating_pos3
fielding_rating_pos4
fielding_rating_pos5
fielding_rating_pos6
fielding_rating_pos7
fielding_rating_pos8
fielding_rating_pos9

Are these calculated fields? I have figured out how the values that show up in the CSV dumps are mapped from the players.dat file, but I can't find a logical place in the record structure for these rating fields. It seems to me like there must be some kind of calculatons for each position where the component ratings are combined in some way and then scaled based on the experience value. For example, every non-zero value for fielding_rating_pos is accompanied by a positive falue for the corresponding experience field.

I'm wondering if I'm on the right track. I don't have anything against doing a little modeling to derive these values if I need to, but I would hate to travel too far down the wrong path if my assumptions are incorrect.

[YoVivoPorSPORTS]

In the game there are two things shown, range and % so shouldn't there be another field (or maybe both are encoded in the field you have there, but that would then make me wonder where the catcher and outfielder arm ratings are). Also I haven't looked at the structure or fields but maybe it is related to speed and stealing ability which are two seperate entitites when displayed in game.

[fearofmimes]

What I have been able to figure out so far is that experience has a linear relationship to the calculated fielding ratings. Beyond that, I'm at a loss. This is probably a result of my poor statistical analysis skills. What I have observed is that for a particular rating, if you vary one of the component ratings (infield arm, infield range, etc.) while holding all others constant, the resulting variation in the fielding rating is proportional linearly. However, the amount of variation for any given increase depends on the values for the other component ratings. I am experimenting with some other analysis techniques to see if I can isolate the relationships that I believe to exist.

I'm wondering if anyone has any insight or ideas about how these ratings are calculated.