Commenting on Shane Spencer ...
IMO, and speaking from years of making rosters and running leagues going back to Microleague and EWB and extending through Tony and BBPRO to HH, is that you HAVE to be consistent.
If you're claiming that REALISM or RE-CREATING the season is your goal, then:
[1] Make Spencer's ratings as good as his stats.
[2] Only play him in September.
To have him use a "stud's" ratings for the entire season is not realistic nor accurate, and it changes it from a "XXYY mod" to a "fantasy league" or a "What If Season?". Spencer is not a stud, nor was he a stud for that whole season. To recreate that in a MVP mod would be silly.
I'm trying to remember the backs of my old 88 Donrus rated rookie cards, but I believe Gregg Jefferies hit .500 and Jack McDowell had a 1.93 ERA, and I tihnk Mark Grace hit high .300, even though they were just Sept call ups. So, if someone were making an 88 mod (I think one already exists), would they make these 3 guys have ratings that would make them 3 of the top 10 in the game?
Yeah, it's your mod, and you do with it whatever you want. We're just talking. Just keep it consistent. If you're going for realism, go for realism. If you're going for fantasy, go for fantasy.
Edited to add that my tone above likely sounds much more brash than intended. I just didn't/don't want users to run into some of the sme problems we used to when using guys with a small data sample or abbreviated season stats.