I am running a replay league. I use lahman's database, and no matter what I do, certain players underperform. I am not looking for players to recreate exactly what they did in 2003, but to be close. I went into the player ratings, and realized that switch-hitters had massive splits. In my mind, switch-hitters should bat the same against lefties and righties, so I set their batting average to be the same against both. Result: statistical accuracy.
Now, I would take a player who would have a real-life batting average of 290, and his splits would be higher in both cases [say 294 and 297]. So I would create another split league from lahman until I found what I wanted. I want lefties to hit worse against lefties, righties to hit worse against righties. Result: Moderate success.
I set all pitchers with the exact same batting average against for lefties and righties, as I did not want to double the effect.
Now here is my problem: Some players, no matter which split I use alway underperform. No matter what split I use, they underperform. How can I fix this? Example: Travis Lee, a lefty has splits at 243 vs L, 293 vs R, and he historically should be at 275. This split looks correct at a quick glance. I simmed approximately 20 seasons, and the highest he ever hit was 249. I have the same problem with Brian Giles, Richard Hidalgo and Geoff Genkins, as well as numerous others. I tried them in parks with 100 factors across the board, no difference.
I enjoy OOTP, so please don't bother with telling me about Diamond Mind or Strat. I am interested in how to fix it with this system. If anyone wants a copy of the league file, you can get it from
www.wearetheshow.com/lgfile.zip
Any help would be appreciated.