Quote:
Originally Posted by Lukas Berger
The league totals being short would be a League Total modifier issue as you note. The difference between the starters and RP's comes down to individual ratings though most likely and that might be ZIPS/player ratings related. I can probably fix that but it'll take some time.
|
I imagine it would.
I have sort of tried to
reverse engineer the correspondence between Stuff/Control ratings and strikeout and walk rates. I'm sort of curious what you're using to generate the ratings from ZiPS projections. What I noticed in my model was that a "50 Stuff" for a starting pitcher resulted in a 19.0% strikeout rate, on average, but a "50 Stuff" for a reliever resulted in a 23.0% strikeout rate, on average. I don't think this is a problem, since an average reliever does strikeout more batters than an average starting pitcher, but I kind of wonder if that's accounted for in your ratings generation scheme? It would explain the discrepancy, possibly.
For example, Cody Allen has an 80 Stuff in the 2018 rosters. In my model, that predicts him for a K/9 around 12.90, despite his ZiPS being for 11.22.