Quote:
Originally Posted by thehef
Curiosity got the best of me so I did a quick test of two good-hitting pitchers from 1930: Red Ruffing of the Yankees (.374, 37 hits in 99 AB's, 4 HR's) and Frank Shellenback of the PCL Hollywood Stars (.279, 36 hits in 129 AB's, 4 HR's). Unfortunately, regardless of adj/wkn settings, they import as anywhere from poor-to-mediocre-at-best hitters.
This is concerning because both of these guys were used almost 20 times in 1930 as pinch-hitters. Not having them available (well, available as the good-hitters they were) is a break from the realism that we're all generally hoping for... I'm pretty sure from what Markus wrote on this topic like 3-4 years ago - "Pitcher ratings are adjusted to sample size, so a pitcher with 100 AB in a season, even if he performed good, will be rated lower than a position player with 600 AB and the same stats" - that this is by design. However, I'm not sure the intent is for these guys to be as poor hitters as they are... I mean, they are not good hitters at all if not adjusted/weakened, but once they fall below the thresholds, they are bad. And that's not realistic.
|
Really, broad realism and one-year recalc via blunt algorithm are basically an incompatible ends / means pairing. There are lots of similar problems. For example: a player who was an All-Star in years X-2, X-1, X+1, and X+2, but got only 100 ABs in year X due to injury is treated no differently for that year than some lifetime minor leaguer who got called up for the final month and feasted on September pitching (including a lot of pitchers who he had just spent five months batting against in AAA, though know these same matchups are given the frisson of “MLB-quality”). Similarly, the 4x multiplier for starter thresholds (vs reliever thresholds) makes modern seasons virtually impossible in terms of having enough decent starters without permitting a number of small sample size relief heroes.