from Amazon
Tuesday, March 16, 2010
How much can a pitcher’s road parks affect his stats?
A lot! We sometimes talk about a team’s strength of schedule and how it might affect its win/loss record (not much, I think). We sometimes even talk about a team’s strength of schedule in terms of how it affects individual players. Of course, if you use BPF and PPF (batter park factors and pitcher park factors), like what you see on B-R, you are automatically adjusting for the pool of pitchers (for batters) and batters (for pitchers) that a player plays against, but even those numbers assume two things which are not true: One, that there is a balanced schedule, which there isn’t, and two, that players on the same team play an equal number of games against the same opponents, which they don’t, especially pitchers.
Anyway, I am not here to talk about the strength of a player’s pool of opponents, although that is an important discussion in and of itself. What about opponent park factors, or the average park factor that a player faces on the road? The same thing that I said above is true for road park factors. For one thing, the unbalanced schedule means that, for example, the Dodgers, Giants, and the Padres play a lot of games in ARI and COL, the two most hitter friendly parks in the NL. And pitchers, especially starters, because they don’t pitch every day, may play an inordinate number of games in one park or parks or another. This can make a big difference in terms of their raw, unadjusted (by their road parks) stats.
IOW, when park adjusting player stats, you really have to know their “average road park PF” rather than just using the “other parks factor” (which is 1 minus your home park PF divided by 13 or 15) which is traditionally used in park factor adjustments.
For example, let’s say that the Padres PF is .85 . That means of course that you would adjust a home stat (in runs) by dividing it by .85. Traditionally, for the road stats, you would use the “other parks factor” of 1+ .15/15, or 1.01. IOW, if the Padres have a PF of .85, all the other parks in the NL must have PF’s of 1.01 for everything to equal 1.00.
But, as I said, because they play lots of games against the D-Backs and Rockies, they play lots of games in two parks that have PF’s of around 1.10 and 1.20 and not 1.01 (the average park in the NL other than ARI).
Anyway, what I did was to compile the average road park factor for all players in 2009. You will be surprised at some of the results:
Here are some Padre pitchers and the average park run factor of all the road parks they played in prorated by the number of TBF in each of those road parks:
Peavy 1.00 So he did in fact play in average road parks (actually not the 1.01 that you would expect)
Mujica 1.03 So if his road ERA were 5.00, that would actually be 4.85 after park adjusting it, which would make a difference of .05 runs in ERA overall (as compared to if you used the generic 1.01 for his road parks)!
Chris Young 1.04 Besides sucking due to a large decrease in velocity, he also played in heavy hitters’ parks on the road, costing himself .075 in overall park adjusted ERA.
Gaudin 1.03
What about some COL pitchers? If you used the traditional “other park factor” you would adjust (divide) their road ERA’s by a factor of .988, since COL has a PF of 1.18.
DeLaRosa was .97, one reason why his stats were so good! He is benefiting overall by about .05 in ERA.
A bunch of relievers actually had road park factors of .94 and .95, even in a decent number of TBF, like Rincon, Daley, and Peralta.
Poor Glendon Rusch, aside from being a bad pitcher, his road park factor was 1.03!
Arizona pitchers have an “other park factor” (again, assuming that they play an equal number of games in all NL parks) of .993. Here are some pitchers who were pretty far from that:
Buckner .95
Petit 1.01
Vasquez .96
For batters, since they tend to play every day, the variations in road PF are not that great.
Here are the players, pitchers and batters, who were helped or hurt the most by their road PF’s, minimum of 150 TBF or PA in their road games:
Batters
Hurt
Carlos Gonzalez COL .96
Fowler COL .97
Iannetta COL .97
Carlos Ruiz PHI .97
J Hamilton TEX .97
A Jones TEX .97
Helped
Delmon Young PIT 1.03
McCutchen PIT 1.03
G Jones PIT 1.03
Pitchers
Hurt
Troncoso LAN 1.04
C Morton PIT 1.04
C Young SDN 1.04
Helped
B Buckner ARI .95
J Martin WAS .96
Sonnanstine TBA .96
Hunter TEX .96
Nippet TEX .96
Once you throw in the weather issues that I talked about the other day, and a player’s personal SOS (the quality of the pool of opponents they happen to face), I think you have quite a bit of context for a lot of players that is not accounted for by traditional park adjustments. Given the ease that these things can be computed with simple database queries, I would hope that sites like B-R will eventually include then in their OPS+, wOBA+, ERA+ and other stats.
And don’t forget about home plate umpires for pitchers. I was wondering when somebody else was going to realize what a big factor these unequal distributions were for projecting starting pitchers.