10 replies to this topic

[Brett]

Something I can't seem to find: when making the park factor adjustment in OPS+ and ERA+, is *only* the player's home ballpark considered? Or does the adjustment take into account every park the player played in that year?

The distinction wouldn't lead to a huge change in the ultimate figure, but say, for example, a player plays in a ballpark that skews heavily offensive, but is in a division where the other parks in the division skew pitcher-friendly (for example, a player on the Diamondbacks or Rockies, maybe). If park factors consider only the home park, that guy would have his numbers adjusted based on being in a hitting-friendly park, but he also plays a slightly disproportionately more games in pitcher-friendly parks than a player, say, in the NL Central.

So, anybody know? I've been reviewing definitions and glossaries, but can't seem to find the answer. Crowd-sourcing is usually more efficient...

[hansman1982]

Take it for what it is worth, but Wikipedia says that OPS+ is adjusted for league factors, so one could infer that opposing ballparks are factored indirectly at worst.

http://en.wikipedia....e_plus_slugging

Then again the article also states that I lead the all-time leaderboard with a 1512 OPS+ and a career OPS of 5.001

[Brett]

Take it for what it is worth, but Wikipedia says that OPS+ is adjusted for league factors, so one could infer that opposing ballparks are factored indirectly at worst.

http://en.wikipedia....e_plus_slugging

Then again the article also states that I lead the all-time leaderboard with a 1512 OPS+ and a career OPS of 5.001

Yeah, but the league factor adjustment takes place independent of the park factor adjustment. You're right, that does take care of part of it, but there could still be a small bit that the formula misses if it only considers home park in the park adjustment.

[hansman1982]

I just submitted this question to the FanGraphs message board and will update with what I find.

[SirCub]

So it sounds like, ideally, you would have a Park Factor Adjustment, a League Factor Adjustment, and a Division Factor Adjustment.

[Brett]

So it sounds like, ideally, you would have a Park Factor Adjustment, a League Factor Adjustment, and a Division Factor Adjustment.

Yup (assuming the division isn't already in there, separated out).

[Brett]

I just submitted this question to the FanGraphs message board and will update with what I find.

Right on, hans. Thanks.

[hansman1982]

I think the question then becomes, is the amount of time a hitter spends in a non-home pitcher friendly park statistically relevant. The Cubs play something around 9 games at Busch Stadium. Lets say Castro plays in 8 of those games resulting in 35 PA's or 5% of his annual total. Is that statistically significant enough to break it out.

[Brett]

I think the question then becomes, is the amount of time a hitter spends in a non-home pitcher friendly park statistically relevant. The Cubs play something around 9 games at Busch Stadium. Lets say Castro plays in 8 of those games resulting in 35 PA's or 5% of his annual total. Is that statistically significant enough to break it out.

Yup. That is, indeed, the question. I'm no statistician, so I can't say.

[hansman1982]

Well they aren't answering my question over yonder. I did check wikipedia and it reminded me about sample sizes and I think that may fit into this a little bit.

Over 35 PA's there are far too many variables to be able to take opposing park factor into account (especially true for a 3 game, 11PA series at Fenway). You may go into an extreme hitters park and go 5-35 with all singles. Well, did you happen to face Halladay and Lee or Rodrigo Lopez and James Russell for 20 of those PA's. The park isnt affecting your stats but the pitcher is. Over the course of 325 PA's that is going to smooth out more.

I guess what I am trying to say, is based on what little I was able to find is that taking into account your Division's stadiums isn't statistically relevant. Which ties into your statement above.

[Brett]

Good stuff. Good talk.