Talking Baseball

Your weekday baseball fix. Some days.

Posted by Dave on Monday, May 10, 2004

The Curious Case of Kauffman Stadium

We often rest on the assumption that objects are more or less immutable. Sure, shirts may accrue stains, a car may acquire dents, and CDs may gather unwanted scratches. But, by and large, objects remain what they were when originally purchased. A cell phone is still your cell phone - even if it has its battery replaced.

Don't be concerned, this is not a philosophical discussion of objects' perdurance or endurance. It is merely an observation of the curious fluctuations of park effects. How did I stumble on such a bizarre observation? Well, Kauffman Stadium has been something of an enigma in recent years, as was discussed recently in a post by Jurgen over at Some Calzone for Derek.

What is a park factor? In essence, it is the contribution a park makes to the inflation or deflation of runs. A neutral park has a park factor of 100. An example of a notoriously good hitters' park is Coors Field - many more runs are scored there as a result of the conditions (i.e. the park creates an effect). Dodger Stadium is a notoriously good pitchers' park, runs aren't easily scored there. It is important to note that a park effect is independent of a team's composition.

Well, one would think that park factors are constant because the parks are unchangeable (aside for structural changes, like moving the fences in). Truth is, however, the park factors are quite fluctuant. Take a look at some particularly strange park factors (courtesy Baseball Reference, the batting and pitching effects were averaged):
Colorado Rockies (Coors Field):

Year: Effect:
2003 112
2002 120
2001 121
2000 130
1999 128
1998 120

Los Angeles Dodgers (Dodger Stadium):
Year: Effect:
2003 94
2002 92
2001 91
2000 93
1999 97
1998 93

Anaheim Angels (Edison International Field, now Angel Stadium):
Year: Effect:
2003 94
2002 97
2001 107
2000 102
1999 100
1998 101

Texas Rangers (The Ballpark at Arlington):
Year: Effect:
2003 110
2002 111
2001 100
2000 105
1999 104
1998 104

Kansas City Royals (Kauffman Stadium):
Year: Effect:
2003 113
2002 116
2001 110
2000 104
1999 101
1998 105

The park factors for Dodger Stadium and Coors Field should come as no surprise. The variability within the last three is strange, however. Edison International Field has suddenly become a pitchers' park, the Rangers are recently enjoying more of a hitters' park, and the Royals have found themselves a hitters' park equalled by only Coors. The Royals, so concerned by their park's effect, even moved out their fences this year.

What I find perplexing is that the park factors can fluctuate so violently if the parks themselves have not changed. For the life of me, I could not track down any alterations to the home fields of the Angels, Rangers, or Royals. So, it would seem that there would be no reason for there to be this much variability in recent years. So, how can we explain this?

I've browsed about, and I've surmised that Kauffman Stadium, Angel Stadium (Edison International Power lost its ad rights following last season), and The Ballpark at Arlington have recently been affected by chance. These three parks were hand-picked amongst the thirty teams in an effort to ascertain what was causing these recent fluctuations. Upon closer inspection, however, many parks have these fluctuations - they're just not as recent. For example, Yankee Stadium played like an extreme pitchers' park in 1999, yielding a park factor of 92. It generally plays as a neutral park, however. Similarly, in 2002, Veteran's Stadium played like an extreme pitcher's park with a park factor of 92. It's safe to conclude that many deviant park factors may just be aberrant for a specific year. An aberrant year could mean that a team would just happen to score far more runs on the road - even though there's no reason for it other than chance. This is likely what happened to the Yanks in '99 and the Phils in '01.

But what about multi-year departures from the norm, as with Kauffman Stadium? Well, there is always the possibility of having the same bizarre circumstance occur each year - but the likelihood is limited for three years. I have a hunch, however. The creation of the park factor formula supposedly should negate the effect of a team's pitchers or hitters. However, in casual observation it would seem that an increase team's runs scored is correlated with an increase in a team's park factor and vice versa. Conversely, it would seem that an increase in a team's runs allowed is correlated with a decrease in that team's park factor and vice versa. I have a feeling that a team's propensity to score or allow runs changes its tactics, and that this may explain some of the externalities associated with runs scored and runs allowed. These are just idle threats for now, but with the help of JC over at Sabernomics, I'll be able to give you and show you some regression results next time (the display is what I'm wrestling with).

A Small Celebration:
This is Talking Baseball's 100th post. I was made aware of this because of Blogger's new, beautiful, and intuitive structure. Recently, the stream of posts has slowed down as the workload for the writers has increased, but we've always tried to adhere to our motto of "Once a day. Every day. Your baseball fix." That was not our slogan formerly, but the consistency is something we've always strived for. In fact, up until today, we've had 100 posts in 112 days. This is more impressive when you dismiss our recent inability to post: Prior to April 22nd, we had 94 posts in 94 days. Some of those were short posts, some were bad posts, but the majority were interesting, thoughtful, well-written pieces. Thanks to the readers, and hopefully the prosperity will continue. Only 1,700 hits away from 10,000.

### So what do you think? We want to know. | | E-mail us ###