October 23, 2007
A Tale of Two Ballparks
Lots of people are excited to see the Colorado Rockies in the World Series, but statheads are probably watching this series with a special glint of joy in their eyes. You see, for the past fifteen years, the Rockies have been the focus of one of the great inquiries in baseball-how do you win at altitude? Performance analysis can be defined as the study of baseball in context, and since 1993, the city of Denver has given the major leagues one of the most fascinating contexts in its history: a relentless high-run environment.
Sure, there have been extreme run-scoring environments throughout the game's history, but usually those have resulted from the vagaries of ballpark construction, especially parks built too big or too small. Often, the situation could be remedied (or at least adjusted) by moving the outfield fences in or back, or adding or removing rows of seats. In contrast, Denver's run-scoring environment is more a matter of atmosphere than of ballpark dimensions. Higher altitude means thinner air, and thinner air means less friction to slow down thrown or batted balls, or to make curveballs curve. The conventional wisdom was that you could move the fences around as much as you like, but absent a pressurized dome, you weren't going to have a normal game of baseball in Colorado.
So the questions that follow from that environment, with a ballpark effect this extreme, have been how to take advantage of it, or how not be prejudiced on account of it. In his tenure as Colorado GM, Dan O'Dowd has famously tried a number of different tactics-obtaining contact hitters, or changeup pitchers, or big whiffers with power-often with little to no positive results. Until this year, baseball in Colorado was a bit like the movie March of the Penguins-an often-heartbreaking story of a dedicated group struggling to survive in a hostile environment.
With the Rockies facing another team, the Boston Red Sox, whose home ballpark is considered a unique environment-and an environment that has from time to time been considered a possible hindrance toward building a successful ballclub-this seems like a good time to take a refresher on the basics of ballpark effects. I'm also throwing in an extra-long "Further Reading" section, so that you can review some of the things that we've written about baseball at altitude.
The concept behind park factors is pretty simple. By comparing the home and visiting team's runs scored or allowed in and away from a given ballpark, you can calculate how much the ballpark increases or suppresses run scoring, above the league average. While the concept and the inputs are pretty simple, the actual calculation is rather challenging, and made more so by unbalanced schedules and interleague play. Park factors are habitually arranged on a 100 scale, where numbers above 100 indicate a percentage of scoring above average, and numbers below 100 indicate a score-suppressing effect, relative to the average. You'll also notice that batter's park factors (BPF) and pitcher's park factors (PPF) are calculated separately, and sometimes diverge by a few points.
Let's look at the ballpark factors of the two World Series sites over the years:
Colorado Year BPF PPF 2006 108.3 108.3 2005 111 111 2004 120 117 2003 112 111 2002 121 119 2001 122 119 2000 131 128 1999 129 126 1998 119 120 1997 123 123 1996 129 129 1995 128 128 1994 116 118 1993 120 122
Boston Year BPF PPF Year BPF PPF 2006 101.7 101.7 1991 105 105 2005 103.2 103.2 1990 105 104 2004 106 105 1989 106 106 2003 105 104 1988 104 104 2002 98 99 1987 103 102 2001 101 101 1986 100 100 2000 101 101 1985 104 103 1999 105 104 1984 105 104 1998 100 99 1983 107 107 1997 102 102 1982 106 106 1996 103 103 1981 106 106 1995 102 102 1980 105 105 1994 105 105 1979 106 105 1993 108 108 1978 111 109 1992 107 107 1977 112 111
As you can see, the numbers fluctuate from year to year, and the range of fluctuation is substantial, more than ten points over the time that MLB has been in Colorado, and twelve or so points over a thirty-year period in Fenway's history.
Why do these fluctuations happen? It helps to keep in mind that since you're measuring against the league average, the target is often moving-other teams change ballparks, plus there's expansion and realignment-but a certain amount of randomness also factors in. To reduce the effect of these fluctuations, park factors are usually measured in multi-year averages, where possible.
That said, 2005-2006 seems to represent a serious change in the trend for Coors Field, a reduction from "an unbelievably good place to score runs, like hitting on the moon" to merely "a very, very good run-scoring environment." In this regard, the numbers don't tell us why the change has happened-park factors are results-oriented. The accepted wisdom is that there's a "humidor effect" from the temperature- and humidity-controlled conditions in which baseballs are currently stored in Colorado, with the assumption that's what is suppressing the park's run-boosting abilities, but the park factor can only tell you that run-scoring in Denver is down, not why.
Increased ball velocity and trajectory data may some day give us some of those answers. Dan Fox, for example, has done some very promising work with pitch F/X data and weather conditions. Nonetheless, because of sample size issues there is no defined park effect for "when the wind is blowing out at Wrigley" or how very late October weather will affect how Fenway, or particularly Coors, will play.
Finally, as some of you might remember from the interminable Non-Contact series, we don't need to restrict park factors to just runs scored and allowed. The same methodology can be used to determine if a ballpark suppresses or increases strikeouts, walks, home runs-basically, any statistic-above or below the league average. Also, we can look at the different effects the ballpark has on left- or right-handed batters, which is a good idea when dealing with an asymmetrical ballpark like Fenway:
HR 1B 2B 3B BB K COL RHB 108.4 108.6 105.7 127.3 100.3 93.4 LHB 105.3 107.5 104.1 113.7 103.7 94.1 BOS RHB 95.3 99.3 113.4 86.4 99.9 101.5 LHB 90.3 100.8 112.8 107.0 99.2 98.6
It's mildly surprising that Coors Field seems to give right-handed hitters more of a boost over their lefty counterparts than Fenway does-the Green Monster's main benefit to right-handers is a significant boost in home runs.
William Burke and Jason Pare contributed research to this article.
Total Baseball: For those of you who want to do the math, the glossary of this site contains a formula for calculating park factors (as well as a number of other sabermetric treats).
Tom Tango, The Book Blog, "Run Impact in Parks": The comments to this post, particularly show some creative ways that component park factors can interact to create a better picture of the ballpark.
Rany Jazayerli and Keith Woolner, Doctoring the Numbers, "Coors Field" and "Coors Field Revisited": In this pair of 2000 articles, Rany and Keith examine and debunk the so-called "hangover effect" of playing at high altitude.