Parallel Parking

Dave Paisley

With a hundred or so games on the books for most teams, I was curious to see how the various park effects are playing out. Early in the season everyone was all excited about the apparently phenomenal offense that Enron Field in Houston was generating. And, of course, every year there's the usual question about how far beyond normal Coors Field will be. Some years it's been fairly modest, most years it's unbelievable.

One other thing to bear in mind is that there are many ways to view how particular ballpark affects play. The usually accepted definition of park effect focuses exclusively on aggregate run scoring. Total runs scored per game by both the home and road teams is divided by runs scored per game by that team and its opponents in all of the teams road games. For instance, to take an example from the table below, if the Cardinals score 6.00 runs per game at home, while their opponents score 5.14, and the Cards score 5.51 runs per game on the road, while their opponents score 4.98 in those same games, then the Busch Stadium overall park factor is equal to:

(6.00 + 5.14) / (5.51 + 4.98) = 1.06

Note: A park factor of 1.00 would indicate the park is league-average for run scoring, while a lower number indicates it's tougher to score in and a higher number means it's easier to score in.

In other words, 6% more runs are scored on average in games at Busch Stadium than in road games involving the same teams. You'll notice in the table that I have computed park factors separately for the home teams and their opponents. I like to see this information because it surfaces trends that are buried in regular park factor. Looking at St. Louis again, you can see that the Cardinals' own park factor is 6.00 / 5.51 = 1.09 (Park Factor Home Team in the table). The aggregate of their opponents when playing against the Cards is 5.14 / 4.98 = 1.03 (Park factor Road team in the table.)

Team	R/G Home	R/G Road	R/G by Opp Home	R/G by Opp Road	Park Factor Home Team	Park Factor Road Team	Overall Park Factor
Col	8.78	3.47	6.96	4.62	2.53	1.51	1.95
Hou	5.60	5.05	6.15	5.71	1.11	1.08	1.09
Atl	5.31	4.98	4.69	4.24	1.07	1.11	1.08
Cin	5.04	5.06	5.42	4.65	1.00	1.17	1.08
StL	6.00	5.51	5.14	4.98	1.09	1.03	1.06
Pit	5.09	4.56	4.95	5.35	1.12	.92	1.01
Phi	4.47	4.56	4.90	5.09	.98	.96	.97
Az	5.04	4.75	4.25	4.82	1.06	.88	.97
Mil	4.28	4.52	5.28	5.35	.95	.99	.97
Fla	4.47	4.75	4.93	5.00	.94	.99	.96
NY	5.07	5.12	4.39	5.02	.99	.87	.93
Mon	4.76	4.96	5.11	5.69	.96	.90	.93
LA	4.78	5.65	4.51	4.98	.85	.91	.87
Chi	4.54	5.37	4.50	6.24	.85	.72	.78
SD	4.33	5.07	4.23	6.04	.85	.70	.77
SF	5.24	6.37	3.82	6.33	.82	.60	.71

Normal park factor calculations assume that a park affects both teams equally or, if it doesn't, that the effects are due to some randomness that can't be accounted for. I beg to disagree. Tailoring a team to a particular park is a common occurrence. Whether it's worth it or not, GMs try to do it.

There's also a general home advantage: most teams perform better at home than on the road, so the home team's park factor is generally always higher than their opponents'. This could be because the offense scores more runs at home than on the road, or because the pitching suppresses opponents' runs better at home. In the above table, only five teams have lower park factors than their opponents, and only one of those by a significant margin (the Reds, who allow themselves to be spanked like naughty children by teams visiting Cinergy.)

The most obvious anomaly in the entire baseball universe is, of course, Coors Field, and this year we have a stunning example of thin air disparity. It's also a very interesting example of radical disparity between park effects on the home team and visiting teams. The Rockies score 2.5 times as many runs per game at Coors as they do on the road -- an unbelievable split.

On the other hand, iheir opponents "only" score 1.5 times as many runs in Denver, which is more in line with Coors' historical numbers. It's interesting to speculate on why this might be, and one obvious explanation is that the Rockies stats come from a relatively small group of players who may have become attuned to Coors, while the road team stats are actually a composite of all 15 other teams (and some AL West teams, too) and thus represent more of a true league average.

Which do you think is more likely to be representative of pure park effects? I know I would assume the composite opponents' numbers. Another explanation for the wide variation is that the Rockies numbers, being a much smaller sample size, is prone to wilder swings around the mean. On top of that there is also the aforementioned tendency of home teams to perform better -- due to home cooking, familiarity, not living out of a suitcase, or whatever other reasons you might be able to dream up.

Another weird anomaly is the curious case of Wrigley Field. It's been a steady hitter's park year in and year out, but this year it ranks tougher to score in than the notoriously pitcher-friendly Chavez Ravine. Has the wind been blowing in all season?

To take another case at random, the Diamondbacks' numbers are somewhat typical. The park is pretty neutral for run scoring overall (.97), with the Diamondbacks (1.06) doing significantly better than their opponents (.88). Given that all Diamondbacks pitching is not created equal, I wondered what effect Randy Johnson might have on the numbers, and if he had pitched more at home or away.

Sure enough, he had pitched 13 games at home and 10 away, indicating that there may be some influence from the random chance that he pitched more at home. However, if we assume that he gives up maybe four runs per game less than another pitcher and it only takes one and a half games to even up the home/road split, the six runs is nowhere near enough to close the gap. It would change the .88 to .90 and the overall park factor from .97 to .98. Interesting to note, but not worth getting excited over.

In fact, I'd say that park factors are generally only good to the nearest 5%, so anything between .95 and 1.05 is essentially neutral in my book.

A couple of final thoughts regarding the new ballparks. Enron is certainly no Coors, but it is right up there at the leading edge of the rest of the pack, and the Astros and their opponents are making just about equal use of it to bang out runs. PacBell, on the other hand, is at the opposite end of the scale.

Note that the Giants and their opponents both score over 6 runs per game in Giants road games (Giants pitching not as good as you might think?). The Giants, however, suffer only a modest drop-off at home, while their opponents appear to run away and hide like Pete Rose from a good hairdresser. This gives the Giants a substantial advantage in runs scoring at home, which they have taken advantage of despite their lousy home start.

There's lots more interesting stuff in there, and with only a third of the season to go, some of those numbers are destined to stick around by the end of the season. It will be fun trying to make sense of it all.

about the author

Dave Paisley is currently computing the Bat Speed Park Factors for both leagues. Ask whether the warm Florida air will make Henry Rodriguez a good hitter at drdjp@strikethree.com.