October 30, 2007
Mailbag at Altitude
I received a decent number of questions about my park effects piece from last week, so I think it's worthwhile to spend one more column rooting through the mailbag and discussing a few loose ends. The extremely short World Series-indeed, the extremely short postseason, with seven series played in just four games over the minimum-has taken some of the urgency out of the long-delayed umpire discussion.
Leading things off, reader C.S. asks:
One question that I've always had on park factors is: is it a straight up comparison between runs scored at home versus runs scored away or is it taken into consideration that the home team does not always hit in the bottom of the 9th, thus the home team's runs scored are suppressed? In other words, the away team always hits nine times (barring the occasional rainout), while the home team many times hits only eight times. If it's simply a comparison of runs scored by a team home versus away, wouldn't it be more accurate to compare runs scored per plate appearance at home versus on the road?
In the classic park factor formula (which we cited in last week's super-sized Further Reading section), the first adjustment that's made is to correct for the disparity in innings between the home team and visitors. Regardless of the details, runs per inning would be a more relevant basis to measure than runs per plate appearance.
Reader M.H. chimes in:
I have no problem with the concept of park factors in general-clearly it's going to be easier to score runs in some parks than in others-except when they change from year to year when the dimensions of the park itself do not. Fenway's dimensions have been basically static for decades, so how are the scoring numbers fluctuating?
M.H. raises two issues. The first is a fallacy that probably stems from the metric's name: park factors are not solely controlled by the dimensions of the playing field or the ballpark; wind currents play a huge role in baseball, and even if your team keeps its fences in precisely the same place over the years, changes to or near the ballpark (like the construction of the auxiliary press boxes at Fenway, or even the construction of high-rise buildings nearby) can affect those currents. Even if you make no changes to the ballpark at all, the weather is going to be variable, with different humidity and temperatures from year to year; that will affect how the ball travels, and (as we see in Colorado) can even change the properties of equipment.
Also, since we're comparing these park effects to the average, "neutral" park, the conditions at every other ballpark in the league affect your ballpark's rating. If a divisional rival builds a new ballpark that's a bandbox, your park's factor will likely decrease in comparison; if cold, gusty weather suppresses run scoring in another rival's ballpark, your own building's park factor is likely to see a boost. So the idea that the park factors should remain static because the ballpark's fences haven't been moved doesn't work.
M.H.'s second issue makes some intuitive sense-it figures that if you suddenly get better personnel in your lineup, your ballpark will look like an easier place to hit-but that's not how park effects work. Take the 1992 Red Sox, for example. That team was probably the worst-hitting Red Sox team in the last fifty years, failing to crack 600 runs scored in a 162-game schedule, and hitting only 84 homers all year. Fenway's park factor that season was 107, well above average and slightly higher than it had been the year before. In 2003, when the team hit 238 homers and came within shouting distance of 1,000 runs scored, the Fenway park effects were actually lower than they'd been in 1992.
How can this be? The important thing to remember is that we're not looking at the absolute run scoring levels-we're looking at the comparison between how both the home team and the visitors perform, both in and away from the ballpark in question.
Shifting our attention to the ballpark where the Red Sox just secured their second World Championship of the 21st century, reader A.C. set out to blow my mind:
In 1,395 games before the humidor (1993-2001) and 973 games since its installation (2002-2007), (what's that, about 100,000 total PAs?-that's a pretty decent sample size)
My numbers came out a little different from A.C.'s (my spreadsheet came up with 46% and 13% boosts, respectively, in the humidor era) but the differences weren't significant enough to obviate the point: that's an amazing advantage. Other teams in the 2002-2007 period enjoyed boosts of up to 15 percent greater run scoring than their opponents at home. The extra boost the Rockies get from their home park is more than twice that.
While A.C. suggests that this disparity might be the result of humidor-related foul play, the fact that there was a considerable advantage prior to the humidor's inception-not to mention the rough treatment the Rockies got over the weekend-suggests otherwise. While the humidor has made a big difference in Denver, the lack of moisture in the baseballs obviously isn't the only side effect of altitude. We could be looking at opposing teams suffering mild (or not-so-mild) altitude sickness, or the home team simply being more familiar with the ballpark.
Regardless of the cause, the Rockies' advantage in runs scored at home is real. I'll leave you with a look at the home field advantage (home winning percentage less road winning percentage) throughout the majors for the pre- and post-humidor periods:
1993-2001 2002-2007 TEAM Win% HomeAdv. TEAM Win% HomeAdv. COL .478 .135 COL .460 .184 FLO .451 .123 TBA .394 .140 TEX .507 .105 TEX .479 .136 SLN .510 .101 WAS .471 .121 PIT .443 .095 HOU .522 .117 SFN .534 .086 MIL .443 .114 NYN .505 .081 SEA .498 .105 DET .438 .078 TOR .495 .100 MON .480 .076 MIN .550 .098 PHI .472 .074 OAK .563 .097 ANA .475 .072 CHA .526 .093 OAK .491 .071 FLO .499 .091 SDN .482 .071 BOS .579 .088 SEA .538 .067 SLN .564 .088 NYA .590 .064 PIT .434 .082 CLE .572 .063 KCA .401 .077 CHN .474 .062 ANA .562 .076 ARI .528 .062 NYA .609 .075 HOU .543 .054 SFN .524 .073 CHA .523 .053 ATL .568 .072 LAN .522 .050 LAN .526 .072 TBA .406 .048 SDN .489 .067 TOR .496 .047 ARI .490 .066 BOS .529 .038 CLE .503 .064 MIL .455 .033 DET .437 .050 MIN .441 .033 CIN .461 .049 ATL .611 .031 NYN .495 .037 BAL .503 .027 BAL .441 .037 CIN .507 .022 PHI .529 .034 KCA .459 .009 CHN .488 .023
Jason Pare contributed research to this article.