September 7, 2009
Ahead in the Count
Home-Field Advantage, Part Five
This is the fifth and final article in this series on home-field advantage. The first four parts of this series have revealed many things. In the first article of this series, we studied what home teams are able to do more frequently than road teams; we learned that they pretty much do everything better, hitting more home runs, reaching base more frequently on balls in play, walking more often, striking out less often, stealing more bases, making fewer errors, and recording more complete-game shutouts. In the second article, we learned that nearly all home teams enjoy relatively similar home-field advantages over time, with the exception of the Rockies, and that the vast majority of year-to-year fluctuations in teams' home-field advantages are random fluctuation. The third time around demonstrated the important role of distance and familiarity in determining home-field advantage, and noted that home-field advantage was much larger in interdivision games than intradivision games, and was especially large in interleague games. We discovered something quirky in the fourth article of this series, that not only was the first game of the series not any more likely to exhibit home-field advantage, but the penultimate game was. More peculiarly, it was statistically significant, indicating that it is not all that likely to be merely noise. I received many e-mails and comments suggesting reasons that this peculiar effect may be real, or expressing skepticism that it is more than merely noise. This indicates that there is probably more to be learned about home-field advantage, and more that is not immediately obvious.
Although the second article taught us that team-specific home-field advantage is small, that does not necessarily mean that player-specific home-field advantage is as well. The first article taught us that home teams are more likely to succeed in hitting home runs, so we are left to wonder whether home-run hitters more likely to have a higher home-field advantage. A thoughtful commenter pointed out in the second article that team-specific home-field advantage seemed to be growing in recent years, and in fact, it has gone up, from about 6.6 percent in 1998-2002 to 8.9 percent in 2003-08. Perhaps teams have been able to improve their home-field advantage by acquiring players who better suit their stadiums. Furthermore, we know that if teams are more likely to hit home runs at home, we naturally are left to wonder whether players who hit more home runs will have higher or lower differentials between their home-run rates at home opposed to those on the road. In this article, we will test several statistics for this trend.
Naturally, testing these type of things requires knowing that players have persistent home-field advantage. This is definitely not something we can assume, because we know that teams do not. Only looking at players who did not switch teams and who were able to acquire 150 at-bats at home and on the road in two consecutive years between 2001-2008, we see that the difference in between OPS at home and on the road for a player has a year-to-year correlation of .1863, which is quite large given our sample of 960 players. It is worth checking whether park factors were playing a role here, so I used ESPN's park factors for the year in question, and normalized the home statistics using park factors listed (hits, doubles, triples, home runs, walks). Since one-year park factors fluctuate a lot (there is no stadium that is going to be twice as likely as the average stadium to surrender home runs), I also developed regressed park factors that were merely halfway between the recorded park factor and 1.00. Either way, this correlation came out large: .1752 for given park factors and .1615 using regressed park factors. The year-to-year correlation for this set of hitters was also pretty high for various other statistics. (Note that all numbers from here out are normalized using the regressed park factors.)
Stat Year-to-year correlation HR/AB .0852 AVG .1038 OBP .1892 SLG .1247 BB/PA .1101 K/PA .0914 BABIP .0830
This year-to-year correlation was not as strong for pitchers who only had a slight tendency to continue being relatively better at home if they had the year before. Using regressed park factor-adjusted ERA on a sample of the 539 pitchers between 2001-08 who threw 40 innings or more both at home and on the road for the same team for two consecutive years, and we get a correlation of .0299. This was not much different with no park adjustments (.0394) and un-regressed park-factor adjustments (.0520).
It seems that if some pitchers do have a persistent ability to be particularly good at home, this would come primarily through their ability to strike people out., and somewhat through their ability to avoid walks.
Seeing as it is pretty clear that hitters and pitchers both have some sort of persistent ability to do better at home, it's worth checking whether home-field advantage is going to be particularly helpful for players who have certain skills. Upon looking at the idea, this seems to be quite true, and for many skills. For example, the correlation between OPS and the difference between OPS at home and on the road (all normalized using regressed park-factors) is .1452 for the 1,824 hitters who got at least 150 at-bats at home and on the road in years between 2001-08. The correlation between ERA and the difference between ERA at home and on the road (again, all normalized the same way) is -.1048. It seems like better pitchers and better hitters will show larger home-field advantages. Consider the correlations between the following statistics and the differential between them at home and on the road, using the same normalized park factors:
Correlation between itself & Stat its home/road differential HR/AB .0707 BB/PA .1129 K/PA .1127 BABIP .1127 BB/K .1429 ISO .0696 SLG .1216 OBP .1517
For pitchers, we get the following:
Looking at these, it seems pretty clear that players who have certain skills are particularly adept at employing those skills while at home. This seems particularly true for nearly all statistics, and especially for those related to strike-zone management.
As a side note, I thought it would be interesting to check the effect of age on home-field advantage. In discussions with David Cohen of TheGoodPhight.com about home-field advantage a while back, he questioned whether older players may have larger or smaller home-field advantages. On the one hand, they may be more familiar with batter's eyes on the road and see the ball more clearly there. On the other hand, travel may be more difficult for them. In fact, the latter effect seems to be slightly more prevalent: the correlation between OPS home-road differential and age is .0884, and the correlation between ERA home-road differential and age is only -.0294. The former is statistically significant, while the latter is not. Perhaps pitchers (in this study, starting pitchers generally, since relievers rarely get 40 innings both at home and on the road) are better able to adjust their bodies on the road to be ready to play, while everyday players may struggle a little more with travel as they age.
This is the fifth and final article in this home-field advantage series. We have learned many surprising but also a few unsurprising things throughout this series. We have learned that teams do not have very persistent home-field advantages, with the exception of the Rockies, and that home-field advantage affects nearly every aspect of a player's game, especially in exhibiting skills in which they are particularly strong. We have learned that travel and distance both play important parts as far as home-field advantage, and we have learned that not only is home-field advantage not especially large in the first game of a series, it is actually more prevalent in the penultimate game of the series by a statistically significant amount.
I will certainly revisit various ideas to test them later, and I may write those up in an article at a later date or an unfiltered post. There certainly are a lot of other possible ways of looking at home-field advantage, and I can see from reader comments and emails that this is a topic that interests many of you. Please feel free to contact me if you have ideas of your own about home-field advantage that you would like to work on, or if you would like to test my conclusions further. I think that I took a pretty big bite out of home-field advantage in this series, but I'm sure that I missed some things. After all, I did most of this research in transit-while on the road. Perhaps you can do better at home.