Petco Park revisited

My first ever article here at Hardball Times examined the effects of Petco Park and how its home team, the San Diego Padres, might better take advantage of it. Now that we have three more years worth of data to work with, I thought it might be time to revisit this question.

We know that Petco Park is a pitchers’ park, but let’s take a closer look at the shape of its bias. The following tables (derived from data at ESPN.com) will serve as a reminder of just how extreme the environment in San Diego is, and also that in spite of this, there are periodic fluctuations. I’ve added hits and walks this time around for the sake of completeness. In the tables that follow, values higher than 1 favor the hitter, while values below 1 favor the pitcher.

Runs

Petco Park, Runs Scored: 2004-2009
Year Runs Rank
2004 .837 29
2005 .803 30
2006 .860 29
2007 .755 30
2008 .796 30
2009 .741 30

That is some kind of consistency. Petco Park has been the least favorable environment for scoring runs in MLB for four of its six years in existence, finishing as runner-up the other two seasons. The scary thing is that even after it had established supremacy in that regard, the park managed to become stingier.

In 2009, the next lowest scoring environment, Cleveland’s Progressive Field, checked in at .838, a full .097 ahead of Petco Park. The gap between those two parks is much larger than that between Coors Field and Chase Field last year (.054) at the top. Here are the gaps between Petco Park and the No. 29 ballpark over the past three seasons:

Runs Scored, Gap at Bottom: 2007-2009
Year Petco Park No. 29 Gap
2007 .755 .833 .078
2008 .796 .842 .046
2009 .741 .838 .097

As a run suppressant, no other environment compares. This is why, if you’re Jon Garland hoping to cash in on one final multi-year contract, taking a little less money for a short-term deal in San Diego makes sense in terms of re-establishing value on the open market.

Homers

Petco Park, Home Runs: 2004-2009
Year Home Runs Rank
2004 .691 30
2005 .750 30
2006 .982 16
2007 .685 29
2008 .743 30
2009 .721 29

One of these seasons—2006—sticks out like the proverbial sore thumb. At the time, I posited a couple of possible explanations for this shift:

  • Buildings constructed beyond the outfield have changed wind patterns that may have kept more balls from leaving the yard;
  • Pitchers have grown accustomed to the way Petco Park plays and are more susceptible to making mistakes high in the zone because they believe—consciously or otherwise—they can get away with it

I neglected to mention a third possibility, which, with the benefit of hindsight, appears to have been the correct one: The spike was anomalous. This leaves unresolved the question of why, but for whatever reason, the Padres and their opponents hit more homers at Petco Park in 2006 than they had before or have since.

Padres, Home Runs at Petco Park: 2004-2009
Year Hit Allowed Total
2004 57 75 132
2005 54 64 118
2006 75 92 167
2007 72 45 117
2008 66 70 136
2009 61 68 129

Note in particular the home runs allowed by Padres pitchers in 2006 and 2007. They surrendered twice as many in the former as in the latter. Take the average of those two years and you get 68.5, which is in line with the surrounding seasons. It’s almost as if the universe, seeking equilibrium, forced Padres hurlers back toward the mean. That’s nonsense, of course, although I’d be lying if I claimed to have a better explanation.

Hits

Petco Park, Hits: 2004-2009
Year Hits Rank
2004 .895 27
2005 .901 28
2006 .905 28
2007 .856 30
2008 .896 29
2009 .805 30

Again, Petco Park consistently ranks at or near the bottom in terms of hits. It has been stingier over the past three seasons than the previous three, when it was already one of the most unforgiving in all of baseball.

Doubles

Petco Park, Doubles: 2004-2009
Year Doubles Rank
2004 .895 23
2005 .832 28
2006 .767 30
2007 .710 30
2008 .780 30
2009 .711 30

The first two seasons downtown were pretty extreme, and then things got downright ugly in 2006. That is the year the fence in right-center was moved from 411 feet to 402 feet, which theoretically could reduce the area available for hitting doubles. That the lower park factor has held for four years running lends further support to this hypothesis. Whatever the reason, Petco Park is where doubles go to die.

Triples

Petco Park, Triples: 2004-2009
Year Triples Rank
2004 1.519 5
2005 1.331 8
2006 1.086 13
2007 1.042 14
2008 .931 18
2009 .778 23

Triples have been harder to come by lately as well. Part of this could be due to the change in dimensions previously mentioned, although the shift has been much more dramatic with triples, which suggests a few other possibilities:

  • As Petco Park matures, teams are getting a better feel for how it plays and doing a better job of positioning their outfielders to minimize the likelihood of triples.
  • Triples are a comparatively rare event and therefore more subject to random fluctuation.
  • Some other factor(s) I haven’t considered.

The suggestion of randomness holds great appeal to me in general because in my experience, people often ascribe meaning to events that have none. However, the fact that the movement has been consistent and steady in one direction leads me to wonder: Do other venues fluctuate so much?

Well, yes, some of them do. The Texas Rangers’ home park, for example, went from 1.840 in 2005, to .780 in 2006, to 1.650 in 2007, to 2.230 in 2008. That’s a lot of movement.

But in terms of heading in one direction, no other ballpark in MLB has seen a consistent decrease (or increase) in triples from year to year over the period 2004-2009. Detroit’s Comerica Park comes closest, having its steady decline (1.840, 1.760, 1.170, 1.600, 1.110, 1.080) interrupted only by a spike in 2007. At the other end of the spectrum, the Florida Marlins’ home park had been becoming increasingly triples-friendly (1.030, 1.110, 1.410, 1.460, 1.480) until last year, when the bottom fell out (.640).

Right now I’m inclined to classify Petco Park’s downward movement as intellectually interesting but difficult to explain and of questionable impact. Still, it bears watching.

Walks

Petco Park, Walks: 2004-2009
Year Walks Rank
2004 1.046 8 (tie)
2005 .942 27
2006 1.012 15
2007 .906 28
2008 1.024 12
2009 1.142 1

This is all over the place, and I have no idea why. What is interesting to me is that there is a lesson here in misusing statistics. Had we looked at this after the 2007 season, we might have concluded that because the ball doesn’t carry at Petco Park and it’s easier to get away with mistakes, pitchers aren’t afraid to attack hitters, leading to fewer walks. However, someone looking only at last year’s data might reach the opposite conclusion and assume that because it’s easier to get hitters out at Petco Park, pitchers aren’t afraid to put someone on base every now and again.

The truth is, we don’t know (perhaps Adrian Gonzalez and the lack of fearsome hitters around him in the lineup are to blame?). It’s an oddity in the record, but I’m not prepared to imbue that oddity with meaning just yet.

The bigger picture

Finally, we return to the question I asked three years ago: “Were the Padres able to use any of this to their advantage?” One way of looking at this is to compare their expected (Pythagorean method, 1.83 exponent) record with their actual record at home:

Padres Actual and Expected Wins at Petco Park, 2004-2009
  Runs Wins
Year Scored Allowed Actual Expected Difference
2004 329 342 42 39 +3
2005 308 318 46 39 +7
2006 315 337 43 38 +5
2007 323 278 47 46 -1
2008 289 332 35 35 0
2009 278 321 42 35 +7

Thomas Thress has shown (By the Numbers, May 2004, p. 19) that there is a bias in this approach, noting that “from 1900 through 2003, the Pythagorean Theorem underestimated the number of home wins for 72.8 percent of all teams.” Thress continues by explaining why this is so:

When the home team wins a game, they bat in fewer innings than the visiting team. In most cases, the home team will not bat in the bottom of the ninth inning. Hence, they have one-ninth fewer opportunities to score than the visiting team.

This is something I had failed to consider in my original analysis. Worse, I’m no longer certain that we’re asking the right question here.

The Padres have outperformed their Pythagorean at home by 21 games in six years (486 games total). According to Thress, teams typically outperform their Pythagorean at home by four games per 162, so we would have anticipated something more along the lines of 12 extra home wins. In other words, the Padres have won about nine more games than expected at Petco Park during its existence. That comes to 1.5 wins per 162 games, or three-quarters of a win per 81 games (the number of home dates in a season).

There are two problems I have at this point. First, how much difference does less than one win a year make? I mean, so what if the Padres have won nine more games than expected over six seasons? That’s not a lot of extra wins, and it’s quite possible that luck is the overriding factor in the difference between expected and actual total.

Second, even if there is some skill involved in improving one’s home record by a microscopic amount in any given season, does this necessarily signify that the team is taking advantage of its park? I’m not sure it does. What might be better to study—and I’m not going to do it here due to space considerations and because I haven’t thought this through yet—is the difference between home winning percentage and road winning percentage as compared to that same difference for other teams.

Conclusion

Petco Park remains the most difficult environment in MLB in which to score runs, and by a wide margin. It appears to have gotten even more difficult beginning in 2007. It has consistently killed hits, doubles, and homers (except for the anomalous 2006 season, when Padres pitchers coughed up an extraordinary number of home runs).

Perhaps most interestingly, Petco Park apparently has gone from a place that initially yielded a lot of triples to one that suppresses them. Why this is the case and whether the current trend will continue remain unknown. Also curious is the yearly fluctuations in walks. In 2007, Petco Park yielded fewer bases on balls than all but two other parks in baseball; in 2008, it was middle of the pack; and last year, it became a walking man’s paradise.

The question of whether the Padres are using Petco Park to their greatest advantage remains open. This may be worthy of additional study.

References & Resources
ESPN park factors

Print Friendly
 Share on Facebook0Tweet about this on Twitter0Share on Google+0Share on Reddit0Email this to someone
« Previous: BOB:  Cubs tickets and a McCourt Update
Next: Why discretion is the better part of valor »

Comments

  1. Rich Campbell said...

    Triples and doubles declined dramatically after they pulled in Death Valley…far more because of that than any other factor.  Yet HR’s and runs scored showed little change.  What this says to me is that I was right to oppose that change.  It added no seating and made the ballpark less exciting.  Triples are arguably the most exciting type of “standard hit.” (Because let’s face it, most in-the-park HR’s are the result of an uncalled error or even an injury.)

    I still believe that the Padres can take advantage of that by emphasizing speed over power and by making an extra-athletic CF’er a priority.

    Great piece, Geoff.

  2. Kahuna Tuna said...

    FWIW, Gonzalez walked 73 times at Petco in 2009, 46 times on the road.

    Overall, Padres batters walked 331 times at home, 255 on the road.  Visiting pitchers had to know how little risk they ran in walking an extra batter or two against such a weak-hitting team — the Padres’ home OPS was .655.  Eeek.

    As for the Pythagorean formula’s predicting fewer home wins than expected, I don’t think the “problem” lies with the number of scoring opportunities per se.  I think it’s more because the home team could be expected to have more one-run victories, since a) most home victories decided after the eighth inning are by one run, and b) home teams can more safely play for a tie-breaking single run in the late innings.

    As I’m sure others have suggested, it might be worth exploring whether to apply separate Pythagorean exponents to home and road run differentials.

    Very interesting piece, Geoff — thanks!

  3. Bob said...

    Interesting piece, Geoff, and I don’t disagree with your conclusions, but I have a couple issues with the math.  First, in 2007 the actual wins of 47 less the expected wins of 46 should yield a difference of positive 1, not negative 1.  This changes the outperformance figure from 21 to 23, which means the Padres won 11 more games than expected at home over the 6-year period.

    Secondly, you correctly identified the number of games in that 6-year period as 486 so to calculate how many more games the Padres won over a 162-game period, you should divide by 3, not 6.  Using 11 more wins instead of 9, this comes to 3.7 wins over a 162-game period or nearly two wins per 81 games.  That’s still not enough to move the Padres up in the NL West so I believe the fact remains that the Padres are not taking full advantage of their park’s characteristics.

  4. Jonathan Sher said...

    Geoff,

    Interesting article. I tend to agree that one needs to ask a different question to better understand if the Padres benefited from their home park.

    At first blush, if your aim is to measure whether the Padres have structured their team to benefit from the extreme pitching environment of Petco, I don’t think it’s at all useful to compare actual wins and wins predicted by the Pythagorean method.

    If the Padres were making great use of Petco, one would expect they would be scoring more runs and/or giving up fewer runs than what they would in another, more average, ballpark. That would increase the number of wins predicted by the Pythagorean method. And if the Padres were neither lucky or unlucky, that would also increase their actual wins by the same amount. There would be no gap between Pythagorean wins and actual wins.

    I agree with your inclination to instead compare how the Padres do on the road versus home, and that’s where the use of the Pythagorean method may prove useful. Compared to other teams in more neutral parks, do the Padres achieve a higher ratio of home to road Pythagorean wins?

  5. Jim T said...

    Considering how bad the Padres have been the past few years, its probably better to evaluate the performance of the away teams to determine how Petco plays.  In the meantime, I agree with Rich Campbell that the loss of death valley was problematic for triples.  Dumb idea.

  6. Tom Waits said...

    @Mike, @Kahuna Tuna

    Petco may make the Padres’ offense look awful, but it’s not. It wasn’t great, but they put up league-average OPS+ the last two seasons (96 and 95 versus an NL average of 94). All teams managed only a 661 unadjusted OPS in Petco.

  7. Marver said...

    2006 was an abnormally hot year in San Diego.  Part of the reason—actually, pretty substantial part—Petco plays as a pitcher’s park is due to the relative low temperature of most games due to the marine layer bringing in cool (the moisture plays a minor roll in the opposite direction) air.

    Since 2006 had an average—during the baseball months—high nearly five degrees warmer than the historical average, the ball carried a lot better that season.

    Global warming will royally screw with park factors. raspberry

  8. Marver said...

    In case you want evidence of that, here’s wundergound.com’s San Diego by month historical weather report.  It’s on July 2006…you can change the month and you’ll see the center column, with average max and average min temperature, is abnormally higher for 2006.

  9. Geoff Young said...

    Thanks, all, for the comments.

    @Rich: I believe that athleticism in the outfield, particularly CF and RF, is a must at Petco Park.

    @Kahuna: Thanks for the info on the walk splits. Regarding the notion that the Padres were a weak-hitting team in 2009, it’s worth noting that their pitchers held the opposition to a 663 OPS at Petco Park last year. The park severely distorts everyone’s numbers.

    @Bob: Yes, you are correct on both counts. As you say, it doesn’t change the conclusion, but thanks for catching this.

    @Jonathan: Agreed. The gap between actual performance at home and on the road vs league norms would be a useful thing to study.

    @Mike: As Tom Waits notes and I’ve alluded to above, the Padres offense was awful only inasmuch as it had to contend with Petco Park. Following are runs per game scored by the Padres on the road (middle column) as compared to runs per game scored by NL teams (right) since 2004:

    2004: 5.42, 4.64
    2005: 4.64, 4.45
    2006: 5.14, 4.76
    2007: 5.16, 4.71
    2008: 4.30, 4.54
    2009: 4.44, 4.43

    Last year, the Padres basically had a league-average offense away from Petco Park. The year before, they were slightly below average. In each of the four seasons preceding that, the Padres scored more runs per game on the road than NL average in all games.

    @Marver: The weather angle is fascinating. I was thinking wind currents at the time, but temperature makes sense.

    Thanks again. You all have given me good food for thought.

  10. Jonathan Sher said...

    Marver,

    Cool link, or should I say hot. Even more meaningful than the daytime high is the night time low, assuming a majority of games were played in the evening. And it appears the gap between the night time low and the average low was even greater than the 5 degree difference in daytime highs.

  11. Marver said...

    As an undergrad Physics student at UCSB, I had to do a research assignment on a California Physics phenomena. Naturally, I chose to investigate how the marine layer affected run scoring at Petco Park. Good to see it have some use to someone/thing besides the GPA.

  12. Brian said...

    This is a really interesting article, but it would seem that most of the variation in these offensive statistics would be attibutable to the performance of the teams that play most in Petco. To get a more accurate measure of park effects you’d have to take into account the quality of the Padres hitters and pitchers in a given year, as well as the power and on-base prowess of the other NL West teams that play lots of games in Petco.

    In 2006, the Padres pitching was very prone to the longball: see Woody Williams and Chan Ho Park. In 2007 it got better because Peavy and Young probably worked on reducing home runs allowed and the Padres acquired Maddux who keeps the ball down.

  13. Geoff Young said...

    @Brian: Thanks for the comment.

    I’m not sure how best “to take into account the quality of the Padres hitters and pitchers in a given year.” Would we look at that season’s road numbers for individual players? This could be too small a sample. Or historical numbers? This could be influenced by too many other variables (players age, get hurt, etc.). Or something else?

    As for the ‘06 staff, you are right about Williams. However, Park was coming off a season in which he surrendered just 11 homers in 155.2 IP while spending most of the season with the Rangers, who play half their games in a hitter-friendly environment.

    It’s also worth noting that both Williams (67.2 IP, 11 HR home; 77.2 IP, 10 HR away) and Park (69 IP, 12 HR home; 67.2 IP, 8 HR away) allowed more home runs at Petco Park than they did on the road that year. If they were so susceptible to the long ball, shouldn’t that susceptibility have carried over to their games played in less pitcher-friendly venues as well?

    The big change was Young, who inexplicably stopped serving up home runs in San Diego:

    2006: 86 IP, 18 HR
    2007: 85.1 IP, 1 HR

    Regardless of how much Young may have “worked on reducing home runs allowed” (and I’m not sure how he would have done that anyway), this represents freakish improvement by any standard.

Leave a Reply

Your email address will not be published. Required fields are marked *

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>

Current day month ye@r *