The Padres moved into Petco Park in 2004. That was a year of high expectations, and it was the best performance of a Padres team since 1998. But
watching the first series against the Giants, which they won, you knew something was different. Ryan Klesko launched 4 or 5 balls to left-center that
looked like easy home runs, only to have them languish in the air and fall as warning track outs. This was going to be a pitchers' park, an intense
one.

But the Padres pitchers have not benefited as much as the Padres hitters have suffered, or at least have not benefited enough to claim a true
advantage. All teams in baseball have a home field advantage, measured by a factor I describe as my Net OPS Gain (NOG). I take the differential
between home hitting OPS and road hitting OPS (Hitting OPS Gain, or HOG, usually positive) and pitching OPS allowed at home vs. the road
(Pitching OPS Gain, or POG, also typically positive). Most teams both hit and pitch better at home. Yes, there are a few parks where this is not true
because the pitchers struggle at home compared to the road (Colorado, Arizona, Philadelphia, Texas, Kansas City), but they are the exception.
There are also six parks where the home team hits worse at home than the road (San Diego, Seattle, Dodgers, Angels, Nationals, Mets).

My conclusion:
My analysis tells me not to build "biased" stadiums. Having a heavily pitching or hitting favored ballpark hurts your team at home. The more
"balanced" the ballpark, the bigger the home field advantage. If you have biased stadium that you can change, I would change it. For Petco, I
would move in the walls, if possible, to make it a neutral stadium. My estimate suggests that the Padres could gain 5-6 wins over the 81 home games. That said, while the 2006 season seemed to conform to my conclusion, in the 2007 season, the Padres have a huge homefield advantage. I will be tracking this.

Contents:

How Does NOG Compare With Other Measures of Park Effects?
Rankings of Parks and HFA
Ballpark Effects on Wins

NOG and Park Effects
As mentioned above, my method for measuring at park's bias is by measuring the OPS differences in hitting (home vs. road) and
pitching (OPS allowed, home vs. road). A park will have a strong pitching bias if the home OPS allowed is much lower than the road OPS allowed, a
park like RFK or Petco. A park would be a strongly hitting park if their home hitting OPS was much higher than their road hitting OPS (like Coors
or Chase).

There already exist park factor measures, ones which measure each ballparks tendancy to allow home runs, triples, doubles, singles, hits, walks
and runs. As a check, I ran the correlations between the OPS Factor and the other factors, it correlated most strongly to the runs, which makes
sense. I think my factor controls better for the relative quality of the teams playing in different ballparks.

The vertical axis is the park factor measure from ESPN (measures teams runs on the road vs. runs at home) and the horizontal axis in my OPS factor. Both are basically measuring the same thing, but given that ESPN measure of park factors is more widely accepted and used, I wanted to make sure both measures were reasonably correlated.

Summary Of Parks & Home Field Advantage
Shown below is a summary of the last three years, averaging the OPS factors for all the ballparks. In the cases where new parks have opened (SDG,
PHL, RFK), the results from the old parks were excluded. I have ranked the teams, left to right, by the size of OPS improvement of pitching
(maroon). The further to left, the better the pitching was at home. I also show the OPS gain (or loss) for hitting (blue). The yellow is adding the two,
measuring the improvement in statistics from playing at home. All, except the Nationals, are positive, so all teams perform better at home.

The Padres, while performing better at home, barely do so, one of the smallest homefield advantages. At first glance, there does not seem to be a
relationship between a particular bias (strongly pitching, strongly hitting) and the overall gain. PHL looks as bad as SDG, but is a hitters park.
Colorado (DEN) (my abbreviations are named after the major city they are near), has the biggest gain. But further analysis shows that the ends of
this chart (biased towards hitting, pitching) do worse than parks in the middle.

One thing that does seem to show up, though, is that the total home field benefit seems to dip whenever a team has a negative bias in either way
(road hitting better or road pitching better). If you separate the teams into those that are "balanced" (meaning both pitching and hitting gains at
home), and those that are "unbalanced" (either negative hitting or pitching at home), you see that the gains are much stronger for balanced parks.

This summary suggests that it is better to be balanced than unbalanced. My hypothesis is that it takes too much adjustment, by either pitchers or
hitters, to fully benefit the ballpark. In the case of Petco, saying "swing normal on the road but do line drives at home" is too much. A strategy may
be to get players that are tailored to the ballpark. For example, getting high on-base, line drive hitters for Petco, or getting ground ball hitters in
Philadelphia.

Taking this further, I charted the gain of "balanced" parks, calculating the weight of the their home gain (e.g. 90% hitting, 10% pitching),
expecting to see the most balanced parks getting the most lift. The chart does not really supports that.

The chart above charts the homefield advantage (measured in NOG) on the vertical axis against the OPS Park Factor (how different a team hits at home vs. the road). My hypothesis was that you would see an inverted parabola-shaped curve, one that had low homefield advantage on the extremes (ballparks that strongly favored pitchers or hitters), and but increasing as the ballparks became increasingly neutral. The shape of the above curve does not support this. However, the shape of the trend line does seem driven by the outlier park to the upper right. This is Coors Field in Denver. It is notorious for its offense, and the Rockies hit so well at home (compared to the road), that it creates a very strong home field advantage. In 2006, the offense dropped way, way down as the Rockies began to humidify the baseballs used in the games, but it is a unique place because of the thinner altitude. If one eliminates Coors Field, a very different curve develops:

While there is still quite a bit of variation along the horizontal axis (OPS Park Factor), the shape of the curve is interesting to me. On the far left are SDG and WDC. These two parks have been by far the most pitching friendly. On the park right are PHX, CHA, TOR & PHL, all parks that strongly favor hitters. This chart suggests that it does not pay to be at either extreme. DAL (Texas Stadium), far upper right, is a counter point, but the other four extreme hitters park hurt HFA. Same with SDG and WDC.

Impact On Wins
So my last step was to quantify how much Petco's bias was hurting the Padres. To do this, I took the difference in the Padres home advantage,
measured in NOG, and compare it to "balanced" ballparks. This gave me a differential in net OPS between Petco and balanced ballparks. Next I
estimated the number of wins the Padres would gain if their home NOG matched the "balanced" parks. The impact is material.

I calculated the relationship between net OPS and winning percentage, creating the formula: 1.3235x + 0.4997, with x being the net OPS figure.
With this, I can calculate the impact on winning percentage by changes in net OPS, and apply this to the 81 games the Padres will play in Petco in
a season.

Balanced Team: teams that have both positive hitting and pitching OPS gains when playing at home (eliminates heavily biased parks)
Highly Balanced Team: within the universe of balanced teams, ballparks where the mix of hitting vs. pitching gains runs within the band of 30/70 to 70/30.

This says that the Padres would gave 5-6 wins if Petco played like a "balanced" or "highly balanced" park. In the tight NL West, that is probably
the difference between the playoffs and sitting home. My vehemence is a bit muted by the variation across parks, but being on the extremes of the equation, especially towards pitching, does not help. It is noteworthy that the two teams on teh extreme hitting side of the equation (Coors, Texas) really struggle maintaining good pitching. So while they get good home field advantage, it is difficult to maintain a good pitching staff when they get shelled at home as much as they do at those hitters' parks.

Next Steps
Update the data for 2006 and 2007. The Padres 2006 season looked very similar to the 2004 & 2005 seasons. So far in 2007, though, they are gaining a huge home field advantage, driven by huge gains from their pitchers pitching at home.

Broaden by years of data. Some of the variation I see I expect has to do with sample size. There is a lot of year to year variation in the home vs. road numbers, and I expect averaging this over a broader number years (when possible; some parks, like Petco, are new so will have a small number of years).