Kevin Schofield's writings, observations, and other pointless distractions
One of the less-well-covered aspects of baseball is what is known as “park effects.” The thirty stadiums (stadia? my Latin is rusty) are not clones of each other; they vary in all sorts of dimensions that are important to the game, such as the distance to the home-run fence, and the width of playable foul territory along the first-base and third-base lines. We can also throw in one-off stadium weirdness, such as the “green monster” in left field in Fenway Park, the ball-eating ivy-covered wall in Wrigley Field, or the uphill slope in center field in Minute Maid Park.
There is also climate — writ broadly here for the purposes of discussion. That includes temperature, prevailing wind speed and direction, humidity, air density (affected by both temperature and altitude, as any pilot can attest to) and sun brightness/direction. Those are also affected, and in some cases mitigated, by ballparks themselves, particularly ones with roofs or that can be entirely enclosed. And sometimes through other means, such as in Coors Field where baseballs are stored in humidors prior to games in order to prevent them from drying out in the thin, dry Denver air.
The sabermetricians spend a fair amount of time trying to determine how much of a difference individual ballparks have on teams’ and players’ performance. Fortunately, at least as a starting point there are some fairly straightforward metrics that we can look at to see what differs.
ESPN has a chart of six of these factors here, which is informative to look at. We’re going to look at stats for the entire 2013 season, since April is a cold, nasty weather month in much of the country and so the 2014 in-years stats, while interesting in their own, are not predictive of the longer-term trend. As a final caveat before we dive in, it’s worth pointing out that every off-season things get changes in some of the stadiums, so these stats frequently change across years: if, for example, the right field fence got moved in, then the park effect for that stadium will change.
A good place to start is to look at the first statistic: runs scored. This is calculated as the ratio of total runs scored by both teams in a team’s home games over the runs scored in its away games. Since teams are scheduled to play approximately an equal number of home and away games against the same teams, this is a pretty fair comparison to make. If the ratio is above 1, that means that more runs were scored in the team’s home park than when they met on the road — suggesting the park overall is biased towards hitters. If it’s below 1, less runs were scored in the home park, and thus it’s biased toward pitchers.
If we look at the stats, we see some expected things and some surprises. Coors Field has had the reputation from day 1 of being a hitters’ paradise (and where ace pitchers’ ERAs go to die), and the statistics bear that out. It’s interesting to look at the Coors stats over the years and see pretty wild variations for park effects on runs scored, even with the humidor. Petco Field also lived up to its reputation as an extreme pitchers’ park. On the surprising side, Chase Field in Phoenix is reputed to also be a hitters’ park, but last year it slightly favored pitchers. Safeco Field has a reputation as a pitchers’ park, but the stats don’t support that; last year it was almost perfectly balanced for runs scored.
Now this hardly tells the whole story, since not all pitchers are alike (nor are all hitters). Some pitchers tend to generate a lot of fly balls; they tend to do poorly in parks where the fences are close in, but do well in Safeco, Petco, and Marlins Park where it’s hard to hit a home run. Other pitchers tend to keep the ball in the infield, and are much less affected by the distance to the home run fence. Some pitchers are strikeout specialists, relying on one or two key pitches like a high-90’s fastball or a wicked slider that can be affected by the climate; and the extreme example of that are knuckleball pitchers. The knuckleball is a specialty pitch that is thrown without spin; where the spin on a pitch provides stability and also determines whether it will curve in any particular direction, intentionally throwing without spin means that it will be very unstable and can wander all over the place on the way to the plate. The best knuckleball pitchers in the Major Leagues were virtually unhittable — in certain parks. The movement on a knuckleball pitch comes from the surrounding air; they would do very well in Safeco Field, or Petco, or probably Wrigley Field, places where there is heavy, wet, windy air. But knuckleball pitchers do horribly in enclosed stadiums, or places like Coors Field with dry, thin air.
The home run stats for last year show that the use of the humidor has definitely brought down the number of home runs in Coors Field — meaning that its bias towards hitters comes from something other than the long ball. Flipping back through the last several years, it looks like there is large variability from year to year on the park effect in home runs — though this is an area that needs deeper analysis.
Triples is an amusing stat to look at. Not a terribly important one since not many triples are hit; but it’s the stat with the largest park effect variation. It correlates to parks with big outfields and particularly funny corners along the first and third base lines where the ball can take funny bounces. The vast majority of triples come from long line drives that rattle around in a corner and cause problems for the outfielder to predict where they will ricochet. AT&T Park is a good example of that; it’s actually a compact park that generally favors pitchers, but it led the league in triples last year. Check out right field — and it helps that it’s a long throw from right field to third base; if the wall goofiness were in left field instead, there would be far fewer triples.
As far as doubles go, there are clearly park effects but they are more nuanced. Large outfields encourages more doubles that go in the “gap” between outfielders. Doubles also come from long fly balls that bounce off the home run fence (or bounce over it). An extreme example is of course Fenway Park with the Green Monster in left field: close-in, super tall. Strangely enough, Rogers Center had more doubles than any other park last year… but it seems to be a one-year anomaly.
There is less clear data on park effects directly on pitching. Partly this is because almost all major league pitchers have at least 3 pitches they throw well, and can select based upon the opponent, the batter and the park — making it much harder to get meaningful statistics. In the last couple of years a new system called Pitch F/X has started supplying much more granular data on every major league pitch thrown, and that has opened up a whole new area for deep analysis. There isn’t enough data yet to draw many conclusions, but I expect we’ll know more soon.