Is Oakland’s Mount Davis Killing Fly Balls?

My favorite part of my job is when players ask me questions. It’s difficult enough to come up with questions on a daily basis, so it’s great to get a free piece — and it’s even better when the question came from someone who plays the game every day. Once you make it to the Show, it’s all about staying in the Show, and that means making the most of your athletic talents. Strategy is often the key component to these questions.

When Athletics infielder Jed Lowrie came bounding across the Oakland clubhouse to me with his question earlier this year, he’d already decided to act on what he had perceived as an issue with his new/old home park. In the spring, he’d connected with his hitting coach, Darren Bush, in order to work on going the other way since he was leaving Houston’s friendly confines for Oakland’s cold. Because fly balls die in Oakland, and opposite-field fly balls are, by nature, less damaging than their pull counterparts, part of that new “oppo” approach was a heavier ground-ball profile. Mission accomplished.

But the reason behind Oakland’s fickle fly-ball play was still on his mind. “I think it’s Mount Davis,” he said back then. His theory was that the wall-like 10,000-seat expansion in center field — constructed in 1996 and nicknamed Mt. Davis in scorn after the Raiders’ late owner Al Davis — was responsible for suppressing fly ball distance in the Coliseum.

Answering his question turned out to be fairly difficult.

For one, we can’t simply examine the advanced ball-in-play data from 1995, to see what it was like before Mt. Davis, because that data doesn’t exist. You also can’t just look at actual results and try to graph them against the outfield, because there are too many variables. One such variable — temperature — is particularly relevant for Oakland. On sum, it’s one of the three coldest parks in the league, and heat is good for distance, and also exit velocity. A few well-hit balls on a warm day would skew the results.

Wind is relevant to the question, as well — and the Bay Area, with its large bodies of water, could be sending up wind currents that suppress batted balls. Prevailing wind numbers are often misleading, though. If you’ve ever talked to an engineer about accounting for wind when designing a big building, you’ll hear about how different the wind can be at different altitudes at the same site — and how much vegetation and surrounding structures can affect that wind. Surrounding structures like, say, a large block of luxury suites and bleachers.

Watch the tarp above the number 43 for evidence of wind currents.

Rather than use wind in our analysis, I thought it would make more sense to try and put everything but wind into our analysis and hope that the wind is the driver of change that we see in the data.

I wondered if we couldn’t use actual exit velocity, launch angle, altitude, and temperature to judge an “expected” distance for fly balls. If we could, we could judge that against actual distance and identify any patterns, if they exist, for Oakland’s park. The idea would be that the effect we see in the difference between expected and actual was from the main variable we left out — namely, wind.

Luckily for a dolt like me, we’ve got Alan Nathan, baseball-loving Physics Professor, in our midst. Nathan was so kind as to send 2015’s batted-ball and temperature numbers, adjusted by expected distance calculator.

Sean Dolinar then helped create a heat map for expected vs. actual distance. Note that there are two tabs in this tableau. In the first, Dolinar binned the balls by actual distance over 300, and the color is the median difference between that actual and expected in that spray angle. The second shows balls grouped by expected distance, but you’ll see that there’s a lot of negative there because Nathan optimized his calculator for well-hit balls (over 90 mph and in the 10-40 degree range).

There’s a lot of data represented there, and it can get confusing, so let’s pull Oakland’s file and annotate it a bit. Here’s Oakland with a picture of Mount Davis arranged behind the data to give it context. Just from looking at this, it looks like it’s possible that Mount Davis robs some of the bigger fly balls to left center of some distance.

Actual minus expected distance, binned by actual distance, with the boundaries of Mt. Davis represented by the dashed lines.

Look also at the bin at 400 feet in center field. The average ball hit to that bin went 37 feet less far than you’d expect, given launch angle, exit velocity, and temperature that day. That’s pretty important! At 400 feet, a fly ball is a 50/50 home-run ball, since that’s the distance of the wall in center. If it had those extra 30 feet, it would be a clear home run.

There’s a decent amount of negative bins in these heat maps, and no positive, really. That might be because we optimized it for home-run swings, or it might have to do with batted-ball spin, which was a variable we didn’t have in this exercise. But, judged by negative bins at 350 feet and beyond, there are few parks that play as strangely as the Coliseum. Oakland features seven negative bins at 350 feet and high; only Miami (7), Milwaukee (7), Kansas City (8) and Toronto (8) were as weird (or weirder) with balls in play.

And that’s the best takeaway here. Oakland is, indeed, weird on balls in play to the outfield. If you take a standard deviation of the difference between actual distance and the expected distance, given the launch angle, exit velocity, altitude and temperature that day, Oakland’s is the highest in baseball.

Standard Deviation of Actual Minus Expected Fly-Ball Distance by Park
Park StDev, Expected-Actual Distance
Oakland 27.3
Colorado 27.0
Cleveland 26.2
Texas 24.8
Milwaukee 24.7
Boston 24.6
Arizona 24.5
Kansas City 23.8
Philadelphia 23.7
Chicago (NL) 23.6
L.A. (NL) 23.5
Cincinnati 23.5
Chicago (A.L.) 23.3
Average 23.2
Expected distance calculated using altitude, temperature, launch angle, and exit velocity, optimized for exit velocity over 90mph and launch angle between 10 and 40 degrees.

Standard Deviation expressed in feet, grouped by home park.

It’s not such an outlier that it seems impossible, but there’s Oakland at the top. Oakland’s difference between actual and expected was higher than any other team’s home park (-10.2 feet, -5.8 feet was average), and it also featured the most variance from batted ball to batted ball.

For Jed Lowrie, who has about average career power, simple park factors could help him decide to employ a more grounder-friendly opposite-field approach. But if you look at his spray charts, you can tell why he was particularly interested in the air flow to the traditional power alleys in right and left center.

Source: FanGraphs

As for the the original question, the answer is “We don’t know.” Despite our best efforts, we can’t say for sure if Mt. Davis, specifically, is killing fly balls in Oakland. But! We can say that almost definitely something strange is going on with the wind patterns in Oakland, because it suppresses batted-ball distance further than you’d expect given the temperature, and it has wilder swings in distance than any other park. It could just be the wind, or it could be the wind’s interaction with the park, but it does look like something is going on here.

With a phone full of pictures of pitchers' fingers, strange beers, and his two toddler sons, Eno Sarris can be found at the ballpark or a brewery most days. Read him here, writing about the A's or Giants at The Athletic, or about beer at October. Follow him on Twitter @enosarris if you can handle the sandwiches and inanity.

Newest Most Voted
Inline Feedbacks
View all comments
7 years ago

Oakland is first in that last chart, but Colorado is second. Is that chart suggesting that Oakland suppresses flyballs as much as Colorado helps them? or is it suggesting that Colorado is the second most flyball suppressing park in the league? If it’s the former, that’s crazy! If it’s the latter, my first instinct would be to say the data is flawed somewhere…

7 years ago
Reply to  Eno Sarris

Okay. I’d assumed as much, but wanted to be sure. Still, Oakland as a Bizarro Mile-high, and with all that foul ground? That’s a tough park!

7 years ago
Reply to  Eno Sarris

The spread is biggest, or the spread is most variable?

You could have big spreads with small variations if the spread was consistent: e.g. if every fly ball in a park traveled exactly 40 feet less than expected the std. deviation of the spread would be 0 even though the size of the spread is still big.

Of course practically, bigger spreads will be more variable because it will give a wider range of spread.

7 years ago
Reply to  jianadaren

highest standard deviation of the difference = spread is most variable