Park Effects in the SEC

Park Effects in the SEC

While reading Jim Albert’s new book Visualizing Baseball, I was motivated to follow his lead and apply some of his methods to women’s college softball. Albert looked at park effects in Major League Baseball. Park effects measure the contribution of a park to the offensive production of a team. In his book, Albert used home runs as his measurement.

Such an evaluation for softball is time consuming. Unlike for MLB where a large amount of data is readily available, doing this type of analysis on college softball is often a matter of manually parsing the NCAA’s data for each team. For that reason, I decided to narrow my scope to the conference games played in the SEC in 2017 for my example.

Here are the total number of home runs hit in each school’s park by the home and visiting teams last season.

The above graphic only tells half of the story. Afterall, a team whose park saw a large number of home runs might have had a power hitting team or a pitching staff that gave up home runs. To evaluate park effects, we also need to look at the performance of each team on the road.

By subtracting each team’s home and road performances, including the performances of their opponents, we can get a better idea of each team’s park effect. I converted the number of home and road home runs to a percentage of plate appearances in order to account for unbalanced schedules (some SEC teams played 24 conference games while others played 23), and then subtracted the two. Here are the park effects according to this measurement in the SEC for 2017.

You can see that the home parks of Auburn and Texas A&M were the most generous in allowing home runs last season, while Kentucky’s and Alabama’s were the stingiest. Some of the possible explanations for park effects are the dimensions of a field, wind, temperature, elevation, or the home team’s lineup being customized to the dimensions of their park (e.g., a short right field will result in a team full of power hitting left-handers).

Here are a few of the possible park effects for the teams in the SEC, as well as their overall park effect.

Team Elev. (ft) LF (ft) CF (ft) RF (ft) Park Effect
Auburn 597 200 220 200 1.3
Texas A&M 367 190 220 190 1.3
Missouri 528 190 220 190 1.0
Tennessee 886 190 220 190 0.7
Georgia 636 190 220 190 0.5
LSU 56 200 220 200 0.2
Mississippi St. 335 200 215 200 0.2
Ole Miss 505 200 200 200 -0.1
Arkansas 1,400 200 220 200 -0.7
South Carolina 292 200 220 200 -0.9
Florida 151 200 220 200 -1.0
Alabama 222 200 220 200 -1.1
Kentucky 978 200 220 200 -1.3

Running a regression analysis on these variables and others was inconclusive. Obviously I’d like more seasonal data. But to evaluate this further I thought I’d take a look at the two parks that limited the number of home runs the most (Kentucky and Alabama) and the two parks with the largest positive influence on home runs (Auburn and Texas A&M) to see if there was something there.

In a study on the atmospheric effect on the carry of a baseball, physicist Dr. Alan Nathan published the article Going Deep on Goin’ Deep. Dr. Nathan found that a 5 mph wind had a significant effect on the flight of a baseball.

Atmospheric Effect Change in Distance
10-degree increase in temperature  3.3 ft
1000 ft increase in elevation  5.9 ft
50% increase in relative humidity at 750 ft  0.9 ft
5.0 mph out-blowing wind 18.8 ft

With Dr. Nathan’s findings in mind, let’s eyeball these stadiums further, with particular focus on the stadiums at the extremes. In the images below, clockwise from upper left are the fields for the University of Kentucky, Alabama, Texas A&M, and Auburn as found on Google Maps.

The fields for Alabama (upper right) and Auburn (lower left) face virtually the same direction and since they share a state and are 150 miles apart, it’s unlikely that there were significant atmospheric differences. A comparison of the two cities on seems to confirm this, with Tuscaloosa typically slightly more humid in the spring but Auburn often just a little windier.

The differences between the stadiums shown above in Kentucky (upper left) and Texas A&M (lower right) on are more interesting. To start, College Station, TX is considerably warmer than Lexington, KY in the spring.

In addition, College Station is typically more humid in the spring.

And finally, these stadiums do experience some wind.

Let’s take a look at some specific dates from last season to see if I’m on to something. Using Weather Underground, first let’s evaluate the wind at Texas A&M for the home dates where multiple home runs were hit.

Date Opponent Total HRs Wind (mph) Max Wind (mph) Direction
3/10/2017 Mississippi St. 2 5 17 S
3/24/2017 Missouri 3 17 24 SSE
3/26/2017 Missouri 2 8 21 SE
4/21/2017 Georgia 3 10 17 S
5/7/2017 Tennessee 4 6 14 SE

It looks like we have something here! Combine the speed and direction of the wind with the orientation of the Aggie Softball Complex and we can likely see why 14 of the 18 home runs hit in College Station last year were on these five dates. Any wind from the south is going to encourage home runs.

In contrast, in 12 home games last season only once did Kentucky have a game where multiple home runs were hit. Overall, only eight home runs were hit at John Cropp Stadium. Here are the seven home games in Lexington, KY where no home runs were hit.

Date Opponent Total HRs Wind (mph) Max Wind (mph) Direction
3/25/2017 Arkansas 0 11 24 SSE
4/7/2017 Tennessee 0 11 22 WNW
4/14/2017 Florida 0 5 9 SSE
4/15/2017 Florida 0 10 23 SSW
4/16/2017 Florida 0 11 28 SW
4/28/2017 Texas A&M 0 8 17 SSE
4/29/2017 Texas A&M 0 9 20 SSE

Any wind out of the southeast in Lexington should limit home runs, and on four of the above dates the wind had that element. The game that stands out for wind direction is the Tennessee game on April 17 when the wind was blowing out from the west to northwest. Though no home runs were hit, Kentucky lost that game 9-0 in just 5 innings. It seems like we could label that game with no home runs as an outlier.

Moving on to Alabama and Auburn, things get murkier. First, let’s take a look at Alabama where home runs were few and far between.  On seven home dates in Tuscaloosa no home runs were hit.

Date Opponent Total HRs Wind (mph) Max Wind (mph) Direction
3/10/2017 Arkansas 0 9 24 NNW
3/12/2017 Arkansas 0 9 17 NNW
3/26/2017 South Carolina 0 6 15 S
4/7/2017 LSU 0 6 17 NW
4/8/2017 LSU 0 2 9 SSE
4/9/2017 LSU 0 5 15 S
5/6/2017 Auburn 0 7 20 WSW

Any wind out of the north should limit home runs at Alabama while winds from the south should encourage them. So why weren’t there any home runs on the four dates that had a wind with something of a southerly direction? On 03/26 there were 15 total hits between the two teams but no home runs, the LSU series matched two of the teams at the bottom of the conference in home runs, and on 05/06 there were 16 hits but no home runs. Hard to find a pattern.

Auburn isn’t any easier to understand. 13 of the 17 home runs hit at Jane B. Moore Stadium were on four dates. Nine of the 13 home runs were to left field, which were at least partially into the wind.

Date Opponent Total HRs Wind (mph) Max Wind (mph) Direction
3/10/2017 LSU 2 11 18 WNW
4/8/2017 Ole Miss 3 9 16 N
4/14/2017 Missouri 4 13 20 NW
4/29/2017 Mississippi St. 4 10 16 SSE

The level of competition could play a role in these results. And some of these results could be associated with randomness, which is significant in women’s college softball. Obviously a larger sample size and more variables would help.

Park effects deserve more study in college softball. While this type of research is enjoyable for me, unless the NCAA changes the way it manages and provides data, studies like this will remain a time intensive process.