Future Shock: Draft Cl... (06/06)
88 comments have been left for this article. BP Comment Quick Links
| |
|
|
June 7, 2009 Prospectus Idol EntryA Tale of Two Teams: The Attendance GameThe most beautiful thing in the world is a ballpark filled with people. - Bill Veeck As the scoreboard programmer for the Trenton Thunder in their inaugural season (1994), one of my daily tasks was to enter in player averages from the compiled statistics faxed to us from the Howe News Bureau. Before I did this, there was one statistic that my fellow press-box compatriots and I craved above all us: the attendance standings. Every day, we wanted to see if we were beating the other two Eastern League franchises with new stadiums: the Portland Sea Dogs and the Bowie Baysox. The memory of these daily quests for the attendance standings recently got me thinking about the drivers of baseball attendance. Currently, bizofbaseball.com (with assistance from Minor League Baseball) offers an attendance database that goes back to 2005. Minor league attendance data broken out by team prior to 2005 are spotty. I decided to look at the attendance figures for all AAA and AA teams to determine if there is some discernible pattern to attendance. I selected AAA and AA teams since they are usually located in mid-size cities with their own market, but far enough away from Major League cities so there is little competition for the fan's entertainment dollars. After examining the attendance figures of this subset, two teams pop out immediately: the Sacramento River Cats and the West Tennessee Diamond Jaxx, the best drawing and worst drawing teams, respectively.
Team League Avg. Attendance per
Opening (2005-2008)
Sacramento PCL(AAA) 10,120
West Tennessee Southern (AA) 1,699
The Sacramento River Cats, who play at Raley Field, are held up as the model minor-league franchise. They have the highest attendance four years running and as a result Forbes has named them the minor league team with the highest net worth (around $29.8M). However, the recession may be hitting the River Cats hard as their average attendance is down 12% through their first 24 games of 2009. On the other hand, the West Tennessee Diamond Jaxx, who play at Pringles Park in Jackson, Tennessee, have been plagued with financial problems and low attendance for many years. They were sold in 2007 to a group of Nashville investors who are still planning to keep the team in Jackson. In 2008, they drew an average of 2,096 fans compared to the previous three years where they averaged just 1,567 fans, a 34% bump in attendance. However, there numbers are still roughly 20% of Sacramento. A discrepancy this large begs further analysis. Modeling Minor League Attendance The first likely driver to consider is simply population. For the 63 ballparks that were used for AAA and AA teams from 2005 to 2008 (there were two team moves, and one new stadium in this time horizon), I found their latitude and longitude from Wikipedia. With the help of Tacitician in Andover, Massachusetts, I looked at the population within concentric circles of 5, 10, 20, 50 and 100 miles from each stadium. The table below highlights the significant difference between Sacramento and West Tennessee:
Population (000s) within
Ballpark 5 miles 10 miles 20 miles 50 miles
Raley Field (Sacramento) 258 760 1,674 3,603
Pringles Park (West Tennessee) 49 89 154 489
At every level, the population around Raley Field is somewhere between five to ten times as much as Pringles Park. But these are just the top and bottom teams. To further understand the relationship, I ran a linear regression with all the ballparks to determine a likely baseline attendance model. I removed the seven franchises (Trenton Thunder, Pawtucket Red Sox, Lehigh Valley Iron Pigs, Frisco Rough Riders, Reading Phillies, Tacoma Raniers, and Bowie Baysox) whose stadiums are within 50 miles of a Major League stadium for two reasons:
Running a regression, I found the following equation which gave a good, if not great fit (R2 = .399), i.e., roughly 40% in the variability of attendance figures could likely be explained by the underlying population differences. The regression equation was: Average Attendance per Gate Opening = .0069 X Population within 10 miles + .0017 X Population between 10 - 50 miles. Specifically, the average attendance per opening at a ballpark was roughly 0.69% of the population within 10 miles of the ballpark and 0.17% of the population between 10-50 miles from the ballpark. In round numbers this essentially means that 7 out of 1000 people within 10 miles of the ballpark and 2 out of 1000 people between 10 - 50 miles decide to attend a game on any given night. Plugging these numbers back into Sacramento and West Tennessee, this simple model would predict that:
Team Actual Avg. Predicted
Attendance Attendance
Sacramento 10,120 10,109
West Tennessee 1,699 1,302
These results highlight the three word mantra of real estate: Location, Location, Location. We may need to update the script for The Field of Dreams to have the haunting voice say "If you build it, they will come....assuming that you put it in a geographically viable location." Ballpark Novelty But is population the whole story? It has been shown at the Major League level that ballpark age plays a factor in attendance (Sport Management Review, May 2005). A similar phenomenon happens in the Minor Leagues as well. I grouped the minor league ballparks into four buckets: New (5 years old or younger), Young (6- 10 years old), Mature (11 - 20 yeas) and Old (20+ years old).
Ballpark Age Group Ratio of Actual Attendance
/ Predicted Attendance
New (< 5 years) 191%
Young (6 - 10 years) 160%
Mature (11 - 20 years) 101%
Old (20+ years) 88%
The data suggest that there is a honeymoon period of ten years where a franchise can expect better than predicted attendance (based on our baseline population model), Around ten years old, however, the attendance typically falls back to numbers that align with the predicted population model. It is interesting to note that Raley Field (Sacramento) is now nine years old. As it starts heading into mature status, maybe some of the 12% drop this year is not just the recession, but the fact that they are no longer the new thing in town. At this point, an interesting question to ask is which teams are beating the predicted performance of attendance and is there a common thread between these teams that suggests another factor to consider? The tables below show the top 5 teams (based on ratio of actual attendance to predicted attendance) and the bottom 5 teams. It should be noted that the predicted attendance column considers both the population AND the age of the facility,.
Team Actual Avg. Predicted
Attendance Attendance Ratio Parent Club
Midland Rock Hounds 3,991 1,862 214% Athletics
Portland Sea Dogs 6,344 3,071 207% Red Sox
Iowa Cubs 7,804 3,354 190% Cubs
Altoona Curve 5,382 2,860 188% Pirates
Springfield Cardinals 7,158 3,996 179% Cardinals
Team Actual Avg. Predicted
Attendance Attendance Ratio Parent Club
Connecticut Defenders 2,830 7,927 36% Giants
Ottawa Lynx 2,078 4,854 43% Phillies
Las Vegas 51s 5,041 9,897 51% Dodgers
San Antonio Missions 4,118 8,027 51% Padres
Portland Beavers 5,491 10,361 53% Padres
Proximity to Parent Club Looking at these two lists, we see a potential common link. For the top-five teams, all but one is located relatively close to the parent club (except for the Midland Rock Hounds), while the bottom five have the opposite problem. These are all teams (except for the Las Vegas 51s) whose parent club is relatively far away, and as a result the potential fans may not have as much of a desire to see the players on the minor league team, since they root for a different parent club. As a next step, I flagged a team as being close to its parent ball club if it met one of the following criteria:
Roughly half of the ballparks housed teams that were affiliated with a close parent club. For those teams where the affiliation changed between 2005 and 2008, I categorized each unique combination of minor league team and major league affiliation separately. The ratios of actual attendance to predicted attendance for the two sets of teams are in the table below: Parent Club Proximity Average Ratio Close 117% Distant 84% So on average, the minor league franchises that were close to the parent club (essentially in the parent club's market) saw attendance 17% over what our model would predict, while those distant from the parent club saw their attendance 84% of what our model would predict. This suggests that for a minor league team to be close to its parent club could have a 33% swing (17% above prediction versus 16% below prediction) Takeaways Based on our analyses, there are four key takeaways
Acknowledgements I want to thank two sources for information that helped in this article:
Tim Kniker is an author of Baseball Prospectus.
|
Tim really nails this piece. Unlike last week where he lost me (and I think a lot of people) out of the box, this one has all the math chops, but is structured in such a way that while I can follow along, I can also ignore things like "population bands" and ratios and still come away learning a lot. It leaves me wanting Part II, where he tells me how much floppy hat day can change things, whether winning matters, and why Memphis is in the Pacific Coast League.