clock menu more-arrow no yes mobile

Filed under:

Tampa Bay Rays Weekly Attendance Trends over Time

As spring training draws to a close, it's time to start turning our gaze towards the regular season and what lies ahead for the Rays this year.  One issue that we've already talked about and is sure to garner attention as the year goes on is attendance.  As the season goes along, we're all going to want to know how well the Rays are drawing, but how do we judge that as the season progresses?  If April finishes and we find that the Rays have drawn approximately 350,000 fans to the park, is that a good thing or a bad thing?  If we see that we drew 30,000 fans to a game against the Yankees, is that higher or lower than we should be hoping for?  It's very easy to look back in retrospect and say, oh, we drew really well this year, but it's much tougher to make this evaluation as the season is still in motion.  And along those lines, how do we define what a "good" average attendance figure for the Rays is?  Is it 25,000 or is that way too low? 

Over the first month of the season I'm going to attempt to answer some of these questions, starting by fleshing out the different individual variables that affect attendance figures.  If we can understand approximately how well or poorly the Rays will most likely draw in certain situations, we can begin to make predictions and judgements based on games that have already been played this year.  With that said, let's start by delving into how the specific day of the week influences the Rays' attendance figures.


Actually, before we get going, a quick word on the data that I'm using.  Basically, I've gone through and compiled the attendance data for every single Rays game ever played into an Excel spreadsheet.  This gives me a lot of data to look back on and to draw conclusions from, but it still comes with a lot of caveats.  Attendance figures are tricky to manipulate because you're constantly dealing with issues of small sample size.  For instance, if I want to go back and look at how well the Rays did last year against the Red Sox on certain days of the week, I have a total of 9 home games to look at - two of which were on a Monday, two on a Tuesday, two on a Wednesday, and then one each on a Friday, Saturday, and Sunday.  Until I get on the computers at my school and plug this data into some fancy stats programs, it's very tough to separate the effects of the different variables that might influence attendance from each other.

To get around that, I decided that in order to increase the pool of data I was drawing from, I'd base this study off of every single Rays game ever played.  Since I want to determine how attendance varies by day of the week, I started by totaling the number of games played each year on Mondays, Tuesdays, Wednesdays, etc and then figured the attendance per game for each of these instances.  I then compared those attendance per game figures for Mondays, Tuesdays, etc. to the average attendance per game for that year, which showed me a percentage difference between the attendance that year on a Monday compared to the average attendance for the whole year.  For example, here's my chart for Mondays:

Games Attendance Attend/Game Percentage
2008 9 170,561 18951.2 -15.3%
2007 8 100,284 12535.5 -26.8%
2006 10 168759 16875.9 -0.1%
2005 9 131224 14580.4 3.4%
2004 5 100725 20145.0 26.4%
2003 5 75224 15044.8 15.1%
2002 6 64251 10708.5 -18.6%
2001 3 32743 10914.3 -31.9%
2000 6 113428 18904.7 4.3%
1999 10 169187 16918.7 -12.3%
1998 11 290794 26435.8 -14.6%
82 Total = -7.07%

In order to get my total percent difference on a Monday, I took a weighted mean of each year, simply dividing the number of games played each year by the total for all years, multiplying those numbers by the percentage figures I had per year, and then adding them all together.  This way, I'd give more weight to the years that I had more data, instead of simply taking an average of every year.

As you can see, there's lots of variablity between years.  Some years the Rays will draw exceptionally poorly on Mondays, while on others, like 2004, the Rays actually drew much better on Mondays than their average for the year.  This variability is where you can see the pitfalls of small sample sizes and the other influences of other variables.  For instance, in 2004 there were only 5 games played on a Monday, three of which were against the Red Sox and Yankees.  That's why I took a weighted mean, though - to try and eliminate as much of this statistical noise as possible so instances like those 3 games wouldn't end up skewing my overall results.

So without further ado, here are the final results per day:

Mondays -7.07%
Tuesdays -7.42%
Wednesdays -10.46%
Thursdays -11.02%
Fridays -1.46%
Saturdays 27.73%
Sundays 14.29%

While this is basically what you'd expect through common sense, I was definitely surprised that the Rays draw below average on Fridays.  I would have thought that'd be a day that would be typically above average as well, even if it wasn't as high as Saturday and Sunday games.  And while these figures are pretty cool, sadly they don't have much predictive value when taken by themselves.  As you saw, these rates vary wildly from year to year depending upon other variables as well, so we can't take a 25,000 Monday night attendance figure and then simply chop 7% off of it to determine what our average attendance will be at the end of the season.

What this data can give us, though, is ballpark figures to keep in mind when evaluating attandance figures that we see this season.  Say the Rays face the Blue Jays on a Thursday night and attendance only tops out at 15,000.  Well, knowing that Thursdays are historically our worst days for attendance and the Blue Jays are not a large draw, it's not something to get worried about  In that vein, though, next I hope to look at how our attendance varies facing different teams, specifically the Red Sox, Yankees, Blue Jays, Orioles, Marlins, and anyone else you guys suggest that might be interesting to see.  With that data and with the rates per day of the week, we should get ourselves in a place to begin making some rough predictions.