Offense sells tickets. Defense wins championships. -Bear Bryant
Many people have taken a shot at trying to prove or disprove the second part of the Bear's adage about defense winning championships. I myself took a shot at explaining some element of it in 2009, though many others have taken more rigorous looks at the issue. The general consensus is that defense might be slightly more important than offense, but you're better off doing well at both.
What I haven't seen is any kind of real attempt to test the validity of the first half of the phrase regarding the ability of offense to sell tickets. That's what I'm going to try to address here on a fairly basic level.
To that end, I looked at average capacity data from the NCAA. That doesn't reflect actual ticket sales, but most schools when announcing attendance count sold tickets as redeemed tickets. Attendance inflation in that way is fairly common. Either way, average capacity smooths out the differences in stadium size around the country. I also capped it at 100%, moving down the figures for all the schools who taunt the fire marshall by reporting greater than capacity crowds.
I also focused on yards per game and points per game rather than the more advanced statistical measures created in recent years. While those stats are great, most people who attend games have never heard of them. I also didn't look solely at home stats without road stats because I don't think anyone looks at that split when deciding whether or not to purchase tickets. Plus, season ticket holders make up a significant percentage of attendance, and they purchase all of their tickets up front without even seeing that year's team.
Some programs' attendance never changes either because they're at capacity every year (Nebraska, Florida, etc.) or because they have low attendance every year (UAB, Tulane, etc.). That means there won't be an perfect link between offense and attendance, but there are seldom perfect links between any two things in any part of football.
Are offense and attendance connected?
The records from the NCAA goes back to 2000, providing 11 seasons' worth of data. I ran correlations and regressions on the stats I picked out. The element of the regression I was after was R-squared, which happens to be the correlation squared. In this case, it tells what percentage of the variance of attendance figures is explained by the variance in the offensive categories.
I found that yards per game has a correlation of .267 with percent capacity, while points per game has a .370 correlation with percent capacity. Those are nice positive correlations, indicating that good offense tends to coincide with higher attendance. The R-squared for yardage is 0.071, indicating that 7.1% of the variance in attendance is explained by the variance in yards gained per game. For points the R-squared is 0.137, indicating that 13.7% of the variance in attendance is explained by the variance in points scored per game.
It does appear that there is a mild connection between offense and attendance. I checked to see if there was a stronger connection between offense of one season and attendance the next, just in case people base their ticket buying decisions on what happened the year before. Each of the correlations and R-squares were about 0.01 less than those for the current year as listed above, so that answer is no overall. It would be more interesting to take a look at offense of one year and season ticket sales of the next, but the latter data is not available en masse.
Intuitively, it would make sense for winning to be more of a driver for attendance than offense. Watching a good offense is fun, but winning conquers all, right?
Based on the data I have, that is correct. Winning percentage had a correlation of 0.507, and the R-squared was 0.257. So, roughly a quarter of the variance in attendance is explained by the variance in winning percentage for college football as a whole. Other factors that no doubt play into changes in attendance include history, tradition, coaching changes (or the lack thereof), marketing, opponent quality, and overall economic conditions.
So while offense does appear to have a small positive effect in ticket buying, creating a kernel of truth for the first half of Bryant's maxim, it gets overwhelmed by winning. That's not all, though.
What about defense?
It's only fair to do the same analysis for defense too. Does defense happen to sell tickets as well?
Based on the same measures, it does. It does to a greater degree, in fact. The correlation for yards per game and average capacityis -0.410, and the correlation for points per game is -0.502. The R-squares for those are 0.168 and 0.252. All four figures are stronger/greater than those for offense. The same holds true when looking at the effects of one year's stats on the next year's attendance (though defense sees the same slight decrease in correlation and R-squared that offense does).
What interests me the most is that the defense figures are nearly identical to those for winning percentage. Intrigued, I tried yardage and point per game margins. Yards per game margin is not as strong as winning (0.472 correlation, 0.223 R-squared), but points per game margin is (0.529 correlation, 0.280 R-squared).
That there's little difference in the points margin and the winning percentage correlations should not be surprising, as the two of them track very closely with each other (which is why Pythagorean expectation works, by the way). That the point margin correlation was slightly higher might indicate on some level that how the team plays independent of winning rate might be ever so slightly more important to attendance than the team's record itself.
To what degree offense sells tickets for your particular team differs from the degree it does for other teams. However for college football as a whole, this much is true in general: offense sells tickets, defense sells more, and winning sells tickets a bit more than that. While the Bear wasn't completely wrong in the first half of his statement, he wasn't completely right with it either when it comes to today's college football environment. Maybe the correlations were different decades ago when he roamed the sidelines, but this is how it is today.