Update #2: College Football Performance and Proximity to Home

How do you measure football performance? How can you compare Ndamukong Suh, a defensive tackle that played for Nebraska from 2005 to 2009 with Cooper Rush, Central Michigan’s quarterback from 2013 to 2016? These questions are incredibly difficult to answer, especially given the nature of the sport. It is impossible, at this point, to accurately and fairly quantify every player’s contribution to the result of a play, much less a game. Advanced statistics in football is always improving and decoding the game, but it is still far behind simpler sports like baseball and basketball.

My solution is simple: I will not try to compare every single player. I will try to compare every quarterback. Authorities in the field agree with the common sense that quarterback is the position with the largest effect on the outcome of plays. A straightforward way to check this is to look at the highest-paid players in the NFL. The 15 highest-paid players and 19 of the top 25 are quarterbacks. The quarterback is the most important player on the field. As a result, game statistics tell you more about how a quarterback plays than anyone else (except kickers and punters, but recruiting data is too sparse for kickers and punters to use them). Basically, the quarterback has the most effect on plays, and the statistics reflect the way the quarterback affected those plays. Obviously, the quarterback is not solely responsible—just the most responsible.

After consulting with various prominent football advanced statistics blogs, I have chosen to use Adjusted Yards per Attempt and Rushing Yards per Game as the statistics with which to compare quarterbacks. Adjusted Yards per Attempt is calculated as
(Passing Yards + (Touchdowns * 20) – (Interceptions * 45)) / Passing Attempts

It is regarded as the best simple statistic for quarterback comparison because it is fair to all offensive systems for evaluating passing efficiency. It is very similar to the way Passing Efficiency is calculated, but it has been tweaked for modern football while Passing Efficiency was developed in 1979. I chose to use Rushing Yards per Game because especially in modern college football, the quarterback is often more than just a passer. This added value must be considered because it is a defining characteristic of many of the recent Heisman winners from Tim Tebow to Lamar Jackson.

Using Adjusted Yards per Attempt and Rushing Yards per Game gives an accurate depiction of a quarterback’s performance, but I also want to consider the total value of a quarterback to his football program. For example, Cam Newton delivered a perfect season and a national championship in his only season at Auburn. But he was only there for one season. The following two seasons Auburn faltered to 8-5 and 3-9 records, although the next year they returned to the national championship game. The point is that longevity is also an important thing to consider for college football players. To that end, I will also use Adjusted Yards per Attempt multiplied by game starts as a metric for quarterbacks.

Tomorrow I will post my data analysis results, as well as my databases and code—open source is cool and all.


  1. Samuel McIntyre says:

    This is a very cool project that challenged my expectations of what research should be. It seems like you’ve done a nice job adjusting to the challenges of your research, and I’m interested to see what your results show. If there is truth to the old wisdom about staying close to home, I’d be curious to know what factors contributed to this result. Additionally, I wonder if there would be a significant difference when these quarterbacks were forced to travel for away games vs. when they played at home. Good luck with the final component of your project!