Offensive efficiency

So, in the world of football, offensive efficiency has been a tough nut to crack.  Obviously, the objective of the game is crystal clear...is a team outscoring its opponents.  When evaluating one side of the ball, is the team contributing to winning games.

We can always look at scoring to get this answer.  (And, I note, we should take out defensive and special teams touchdowns when we do it.)

The next measure also seems pretty clear....yards per play is a key measure of how efficient an offense is.  Teams that gains yards should score.  When they don't...or when they score without getting yards, or out of proportion, then we might start to wonder why?  Turnovers, 3rd down plays, poor red zone execution---these could all be potential culprits.

So, people are trying to figure this out.

For example, the New York Times evaluated Oregon's offense on a yards per minute of possession metric.

Phil Steele has been weighing in over the past week on two other possible measures...points per play and yards per play.

Note:  I do know that the football outsiders probably have the best measure...they evaluate each play and the progress made toward a first down, and develop efficiency that way.  It requires play level datasets, however, which I don't have, so the search is on for meaning in aggregated data.

Also:  this is so much easier in basketball, where the Four Factors have kind of ended the debate.  I'm looking for something on that order here.

One of the risks we run when we have data is the irresistible urge to start dividing shit into other shit and then stroking our chin and looking thoughtfully at the ratios that result.  We may or may not have learned anything.

This is a long standing interest of mine, so I'm going to take some time during the off-season to dig into it.

So, we start.  I ran the MAC stats for the measures proposed above, just to see what we came up with.  Stats are MAC-only games, time of possession is rounded down in all cases (ie, seconds are eliminated) and we are using (contrary to my statement above) all scoring.

Here are the rankings for each team.

So, a few thoughts.

The reliability measure shows the number of places, in total, the rankings are.  So, on points per minute, the rankings are across 13 teams are off only six....and for points per play, there is only a difference of 2.

Conclusion:  There's not much to be gained by dividing points into other stuff.

We do see some differences on yards per point, and yards per minute.

Yards per minute is a measure I am not very enthusiastic about.  It might show what Oregon do with the cardio-offense, but it is only one way to win a game.  You can just as easily control the ball, score the same amount, and actually have a better shot at winning.

Next, a quick comparison how of teams rank in points scored vs. yards per play....who overachieved....who underachieved.