Forum : SEC Stats: The Fife Factor, Week Eleven

This is statistical nonsense.  You should not try to cram all these things together this way.  Having defensive plays in the divisor when offensive outcomes are in the numerator is incoherent.  Florida gets cheated in this measure.  Their number of defensive plays is naturally low, because their defense is awesome.  A team like Arkansas has many more defensive plays, and that number goes into the divisor & helps them in this calculation.  They're actually worse than they appear.

Offense and defense need to be divided.  You'd have to work hard to get meaningful data.  To begin with, you need to separate offensive penalties from defense and special teams.  Further, if you're looking at total blunders, all fumbles are blunders even if they aren't recovered.  Lost-yardage plays are blunders for an offense.  For a defense, busts are blunders.  The number of 20+ yard gainers allowed is a good proxy for how good a defense is.

Finally, you need some sort of risk-reward adjustment.  If the offense is making big yards and scoring touchdowns, it's doing fine even if the risk factor is elevated.  Arkansas, with 7 yards per offensive play, is obviously taking more risk than LSU, with 5.2 yards per play.  Which offense would you rather have?