STAT RELEVANCE REVISITED: PASSING=PENALTIES?

Four things can happen when you pass, according to Saurian Sagacity's latest look at the effect of penalties on winning and losing, and turns out the fourth element is also bad. Er, maybe...

Last winter, I found over and over again, in game after game, conference after conference, that penalty yardage had no correlation whatsoever with winning in games between two BCS conference teams, when every other stat category - even which team was at home and which team scored first - showed at least some non-trivial, positive correlation to victory. With penalties, in fact, the correlation was slightly negative: on a macro level, the twenty most penalized teams had a better aggregate record than the twenty least penalized teams, while on a game-by-game level, the team that was penalized more than its opponent had a slightly higher winning percentage.

Some mild gnashing of teeth occurred in the face of such counterintuitive data, or, in Lou Holtz's case, some impressive fits of slobber in defense of an ancient coaching point. Are more penalized teams more aggressive? Are refs subconsciously "leveling the field" by throwing more flags against better teams? Do teams that hold the ball longer on offense commit more penalties, since most penalties are called against the offense?

None of that explained much, as Mergz points out in detail. He persisted with the idea that the key to the answer lies with the offense, and when he looked specifically at penalty yardage of the top passing teams against penalty yardage accrued by the top rushing teams, he found an interesting result, if not a conclusive one:

In a nutshell, the more you pass, the more you're flagged.

That's not a foolproof summary - maybe heavily-penalized teams are passing because penalties have put them in a hole they have to throw out of, total yardage does not necessarily reflect how often a team throws or how efficient it is at passing - but it will do for a headline. Mergz also looked at the records of last year's top 20 passing offenses:

1. Hawaii 11-3
   
2. New Mexico St. 4-8
   
3. Texas Tech 8-5
   
4. Brigham Young 11-2
   
5. UTEP 5-7
   
6. Purdue 8-6
   
7. Louisville 12-1
   
8. Houston 10-4
   
9. Kentucky 8-5
   
10. Missouri 8-5
    
11. Baylor 4-8
    
12. Tennessee 9-4
    
13. Notre Dame 10-3
    
14. Southern California 11-2
    
15. Washington St. 6-6
    
16. Ball St. 5-7
    
17. California 10-3
    
18. LSU 11-2
    
19. South Carolina 8-5
    
20. Pittsburgh 6-6

Thirteen winning records out of twenty in the most penalized cross-section.

I can think of a few reasons passing might increase penalties: increased (or at least more obvious) instances of holding, the added possibilities of intentional grounding, illegal touching and ineligible men downfield and increased instances of false starts by linemen in a slightly less stable stance who must react quickly to prevent getting beaten around the end.

Problems: passing offense was not especially highly correlated to victory, and slightly less so than rushing. Of the top 20 in each category, there were 13 winning teams in both cases, but the top 20 rushing teams had a better aggregate record and a greater gap between the success of the teams at the top of the category and those at the bottom. If the most-penalized team was more likely to win last year (it was, according to the comprehensive sample, almost 60 percent of the time), that doesn't jibe with the same team being unusually passing-oriented, as the team with the most passing yards in any given game only won about 55-56 percent of the time, less often than the team that won time of possession or was first to score, and far less than the best rushing offenses.

Still, by showing an obvious correlation between passing and increased penalties, Mergz may have begun to crack the mystery, insofar as the solution is crackable and not just a random jumble with margins from team to team too small to draw any meaningful conclusions. That would be so bogus.