Photo by John Korduner/Icon Sportswire via Getty Images
Have Iowa fans been right all along?
I just published this over at my gambling website, because if there’s anything I love, it’s risk-taking, degeneracy, football, and statistics. (I began my academic life hustling with the nerds at Vandy trying to scrape out a PhD in psychometric statistics. Turns out you still have to see clients, and as I hate people, I dipped. But the stats love never went anywhere).
Anyway, the analysis is too interesting to not share with you folks. Call it “punting is winning” or “vindication of percentage football,” but for an average team, punting matters. Not as much as it does in the NFL, and not for CFB’s most talented teams, but it is nevertheless real.
Inspired by a discussion that I saw earlier today, I decided to try and crank some preliminary data for a new metric.
Punt yard differential field position.
The way that this works is that most punts land between the 15 and 20 yard line, both in college and pro. Almost 3/4 hit between that spot. What some data nerds found was that for every yard differential of field position one team had in punting versus another, they gained almost .4 points.
I have a punter that’s kicking from my 35, and it lands at your 19. You punt from your 35 and it lands at my 18 — Your net field position is one total yard gained from swapping punts. Therefore you gain .4 points for that one yard of starting field position differential.
In the NFL, it is very powerful. However its power seems to be most predictive between the receiving team’s 10-25 yard line. In other words, it’s not just the yard differential, it’s also where that differential is occurring.
So, using that as an initial parameters, I went through 50 punts each in the SEC and Big 10 and crossfit it with the final results as part of the composite model. (And we have had a lot of variability this year, about 32% is unexplained mathematically as to why a particular score happened, which is significantly higher than usual.)
I assessed the relationship between punts landing at the opponent’s 10 to 25 yard line, and then yard differential (as was done in the NFL). But it is not generating .4 points per differential yard — the effect is far smaller in CFB, about .18 points per yard. But it certainly did correlate with some of the missing margin in the final score, and it is statistically significant at p=.05.
Why are the smaller result than the NFL?
I suspect a lot has to do with the inherent difference in college ball, the huge disparity in caliber of opponents, and just the overall difference in schemes.
One of the things that makes college football unique is if you want to see a 3-3-5 one side of the field, and a Buddy Ryan 46 on the other sideline, and an air raid by one team, and a triple option by the other, it’s pretty easy to: just change the channel until you see what you’re looking for.
But in the NFL, schemes are far more homogenized, the talent disparity is minuscule. What separates an all pro corner from the guy bagging groceries could be 2/10 of a second, or the few hundredths of a second reaction time lost when he buys a head fake.
It is what we see with political movements as well, it’s called the fallacy of exaggerated difference. It’s why a Bernie Sanders bro may look at a Joe Biden voter and see a literal Nazi. Just as in politics, in the NFL, small differences can become multiplicative. And when offenses are getting inside the 10 to 25 yard line, you pretty much know what play is gonna be ran, how they’re going to manage the clock, you have hundreds if not thousands of snaps of tendencies with the same core group of players and coaches.
Every yard in the NFL is desperately important because of that lack of disparity.
HOWEVER, just as in college, when there is a tremendous talent gap or a tremendous schematic advantage, this punting yardage differential evaporates for both college and pro. The last two years, for instance, Buffalo and Kansas City had the worst punt differential under this equation. One won the Super Bowl, the other led the league in scoring margin for/against.
That is why punting may not be important to teams like Ohio State or Georgia or Alabama as a result, but when Iowa Man tells you that punting is winning, when Aggie praises their kicking game, and you see the razor-thin margins by which the Hawkeyes or A&M are winning and losing games, know that analytically yes, they likely are notching Ws because of superior punting yard differential; and that differential is translating in a real way on the scoreboard at the end of the day.
It is too late this year to incorporate it into the data set, and I certainly want to do some more exploration and fine-tune it, but next year I certainly will have some variable of punt yard differential x field position…at least in one model.
Interesting stuff. Thought I would share. Let me know your thoughts below.