When Economists Talk, Pulaski Academy Listens

A few years back, the economist David Romer wrote an academic paper suggesting that teams in the N.F.L. should punt less frequently. While the league’s coaches mostly dismissed his findings, I suspect that teams are a bit more likely to go for it on fourth down in marginal situations today than they were before he wrote the article.

N.F.L. coaches may have laughed at Romer, but Kevin Kelley took note. Kelley is the football coach at Pulaski Academy, an Arkansas prep school, and his team no longer punts except to avoid running up the score. As Gregg Easterbrook reports on ESPN.com:

Kelley says that when he began to shun the punt, people thought he was crazy: “It’s like brainwashing, people believe you are required to punt.” Players and the home crowd needed to get acclimated to it. “When we first started going on every fourth down,” he says, “our home crowd would boo and the players would be distressed. You need to become accustomed to the philosophy and buy into the idea. Now our crowd and our players expect us to go for it, and get excited when no punting team comes onto the field. When my 10-year-old son sees N.F.L. teams punting on short yardage on television, he gets upset because he’s grown up with the idea that punting is usually bad.”

So far, it seems to be working pretty well. The Pulaski Bruins finished the regular season at 9-2-1 this year.

(Hat tip: Blake Howard)


Mike

When I played Pop Warner as a kid, our PAT rules were reversed - 2 pts for a kick, and 1 pt for a conversion. This was mainly because most elementary school and jr. high kids couldn't kick field goals from any distance. My team had a decent kicker, who was accurate maybe a little over half the time. So yeah, in our case, going for 2 was worth it.

Andrew

Rolub, that's exactly what I was going to say. Belichick seems to go for it quite often in situations where other coaches would not.

James

What about going for the 2-point conversion as opposed to the PAT as the default strategy? Even if a team converts the PAT %100 of the time, any 2-point conversion rate over %50 would make it the obviously superior choice in all but the most select situations. As the state of football conventional "wisdom" currently exists, the opposite is true. Of course, one of the fundamental ideas of Freakonomics is that "The conventional wisdom is often wrong".

David

James (re: 2-point conversion),
You have to consider the marginal value of a point. At the end of a game, the one point you score that's greater than your opponent has 100% of the value and all additional points are meaningless (except to gamblers). Likewise, failure to score that 1 point can have dire implications.

Over the course of a season (total points scored), your suggestion would be accurate, but it would not be an appropriate strategy for individual game management.

Note: I wonder my statement about would not hold true for a team like Texas Tech, who scores lots of points but lacks a stout defense. They have high scoring games, where lots of touchdowns allow for lots of 2-point attempts. This also depends on other factors (e.g., their red zone TD%), but might be an instance where hoping for 50% could work.

Doesn't Matter

Regarding comment 1, the idea of more than 50% success only holds if the value of getting 2 points is always identical to the value of gaining 1 point twice. That is, even ignoring risk (which few coaches do) a 100% probability PAT will have a team better far better off than a 50% probability conversion, most notably late in the game (or a perceived late occasion for scoring) and when tied, trailing by one, three, or four (to set up the win, the tie, the FG win or the FG tie). Since so many situations in football do fall under this set of circumstances, it makes sense that the PAT is still the standard.

However, I do believe that use of the 2-point conversion attempt is probably still below equilibrium.

rolub

It appears Bill Belichick (who Easterbrook is not the biggest fan of) may be taking notice to this as well.

schadenfreude

But how could you determine the marginal value of a point in the first quarter? Basically, you can't. The best strategy is to maximize your expected points and that is to go for the two point every time. The only exception is late in the game in a situation where the second point would be meaningless as there is only time for a single possession for the other team.

In the NFL, teams make the two point conversion sixty percent of the time.

ponderingfool

Regarding the PAT, you also have to factor in the fact whether the defense can return an interception/fumble for points as well (different rules apply at different levels of the game).

Shan

It would be interesting to see a study like this involving Texas Tech... last time I watched one of their games in its entirety, the only time they punted was when they were 4th & 44.

Will

It seems to me the coach is ignoring the average-marginal distinction. Although the teams have, on average, a 75% chance of converting on 4th downs, this would probably change if punting became rare.

luke

in the canadian football league, there are only three downs, but the players on different teams line up further appart (the defence has to be a full meter from the ball). as a result, the offence has a better shot at converting to a 1st down with short yardage.

anyhow, i'm with james on the 2 point conversion, mostly. the odd way football is scored means that single points don't always matter. 2 TD with two-point conversions still looses to 2 TD with one-pointers and a field goal

this suggests a research topic from prior data: find out how often any game is won by 1 point, or even better, check if the margin of victory is within some 2-p.c. factor... 2*(% of completing a 2 point conversion) - 1. could apply this to either the loosing team, or to both teams, and also check to see if it would force overtime, or if it would have prevented an overtime. (so really the win doesn't matter, just the score at the end of the 4th quarter).

Read more...

Arjun

ha, awesome. Great post and brilliant paper.

sqjtaipei

Great post. I think there are some differences for the various levels of football. For a couple seasons, my Dad's HS football team (my dad was a coach) didn't have a decent kicker... and had an above average defense. So... they kicked onside kicks after EVERY score! It was brilliant. They were able to save a lot of time by not working on long kickoffs and put that time into working on the onside kicks. They recovered over 50%.
In the NFL you'll always have a good kicker and you kick from so far back that the penalty for not recovering is probably at least a FG for the opposition.

stuart

In response to the person who argues for the 2-point conversion, Pulaski Academy goes for two often. Honestly if kickoffs weren't we required, they wouldn't need a kicker.

bobbye

check out an exclusive interview with Coach Kevin Kelley following Pulaski's recent 5A state championship only on www.straitpinkie.com

www.straitpinkie.com/sports/the-pinkie-interviews-pulaski-academy-head-coach-kevin-kelley

Coach

There is a major problem with the argument Romer presents. What is the opposite of making the first down? A TURNOVER. What coach in his right mind would call a play that risks a TURNOVER 40% of the time. Very few at the NFL level as we can see by statistics. So few in fact that you cannot even get a scientifically significant sample. I think most of these coaches can understand a complex problem better than this scientist. As the yards you need for the first down increase the odds of the turnover increase too. Not a good decision in most cases.

As far as a two point conversion goes. It is meaningless because a two point coversion has a 0% chance of a turnover. That is a decision you make according to the score of the game and other factors.

Mike

When I played Pop Warner as a kid, our PAT rules were reversed - 2 pts for a kick, and 1 pt for a conversion. This was mainly because most elementary school and jr. high kids couldn't kick field goals from any distance. My team had a decent kicker, who was accurate maybe a little over half the time. So yeah, in our case, going for 2 was worth it.

Andrew

Rolub, that's exactly what I was going to say. Belichick seems to go for it quite often in situations where other coaches would not.

James

What about going for the 2-point conversion as opposed to the PAT as the default strategy? Even if a team converts the PAT %100 of the time, any 2-point conversion rate over %50 would make it the obviously superior choice in all but the most select situations. As the state of football conventional "wisdom" currently exists, the opposite is true. Of course, one of the fundamental ideas of Freakonomics is that "The conventional wisdom is often wrong".

David

James (re: 2-point conversion),
You have to consider the marginal value of a point. At the end of a game, the one point you score that's greater than your opponent has 100% of the value and all additional points are meaningless (except to gamblers). Likewise, failure to score that 1 point can have dire implications.

Over the course of a season (total points scored), your suggestion would be accurate, but it would not be an appropriate strategy for individual game management.

Note: I wonder my statement about would not hold true for a team like Texas Tech, who scores lots of points but lacks a stout defense. They have high scoring games, where lots of touchdowns allow for lots of 2-point attempts. This also depends on other factors (e.g., their red zone TD%), but might be an instance where hoping for 50% could work.