Go For It, Please
In a recent edition of SIVault, L. Jon Wertheim writes about the decision-making of the football coach for the Pulaski Academy Bruins (a high school located on the west side of Little Rock AR). For me the article, brought to my attention by my son, is more about life than sports.
The coach, Kevin Kelley, does not list a putter or a kicker on his roster. He never punts. And on kickoffs he either on-side kicks or kicks it out of bounds. Pulaski has not punted since 2007 (when it did so as a gesture of sportsmanship in a lopsided game). According to Kelley, “The average punt in high school nets you 30 yards, but we convert around half our fourth downs, so it doesn’t make sense to give up the ball. Besides, if your offense knows it has four downs instead of three, it totally changes the game.” The coach also does not run back punts.
Consider a worst case scenario, fourth-and-long near your own end zone. According to Kelley’s data, when a team punts under these conditions, the opponents take possession inside the 40-yard line and will then score a touchdown 77% of the time. If they recover on downs inside the 10, they’ll score a touchdown 92% of the time. “So [not punting], you give your offense a chance to stay on the field. And if you miss, the odds of the other team scoring only increase 15%. Kelley has similar logic for on-side kicks and for not having punt returns.
According to the SI article, “Much of Kelly’s analysis has support among number crunchers. In 2005 David Romer, a prominent Cal economist, published a study that argued that over the course of the three NFL seasons he studied, there had been 1,068 fourth-down situations in which teams, mathematically, would have been better off going for it. In all but 109 cases the teams either kicked or punted.”
Personally, it seems like Kelley’s strategy coupled with random offense play selection (so it is harder for the defense to guess what you are doing) is worth serious consideration. For example, on third or fourth down and short, if you were willing to try almost anything, your odds of making it should improve dramatically. If you just crash the line everytime, it just muscle against muscle. It’s important to keep the defense guessing.
Coaches are too afraid of being second guessed by the damn color commentator, rabid super fans, sport talk shows, newspaper sports columnists, and their ilk. This causes the head coach to stay with “traditional” ultra-conservative responses to situations. If coaches were ocassionally willing to take the heat for an “usual” loss, they would probably improve their winning percentage significantly. Coaches play “not to lose,” instead of “to win.”
If you can score a 2-point PAT over fifty percent of the time, you probably ought not to be kicking PATs. And why all the settling for 3-pointers? If you can score a touchdown over 50 percent of the time, why kick? Let’s open the game up. These “new” strategies would be much more interesting to the fans than what the game of football has devolved into.
The judicious use of models and statistics can also be used in recruiting and team development. To quote my brother Lars Hansen (from a graduation speech at the University of Chicago): “The practical value of statistical theory has been documented in a variety of contexts, most recently in a popular treatise on the success of Oakland’s professional baseball team. Michael Lewis showed how the formal use of probability models and data on the performance of baseball players helped to allow the Oakland Athletics to field a highly competitive team while spending far less than teams like the New York Yankees or even the Chicago Cubs.”
Sports are obsessed with statistics; teams need to use them better. Bureaucracies have the same problem. Individuals have such a great fear of being second guessed, decision-making becomes painful and frequently needlessly delayed while it goes through unnecessary reviews. And, in many cases, the product becomes heavily compromised and not best result. But the decision is “safe.”
September 30th, 2009 at 3:27 pm
I’m a bit leery of running an entire football game on statistics. Especially high school ball. What about old fashioned play, gut instinct, going for the “play of the day”. Why does high school football now turn into a statistical nightmare? What am I missing? Sometimes we need to just go back to good old fashioned clipboards, plotting the plays in a sweaty, stinky lockerroom. Not high-tech computers that calculate each player and every move in the Sports Arena. I recently watched “Hoosiers” on TBS. One of the best sports movies made, in my opinion. I am confident that the majority of that scenario was based on good old fashioned teamwork and values. I’m starting to sound like a prude.
September 30th, 2009 at 4:09 pm
This page is a bit convoluted and I don’t necessarily disagree with your point. First, I think the game of football would be much more interesting if teams were not so quick to punt or kick. They would also end up with higher scores and help keep their defenses off the field. Second, if teams would ramdomize their play selection, they would do better. The latter is not meant to demean the role of the quarterback, it is meant to help him with play selection. But the quarterback would still have to read the defenses and react accordingly.
As for my brother’s point: he is just restating a way for teams in small markets to stay competitive using statistics. But some of the same methodology could be used in other facets of sports.
November 19th, 2009 at 11:47 am
According to Time Magazine, the 33rd best invention of 2009 is “The No-Punt Offense:”
As Sports Illustrated explaned in a recent story, Kevin Kelley, coach of the Pulaski Academy football team in Little Rock, Ark., has called for only a single punt in the past two years. Like a seasoned gambler, Kelley has figured out that punting on fourth and long near your own end zone decreases the odds of the other team’s scoring by only a relatively slim amount. So going for it will pay off in the long run: Pulaski won a state championship last year and is in the hunt this year too.
November 30th, 2009 at 2:20 pm
Over the last two weeks, the NFL has had three good examples of the pathetic situation surrounding punting or not punting on fourth down. On all three examples the pundits were wrong, dead wrong. The one example eight days ago provided fodder for the arm-chair pundits for an entire week. It involved the Pats going for a 4th and 2 (Patriots vs Colts) on their own 40 with just a few minutes to go. The Pats were ahead, but still decided to go for it. After the play was over, they got a bad ball placement, and turned the ball over. The Colts then went down and scored. The Pats lost the game and had to live with 24-hour damning from everybody for an entire week. The Pats probably deserved criticism, but it was not for not punting, it was for the way they handled the clock and the previous plays. They were correct for not punting. If they had made the 4th down and two, the game would have been over. By punting they would have merely made the Colts’ drive longer. And Payton Manning is very adept at last minute scores, whether over short or long distances.
The next example was during the Viking-Bears game yesterday. It was halfway thru the 3rd Quarter, the Bears punted on a 4th and 1 near midfield. The Vikings welre killing them, running and passing them into the ground. The Bears needed to give their defense a rest. They needed to change the momentum of the game. Why the hell would you punt? Yet the announcers praised the Bears’ coach for making the right decision. The Viking, after getting the ball, went down and scored a TD. The Bears ended up getting hammered. Any pundits second-guessing this decision? Of course not, and that is why coaches are so conservative.
The last example is perhaps the strangest. The Steelers were playing the Ravens. The Steelers, who were playing with their 3rd-string inexperienced quarterback, were surprisingly ahead by 3. There was 4 minutes left in the game. The Ravens were just inside Steeler territory. Fourth and 4, they decided to go for it. The announcers were predictably apalled. Baltimore didn’t make 4, they made 20 plus. Their drive then stalled and they kicked a field goal, tying the game. They went on to win in overtime.
Coaches are way to “conservative” because of the Monday-morning backstabbers. They punt way too often and they settle for 3-points way to often. And I suspect they don’t go for 2 often enough. Please teams, hire real statisticians. You will win more games if you forget the alleged pundits.
January 3rd, 2010 at 10:39 am
On New Year’s Day morning I watched the Florida Citrus Bowl from Orlando. It was raining and the field was in terrible shape. It was in embarassingly bad shape. As the players tried to play, large chunks of grass would move under their feet making for a muddy mess. After a short time, it looked as if the field had been invaded by a herd of moles. In this day and age, this in inexcusable. To make matter worse, the officiating was bad. The group of officials at the game were the definition of incompetance.
But worst of all, JoPa is so conservative that he deserved to lose. Unfortunately he won. On a 4th-and-a-yard (in the first half) on the LSU forty, he punted. The Penn State kicker punted it into the endzone, for a net of 21 yards. Big deal. In the second half, near his own 40 with less than a yard to go, JoPa again decided to punt. This time the punt netted less than 10 yards as the LSU kick-return man made a nifty run. And of course, let’s always settle for 3-pointers when the opportunity presents itself. JoPa please hire a real statistician, and quit listening to the stupid pundits.