Belichick had the numbers on his side
FOXBOROUGH - Among the countless criticisms hurled at Patriots coach Bill Belichick for his decision to go for it on fourth down Sunday night, former Colts coach Tony Dungy summed up the most popular when, speaking on NBC, he said, “You have got to play the percentages and punt the ball.’’
What Dungy did not realize, though, is that “the percentages’’ dictated that Belichick do exactly what he did.
The Patriots coach has been lampooned by experts, fans, and former players for his call. All of them followed football’s rigid dogma and not what mattered: Did the Patriots have a better chance to beat the Colts if they went for it on fourth and 2 than if they punted? The answer is yes.
While Dungy spoke and football-watching mobs sharpened pitchforks, a former Navy pilot named Brian Burke ambled to the computer in his Reston, Va., home so he could determine if Belichick had made the right decision.
Burke is a football-crazed, math-inclined single father who works for a military contractor. Three years ago, he found himself with nothing to do once he put his children to bed. “I was pretty bored,’’ Burke said. He used software leftover from grad school to create a simulation tool that could solve football arguments for his friends at work. When they stopped paying attention, he put his findings on a website.
The result is AdvancedNFLStats.com, a site devoted primarily to determining win probability for football teams. Burke’s simulation engine is perfectly suited to solve debates such as, say, whether a team should go for it on fourth and 2 from their 28 with two minutes left while leading by 6.
Late Sunday night, Burke compiled the data and ran the numbers. When Burke began, even though he believes coaches are too conservative, his gut told him Belichick had made the incorrect decision. His outcome proved him wrong.
According to Burke’s tabulation, going for the first down gave the Patriots a 79 percent chance of winning. Punting gave them a 70-percent chance to win. Even after Burke made tweaks, the win probability never dipped in favor of the punt. If anything, factoring in how explosive the Colts’ offense is, the team-specific adjustments only made going for it more favorable.
“A lot of criticism is probably way over the top,’’ Burke said. “At the very least, it’s defensible. It’s not crazy. It’s not reckless.’’
Sheer data, which Burke has compiled and stored like a librarian, argues the point. On average, an offense operating outside the red zone will make the first down on fourth and 2 60 percent of the time. When teams face the situation the Colts would have had if the Patriots failed - two minutes left, needing a touchdown, at roughly the opposing 30 - they score 53 percent of the time.
The Patriots would certainly win the 60 percent of the time they convert. They would also win the 47 percent of the time they’d stop the Colts. Overall, going for it gave them a 79 percent chance to win.
Now, what if the Patriots had punted? On average, the net punt would have been 38 yards, and the Colts would have taken over on their 34. Statistically, teams will score 30 percent of the time in that situation, meaning a punt gave the Patriots a 70 percent chance to win.
Burke is not the only statistically minded football source to draw the conclusion. The ZEUS program, developed by a pair of champion backgammon players, was made to simulate specific football situations and spit out probability. One of the developers told the New York Times that Belichick had made the right call, their numbers similar to Burke’s.
Critics of Belichick made two mistakes. First, they underestimated the chances of converting a fourth and 2 and overestimated the difference a punt makes. Playing with abandon against a preventative defense, an offense can typically pick up the yardage from a punt in a matter of three plays and 30 seconds, Burke said.
(Burke, it should be noted, did not wholly absolve Belichick. Burke believes Belichick, who knew he was going to go for the fourth down if necessary before third down, should have run the ball rather than pass on third down.)
“The one thing people aren’t looking at is that third-down call,’’ Burke said. “An unsuccessful pass on third down gives you fourth and 2. An unsuccessful run is going to give you a real short fourth down and make your chances of winning better.’’
Second, they mistook convention for truth. Football is a conservative sport coached by conservative men. In moments of uncertainty, when a fast, important decision is required, they revert to what they know, what has been passed down, what is safe.
“The traditional decision has always fallen back on the punt,’’ Burke said. “That’s how human beings have evolved to make decisions. They fall back on the punt.
“Maybe 30 years ago, that made a lot of sense. You didn’t have a lot of high-powered offenses. The math was different. Punting on fourth down made a lot of sense then. Now, you have guys like Peyton Manning and Tom Brady.’’
Burke is often frustrated by the unwillingness of coaches to break their tendency for the safe. “Belichick,’’ he said, “is the exception to the rule in a lot of ways.’’
And Sunday night, that helped him make the right decision. All he had to do was play the percentages.
Adam Kilgore can be reached at email@example.com.