Regarding that fourth down decision by Pat Shurmur on Sunday, a fine gentleman named Tim Smith contacted me on Twitter and pointed me in the direction of a site called advancednflstats.com.
Tim is a fine gentleman, very instructive and patient with a curmudgeonly Irish dude who has trouble with on-off buttons.
So I checked out what he said.
This site is an offshoot of all the baseball sites that look at things in a Bill James/sabermetric way. Stats, calculations, that kind of thing.
The fine Mr. Smith pointed out this gizmo (technical term) called a “fourth down calculator.”
In that calculator, you punch in the situation and the gizmo punches out statistical probabilities for going for the first down vs. punting or trying a field goal. It takes things a step further by determining the success rate of that down and distance based on results since 2002.
That’s 10 years of results, which is not insignificant.
In the scenario Shurmur faced on Sunday, the Browns trailed by seven and had fourth-and-2 with 3:53 left.
Punch those numbers in and the following results are spit out:
–Teams that faced fourth-and-2 since 2002 made the first down 60 percent of the time. Clearly the Browns have dragged that figure down a tad. As gentleman Tim points out, the numbers do not take into account specifics like how the Browns offense is playing and the quality of the defense they are facing. But it does give an idea that the odds of making the first down were better than 50-50. (Unless you walk out of the huddle like you’re in dazed and confused and run a play terribly.)
–The site then spits out the probability of winning with either decision, going for it or punting. In this instance, going for it produced a 10 percent chance of winning; punting produced a six percent chance of victor. Neither are very high, but going for it increased the likelihood of winning by four percent.
–Making the first down doesn’t exactly guarantee much. Making it would have increased the likelihood of winning to 15 percent, but missing it dropped it to three percent.
–The site then produces something called a Win Probability Total, which takes into account the six percent chance of winning with a punt, 10 percent chance of winning by going for it and the 60 percent likelihood of making the first down. And probably takes into account some numbers I’m too dumpy to understand.
This Total produces a bottom line that the break-even number is 25 percent. Which means that if Shurmur felt he had a better than 1-in-4 chance of making the first down, he should have gone for it.
In short, the site and the numbers say he made the right call.
But … one thing that the numbers do not consider is the timeout situation. Which is kind of important, since the Browns had two and the two-minute warning.
It also does not consider flow of the game, as in the Browns defense had been playing far better than the Browns offense. And that the Ravens offense would probably have been conservative had it gotten the ball back. In my mind, the Browns probably get the ball back with about 1:50 left and no timeouts, which is plenty of time for a real NFL offense to drive and score — or at least have a chance to score (the Browns had the chance with less time in the game in Baltimore).
In my mind, those intangibles outweigh the four percent difference between going for it and punting, which to me made the punt more logical.
But in this case, these numbers, these cold, hard numbers, say Shurmur made the right decision by going for it.
(With great thanks to Tim … a fine and true gentleman)