THE PLANK NOVEMBER 16, 2009
-
Read Later
READ LATERAvailable only to subscribers. SUBSCRIBE TODAY
-
Listen
ARTICLE AUDIO
- Font Size
For those who missed the football game and didn't watch SportsCenter this morning, last night's contest between the Colts and Patriots turned on an already controversial fourth-down call.
The Partiots had the ball, fourth and two, at their own 28-yard-line. They were ahead but only by six points, 34-28, following a fourth-quarter Colts touchdown. The clock showed 2:08.
The conventional call was to punt. And although Patriots coach Bill Belichick is famous for eschewing punts on fourth down, it was hard to believe he'd go for it in that situation, so close to his own end zone. But he did. And it almost worked. Brady, throwing from an empty backfield, hit Kevin Faulk right at the first down marker. But it appeared that Faulk bobbled it for a moment; by the time he'd gained full possession, an Indianapolis defender had driven him backwards. The spot put him short of the marker.
At that point, Peyton Manning and the Colts took over, methodically moving down the shortened field and running out the clock. They scored the winning touchdown with just ten seconds left.
The second-guessing started with the Belichick post-game press conference and hasn't stopped since. Maybe it's the shock of seeing Belichick, widely considered the smartest coach of his generation, have a call backfire. Or maybe it's the shock of seeing such a rare decision. Whatever. Nobody seems to think he got it right. "Belichick call unrivalled," writes the Boston Globe's Dan Shaugnessy--employing gentler language than what, I'm sure, callers into Boston talk shows are using this morning.
But statistics show pretty conclusively that football coaches are far too conservative about fourth down decisions--that they should go for it, rather than punt, far more frequently. Apparently, statistics also show that Belichick made the right call here, notwithstanding what everybody thinks. From the website "Advanced NFL Stats":
With 2:00 left and the Colts with only one timeout, a successful conversion wins the game for all practical purposes. A 4th and 2 conversion would be successful 60% of the time. Historically, in a situation with 2:00 left and needing a TD to either win or tie, teams get the TD 53% of the time from that field position. The total WP for the 4th down conversion attempt would therefore be:
(0.60 * 1) + (0.40 * (1-0.53)) = 0.79 WP
A punt from the 28 typically nets 38 yards, starting the Colts at their own 34. Teams historically get the TD 30% of the time in that situation. So the punt gives the Pats about a 0.70 WP.
Statistically, the better decision would be to go for it, and by a good amount.
Brian Burke, writing at the New York Times, and Nicholas Beaudrot agree.
All of this comes courtesy of Matt Yglesias, who--I know--took little joy in making this case. For what it's worth, I feel the same way. Although my devotion to another Boston franchise is a matter of public record, my pro football loyalties remain with the team I grew up watching in South Florida.
Update: Readers are already writing in to suggest that the pro-Belichick analysis fails to account for the specifics of the situation--it was Peyton Manning at QB, etc. Actually, that's not true. I just didn't excerpt those parts. Here's more from Advanced NFL Stats:
You'd have to expect the Colts had a better than a 30% chance of scoring from their 34, and an accordingly higher chance to score from the Pats' 28. But any adjustment in their likelihood of scoring from either field position increases the advantage of going for it. You can play with the numbers any way you like, but it's pretty hard to come up with a realistic combination of numbers that make punting the better option. At best, you could make it a wash.
And now, here's Beaudrot:
New England has gained a first down on approximately 66% of its attempts with Tom Brady as quarterback. The Colts had no timeouts. If the Patriots gain a first down, the game ends; they can simply wait for the two minute kneel on the ball three times to end the game. If they don't gain a first down, the Colts would still need to score a touchdown to win the game. Let's give the Colts a probability P of getting the six if the ball starts at the 28 yard line. So if the Patriots try for the first, their chance of losing is
(Probability of 4th down failure) x
(Probability of Colts scoring a TD from the 28 Yard line) = 0.33P
The average New England punt nets about 40 yards. Let's give the Colts a probability Q of scoring a TD on a driving starting at the Indianapolis 32. Then, the chance of the Patriots losing is simple Q. For Belichick's decision to make sense, we just have to believe that he gave his team a higher chance of winning. In math terms, that would mean 0.33P > Q. Doing some algebra leaves you with P > 3Q. In other words, you would have to believe that the Colts odds of scoring a TD on a drive starting 28 yards from the end zone are more than three times the odds of the same outcome starting from 68 yards out. The win probability graph for the game suggests that, given 1st-and-10 from New England's 29, the Colts had roughly a 51% chance of winning in the actual situation. We would have to believe that their chances under the punt scenario were below 17% for Belichick to have made a bad decision. Considering the Colts' have scored touchdowns on 30% of their offensive possessions, my guess is that this was a good one.
14 comments
But do those statistics suffer from selection bias? Most teams only go for it on 4th down when either it's so short to be nearly a guarantee (4th and 1, 4th and inches) or they have to, late in the game in four down territory. You don't have the results of people going all the time, anywhere, to make a truly accurate statistical picture. And the idea that Manning/Colts only get a TD from the 28 53% of the time seems to strain credulity. Giving the Colts the ball at your own 28, with the incipient momentum that comes from your 4th down failure, confidence from the short field meaning that they don't have to rush the 2 minute drill, etc. makes for a far more likely TD than 53%. The statistics seem underspecific to the actual situation, and over generalized in the likelihood of success.
- Crock1701
November 16, 2009 at 11:48am
What Crock said: "Giving the Colts the ball at your own 28, with the incipient momentum that comes from your 4th down failure, confidence from the short field meaning that they don't have to rush the 2 minute drill, etc. makes for a far more likely TD than 53%." Simply put, Belichick was gambling. And anyone who knows anything about football knows that the odds were absolutely not with him. The statistics Isaac C cites do not take into consideration the specific game conditions of this particular play. When you do take these game conditions into consideration, it's obvious: you don't go for it on your own 28 on 4th and 2 with two minutes left in regulation unless you are TRAILING, and MUST score to tie or win. Or unless you have ABSOLUTELY no confidence in your defense. Belichick seemed to think his offense had a better chance against the Colts' defense than his defense had against Peyton Manning. But even if this were the case, it only increases the odds--vastly increases the odds--that his defense won't be able to keep the Colts out of the end zone if the 4th-down play doesn't work. Which is to say, Belichick was gambling, big time, like a poker player risking everything to draw to an inside straight. And, well, he lost.
- JosephCuomo
November 16, 2009 at 12:16pm
No statistical analysis that is preceded by the word "momentum" can be taken seriously. The arguments for going for it on 4th down are compelling, but selection bias will always punish an innovative coach who puts a go-for-it strategy into practice. In close games, fans and team ownership will remember the one time it lost you the game and not the three other times when it saved the game. Given that there are still many people who believe that it ever makes sense to intentionally walk a batter in baseball -- it never does, except for bottom-of-the-ninth ties or one-run leads with runners on, first open, and less than two out -- the conventional wisdom of always punting will persist forever.
- rhubarbs
November 16, 2009 at 1:06pm
I'm with Tony Dungy: they should've punted plain and simple.
- basman
November 16, 2009 at 1:12pm
rhubarbs- You write: "In close games, fans and team ownership will remember the one time it [going for it on your own 28 on 4th and 2 with two minutes to go in regulation] lost you the game and not the three other times when it saved the game." Okay, rhubarbs, I'll bite: where's your statistical analysis that says that going for it in that situation works three out of four times? And how do you factor in other relevant statistics, such as Peyton Manning's career percentage of scoring a touchdown inside his opponent's 30-yard-line with two minutes left in regulation and the game on the line? Or such as the Pats' career percentage of stopping their opponents from marching 65 or 70 yards down the field to score a touchdown with two minutes left in regulation and the game on the line? When you've looked up those stats, and factored them into your analysis, well, then, please get back to me.
- JosephCuomo
November 16, 2009 at 1:24pm
Belichick was wrong. One purpose of sport is to provide a reality different from that governed by rational laws. Not necessarily a reality that is 100% irrational, but one in which reason and logic are pimp-slapped by results. (A correlative beauty of sport is that it's supposed to be fair, but often isn't; curiously, a correlative beauty of life is that it's not supposed to be fair, but often is.) Working backward from the Colts/Patriots result, therefore, one can conclude only that Belichick was wrong. This, regardless of data; had 99.9 percent of coaches who went for it on fourth and two been successful, Belichick still would have been wrong. Winners never suck; losers always do.
- williamyard
November 16, 2009 at 2:10pm
The relevant analysis on this very page indicates that it's about a three-to-one advantage to go for it in that situation. Even aside from the algebraic proof that Belichick made the right call, we must also consider the aesthetics of the game. The NFL is boring enough to begin with, so I approve of any coach who rather than playing to avoid losing, plays to win. Punting was the play you call if you're afraid to play to win. It's the play for coaches who would rather lose respectably than risk their jobs to win. Punting makes the game about whether the other guy fails in his effort to win -- and the Colts probably win even after a punt. Going for first down was playing to win, rather than playing to minimize the chances of losing. Punting is the play for cautious sissies who aren't really man enough to play football anyway. It's the croquet fan's move. It's what a soccer coach would do. In that situation, Marshal Petain punts. General Patton goes for first down. Going for a first down against the Colts inside the Pats own 30 was glorious -- even beautiful -- football, and there's precious little of either in the NFL these days, so even if it weren't the best-odds call, I'd still approve of the play.
- rhubarbs
November 16, 2009 at 2:40pm
Belichick was right to go for it, but wrong by some mileage to throw the predictable, shotgun formation, 5 receiver set, none yards pass. That let the Colts pass rushers pin their ears back, opened up myriad potential blitz packages the Patriots had to scheme against, and dbacks looking for the extremely short yardage, out or in. Well- they got what they asked for. In for a dime, in for a dollar. If you're going to strut your brass cahones, don't follow that up with a play call fit for a pantywaste. See for further example: off-guard dive, USC and Texas national championship game, White, Lendale.
- I Majorajam
November 16, 2009 at 3:06pm
p.s. Brett Favre: M.V.P. Analyze that motherfuckers.
- basman
November 16, 2009 at 3:59pm
I do think Rodney Harrison made an good point about the play looking like a vote of no confidence in your defense. How do you quantify the impact of that on the rest of your season? I think they still would have lost if they punted though. The defense was playing really well, but they were tired and guys were getting injured as the game wore on. The Colts ripped them apart in the previous drive. Even so, Reggine Wayne had to make a hell of a catch for the Colts to win.
- ClumsyMohel
November 16, 2009 at 4:25pm
rhubarbs- You write: "The relevant analysis on this very page indicates that it's about a three-to-one advantage to go for it in that situation." But the stats that Jonathan C cites get at least a few things wrong. First, Jonathan C cites this stat: "Historically, in a situation with 2:00 left and needing a TD to either win or tie, teams get the TD 53% of the time from that field position." But according to today's NYT: "The Patriots. . .are 17th in the N.F.L. in fourth-down efficiency. They have gone for it 11 times on fourth down and converted five times." That's a statistical probability of less than 50% (not 53%). And this stat does not take into account several other factors: such as Peyton Manning's career percentage of scoring a touchdown inside his opponent's 30-yard-line with two minutes left in regulation and the game on the line, as opposed to Manning's career percentage of scoring a touchdown starting at his own 30-yard-line (or 35- or 25 yard-line) with two minutes left in regulation and the game on the line. Or such as Belichick's career percentage (as head coach) of stopping his opponents from marching 65 or 70 or 75 yards down the field to score a touchdown with two minutes left in regulation and the game on the line. Without those stats factored into the mix, the analysis Jonathan C cites is just plain meaningless. If, though, as you suggest above, you think that it was more exciting and ballsy for Belichick to go for it in that situation, I would say fine, yes, it was. It was also more exciting and ballsy for McCain to have chosen Sarah P as his VP. But it was also ineffective and stupid, if he truly wanted to win. A similar thing could be said about Belichick's 4th down decision last night.
- JosephCuomo
November 16, 2009 at 5:02pm
Sorry, rhubarbs, what I meant to say was that the stats Jonathan C cites start with one basic presumption: "Historically, in a situation with 2:00 left and needing a TD to either win or tie, teams get the TD 53% of the time from that field position." But this stat doesn't appear to factor in field position, and the degree to which said field position affects the percentage of success. (It's likely, for instance, that the chance of scoring a TD from inside one's opponent's 30 is much higher than one's chance of scoring a TD starting a drive from inside one's own 30.) It also doesn't factor in the specific histories of the participants in this game, ie, Belichick's career history (as head coach) of stopping his opponents from scoring a TD from inside his own 30 with two minutes left in regulation with the game on the line, versus his career history of stopping his opponents from scoring a TD from inside their own 30 with two minutes left and the game on the line. My point, as I suggested above, is that the statistical analysis Jonathan C cites (and you support) is faulty in the extreme. And doesn't present a true picture of the odds in this given situation. But, as I also said above, it was indeed an exciting and ballsy play for Belichick to call. It just wasn't very effective.
- JosephCuomo
November 16, 2009 at 5:18pm
A 5-for-11 record has no predictive value for attempt number 12. The sample size is obviously too small to have predictive value, and furthermore the relevant data set is not fourth-down conversions but yards per down and/or average gains on all 2-yard downs. In this case, the Patriots entered the game averaging a gain of 5.9 yards per offensive play, and convert the large majority of all 2-yard downs. It's OK to engage in innumerate magical thinking about a subject as trivial as sports. No need for anyone who doesn't to pretend to know or care about statistical analysis; it's fine to go on talking about "momentum" and other such conventional-wisdom claptrap and leave the numbers to the boring geeks we'd rather ignore anyway. I still believe in jinxes and batters being "due", and to heck with the killjoy statheads who prove that neither exists. What I don't do is make the mistake of fabricating numbers to "prove" the existence of jinxes and due-ness and argue numbers with the spoilsport geeks.
- rhubarbs
November 16, 2009 at 5:36pm
rhubarbs- You write: "What I don't do is make the mistake of fabricating numbers to 'prove' the existence of jinxes and due-ness and argue numbers with the spoilsport geeks." You may not do this, rhubarbs, but the sources Jonathan C cites above (and of whom you apparently approve) certainly claimed that their numbers "prove" that there was a statistical probability that Belichick made the right call. As I suggested above, I think their stats and their reasoning and their analysis was, well, just plain silly. As for, as you put it, "conventional-wisdom claptrap," it's also conventional wisdom that drawing to an inside straight with one card to come in hold em is a very low percentage play (with about a 10% success rate). While it might make sense to gamble for a small percentage of your money on such a play (with the possibility of winning a lot more if you make it), I've seen poker players risk everything on just such a play and, yes, hit the straight, but I've also seen them lose everything a whole hell of a lot times more. Which is to say, I love playing poker with such risk takers (it can be very lucrative!). And if my team were playing the Pats, I'd love to see Belichick make the same call on 4th and 2 on his own 28 over and over and over again.
- JosephCuomo
November 16, 2009 at 6:55pm