Writers of Pro Football Prospectus 2008

01 Oct 2012

The Panthers' Fourth-Down Decision

Tommy Craggs asked me to contribute to the Slate/Deadspin weekly roundtable today, bringing the fight against the NFL's Punt to Win philosophy. In case you couldn't tell from my comments in Audibles, I think Ron Rivera's decision to punt late in yesterday's game was pretty damn awful.

UPDATE: As long as we're talking about fourth downs, sigh, I might as well link to this article about Jim Schwartz by Jason Lisk. This article makes me sad.

Posted by: Aaron Schatz on 01 Oct 2012

29 comments, Last at 03 Oct 2012, 6:13am by Pat (filler)

Comments

1
by JasonK :: Mon, 10/01/2012 - 3:14pm

FYI, the version of this article that is cross-posted on Slate has an amusingly out-of-date author minibio:

"Aaron Schatz is editor in chief of Football Outsiders and the lead writer of Pro Football Prospectus 2007."

2
by James-London :: Mon, 10/01/2012 - 3:14pm

Yep. Pretty much this. Awful decision.

Phil Simms is a Cretin.

3
by herpderp (not verified) :: Mon, 10/01/2012 - 3:17pm

It's also the "metadecision" which is highly questionable. You've got a young quarterback who's just fumbled the ball. If you let him go get it, he gains redemption. If he doesn't get it, the media ire goes on you. Instead, Rivera punts, and the story of the Panthers week is how Newton choked.

Of course, all that is much less important than the fact that the decision made it much less likely that his team would win the game.

4
by billsfan :: Mon, 10/01/2012 - 3:37pm

Each team in the Pats/Bills game punted from the opponent's 41 a nonzero number or times, and at least one of those coaches probably knows better. Even so, Belichick still sent in his 0/1 on the day kicker on 4th-and-1 from the 25.

Of course, all this happened in the first half, but 2/3 of those decisions were surprising.

(I also like the Eagles)

5
by Richie :: Mon, 10/01/2012 - 4:01pm

Sometimes the games start to blend together on me. But didn't Atlanta punt on the possession before the Carolina possession, in a situation that they should have gone for it as well?

(I guess I should check the PBP log.)

6
by Richie :: Mon, 10/01/2012 - 4:06pm

Nope. I must be thinking of a different game. Looks like Atlanta had 4th and 11 at ATL 28 with 2:26 to play. Not a great spot to go for it.

7
by langsty :: Mon, 10/01/2012 - 5:15pm

I guess I didn't think it was a terrible call by the Panthers. According to Elias, "no other team in the past 20 years has conducted a game-winning drive starting inside their own 5 with less than one minute to play." As a Falcons fan, I was dreading the punt and hoping they'd just go for it.

8
by tuluse :: Mon, 10/01/2012 - 5:32pm

That sounds awfully specific. Does that mean a team starting from it's 6 yard has? Do those 5 yards really make the difference?

Does it mean a team with 1:01 has?

10
by Led :: Mon, 10/01/2012 - 5:44pm

You don't know you're going to down the ball inside the 5 before you punt it! I suspect the odds of downing inside the 5 are comparable to the odds of picking up the first down if you go for it, if not worse.

27
by Dennis :: Tue, 10/02/2012 - 4:08pm

Right, there's no guarantee you're going to get them pinned back like that. This is one place where I actually agree with TMQ - teams don't go for it on fourth down nearly as often as they should.

11
by Richie :: Mon, 10/01/2012 - 6:12pm

I was rooting for Atlanta and very happy when Carolina punted. I was less happy when the ball was downed at the 1. (And Carolina was about an inch away from it turning into a touchback.)

9
by WeaponX (not verified) :: Mon, 10/01/2012 - 5:40pm

Nakamura's poor play is obviously a bigger force than 20 years of history.

12
by BigCheese :: Mon, 10/01/2012 - 6:45pm

And the ESPN article you linked to whose point was basically "Every single stat PLUS the actual outcome of the game PLUS the actual strengths and weaknesses of the teams involved tell us that punting was abolutely the worst call. But I can't set myself loose from so-called conventional wisdom for even a single second so I say it was the right call" DOESN'T make you want to brak down sobbing uncontrollably???

- Alvaro

Phil Simms is to analysts what Ryan Leaf is to NFL QBs

13
by Rots (not verified) :: Mon, 10/01/2012 - 8:19pm

If i were the owner of the Panthers what Rivera did was an instant termination offense. Inexcusable.

Ive said it a thousand times but on gamedays at least almost every non Pats/49ers team would be better off being managed by some 18 year old with 10000 games of Maddens under his belt.

My beloved Broncos would be better served if the worlds most conservative coach grew some testicles and went for it on 4th and 1 now and then as well.

15
by jackiel :: Mon, 10/01/2012 - 9:13pm

The great thing about Fox is that he puts his faith in such a mediocre defense.

20
by Rich A (not verified) :: Tue, 10/02/2012 - 12:29am

I'm not a top ranked player madden player anymore but a few years ago I was in the top couple hundred. So much of madden unfortunately comes down to knowing what is broken in the game and exploiting it, be it putting CB's at WR positions or viceversa (Hester back in the day), or constantly attacking weak spots of the field that AI has trouble defending.

Beyond that, most players only run about 5 or 6 plays, and usually to the same players. It's actually incredibly easy to play most players once you've read some smartfootball and blitzology as well as knowing what's broken in the game and how to cover that weak spot manually as well as hide your teams inherent weakness. This is because they almost never give the same look and then do something different be it PA, a screen, or whatever, it's just one play from one formation, then one play from a different formation then back to the first formation. And you can also find 'the money plays' on youtube videos and so you can know how they're going to attack you before you even play because you know what they're trying to do based on team and formation.

I was always able to score at will but when it came to another person that could score at will, in 7 min quarters the final score became 106-105 or something ridiculous like that. Sometimes you would just run into a razor of a player. Those guys in the top 50 have it all figured out.

Interestingly enough this kinda sounds a bit like what Buffalo was saying about their coaches yesterday, they just thought that they had a decent personnel grouping and then the Pats just kept blasting away with 2 or 3 different running plays. Of course the Pats also went PA to Gronk for a TD off their running formation so they have a bit more creativity than the average madden player.

If you want to get scheme diverse playcallers find the top rank NCAA football videogame players. That game is much more creative with play calling. And always audible your manually controlled player to blitz, even if you aren't blitzing him, it messes with how the O-line will automatically slide, giving your actual rushers better pressure. Lastly, in the red zone, blitz 7-9 or drop 9 into coverage from showing blitz, the AI is ridiculous to pass against with that many in coverage.

If you want to see good commentary on a game, follow the starcraft 2 community. Day9 has some incredibly insightful things to say about strategy. Plus that's where the future of games is going. And I'm not just talking videogames, athletes too are taking to games rather than sports, which I think is unfortunate.

21
by tuluse :: Tue, 10/02/2012 - 12:35am

I don't think he mean picking which plays to run, but rather time out usage, when to go for it on 4th down, those types of decisions.

23
by The Ninjalectual :: Tue, 10/02/2012 - 9:31am

It's still a good argument for why those kids might not be the best judges. Too much going on in the video game that isn't there in real life.

25
by zlionsfan :: Tue, 10/02/2012 - 12:10pm

If "real life" means trying not to lose by too much instead of trying to win, then perhaps the kids are the better judges.

Certainly there are some aspects of Madden strategy that wouldn't translate - playing players "out of position", for example, as you can't simply tell your #1 WR that he isn't going to start; that's more an exploit than strategy anyway - but the overall idea of doing whatever you can to win is something that most NFL GMs would do well to consider. (Coaches, I think, don't have that option yet because they're graded based on the risks they take rather than the results they generate.) I don't think punting on fourth and one in your opponents' half of the field, for example, is more valid simply because real-life coaches default to that option.

28
by Rich A (not verified) :: Tue, 10/02/2012 - 6:23pm

Timeout usage is awful online, except for the top 50 players

14
by mansteel (not verified) :: Mon, 10/01/2012 - 8:37pm

Question: do we have data on the success rate of teams going for it on 4th down while winning (or even tied) late in the 4th quarter?

It seems to me that this situation might be qualitatively different than general 4th-and-1s in that the defense can completely sell out to stop a QB sneak/running play because a long TD is often no worse for them than a 1-yard gain.

FWIW, the plays that fall into this category that I can remember generally haven't worked, although this could simply be a function of the availability heuristic on my part.

16
by MJK :: Mon, 10/01/2012 - 10:26pm

Good point. It brings back to mind Belichick's "infamous" 4th and 2 decision. After the game, the Colts commented on how they completely sold out to control the short field, because a deep pass was just as damaging as a 2 yard pass (either one would end the game). So their chance of stopping the conversion was considerably greater than the average for 3rd or 4th and 2.

18
by BJR :: Mon, 10/01/2012 - 11:12pm

If it's the effect of the defense selling out you want to isolate, why not just look at all goal-line plays?

17
by Joseph :: Mon, 10/01/2012 - 10:50pm

1. To me, the counterpoint to comments #14 & 16 above is that if you can't pick up that 1 yd, you don't deserve to win.
2. Considering the relative strengths & weaknesses of the Panthers' offense and defense, this just isn't debatable. CAR's offense is definitely their stronger unit, and the Falcons offense is their stronger unit.
[Now, if there was less than 1 minute, or the Falcons had no timeouts left, then I can see the point. Tell the punter to make sure to get it out of there, and try to kick it high and to the side.]

19
by jebmak :: Mon, 10/01/2012 - 11:49pm

Loved this reply in response to random idiot bashing stats:

"You're not a fan of science and facts? Your "gambler's instinct" is a synonym for uninformed and ignorant opinion. Stats aren't a manipulation of the world, they represent it. Your complete inability to grasp scientific concepts is frightening. I'm taken aback, astounded, dumbfounded. This is what someone would write if they were trying to appear stupid on purpose."

22
by Sophandros :: Tue, 10/02/2012 - 8:47am

How about McCarthey's decision to run a fake punt from deep in his own territory?

-------------
Sports talk radio and sports message boards are the killing fields of intellectual discourse.

24
by DEW (not verified) :: Tue, 10/02/2012 - 10:36am

I'm further appalled by Mike Golic and Cris Carter on ESPN this morning sitting there and saying Rivera was "playing the percentages" by punting. There are arguments to be made for punting the ball, shaky as those arguments may be. Probability theory is not one of those arguments! The sheer ignorance of not merely how to *apply* statistics to the decision-making process but of what the math actually *says* from these ex-players tells me more about why coaches make the foolish conventional-wisdom decisions than anything else. It's not because they're risk-averse with regard to the negative consequences to their own job of bucking conventional wisdom and failing, it's because they actually believe that the conventional wisdom is in fact wise!

26
by dbostedo :: Tue, 10/02/2012 - 1:11pm

I took a couple of things from hearing that this morning :

-- Here is what those two ex-players think "playing the percentages" means : If you go for it, the game result rests on one play. If you don't go for it, there's a punt and a few more plays. "A few" is greater than 1, therefore, punting is playing the percentages.

-- I think coaches are obviously risk averse, and I've always assumed/read it's likely because of execs and media and bowing down to conventional wisdom. What I hadn't considered, is that if all of your players buy into the conventional wisdom too, then you might be in trouble from credibility with your own players. This is of course related to their failure to understand/implement "playing the percentages".

For instance, in the discussion this morning they immediately jumped to mentioning times when going for it on fourth down failed. Presumably, the players feel better if the end result is dragged out because they have more chances to do something about the result. (And every team probably thinks the really should be able to stop the other team, no matter what the percentages actually say.) So it's psychological regardless of what the math actually says. In this case, some of the players might even understand the math... but it's awfully hard to get past your gut feel even then.

I know this is probably just recycling where the conventional wisdom comes from, but I, personally, hadn't considered the players reactions in this; I figured it was pretty much entirely that the coaches are afraid to take risk.

29
by Pat (filler) (not verified) :: Wed, 10/03/2012 - 6:13am

I'm sorry, I just find this attitude amazingly arrogant. The math is not as clear as you're making it, and Golic and Carter have a point.

1) The outcome of the plays aren't random: it's a game. It's being played, and in fact, there's strategy involved between the coaches, and every new play is a chance for a coach to exert a strategic advantage over another team. A coach who thinks he sees an advantage he can exploit shouldn't risk the entire game on one play. He should extend the game as much as possible to allow that advantage to build.

2) Similarly, strangely, if the team is noticeably better (on both offense and defense) than its opponent, they should probably punt. This sounds completely counterintuitive, but it's not: the variance on 1 play is huge. The variance on the result of 10 plays is less, and if the team's better, well, that means that the average result is good.

Giving up a bit of the average win probability to reduce the variance isn't necessarily a bad thing.