Writers of Pro Football Prospectus 2008

08 Oct 2013

Pleading the Fourth

Hooray! Another article about going for it on fourth down, with which to poison Mike Smith's brain! But this one is armed with years of Football Outsiders data, and it breaks down the plays which have been most successful on 4th and short. That way, you can figure out what Mike Smith is doing wrong (besides failing to mount a pass rush or hire defensive backs who can tackle or developing decent special teams units or ...)

Posted by: Mike Tanier on 08 Oct 2013

12 comments, Last at 10 Oct 2013, 7:24pm by Mr Shush


by foxlies (not verified) :: Tue, 10/08/2013 - 8:35am

perhaps not my favorite subject or type of coaching criticism;but great title;read it even before i knew it was u....nflalternative

by foxlies (not verified) :: Tue, 10/08/2013 - 8:35am

perhaps not my favorite subject or type of coaching criticism;but great title;read it even before i knew it was u....nflalternative

by Aaron Brooks Go... :: Tue, 10/08/2013 - 10:06am

I think they fail to note the distance-from-goal line trend accurately at the end, because they are blinded by their own narrative.

The red zone is harder for the same reason the goal line is harder -- because the defense has so much less vertical space to defend, it can deal with the short stuff much easier (i.e. -- there is no long stuff).

Come on guys -- this is young Al Davis-era stuff. We were still running cars on kerosene back then and playing football with human skulls.

by Bobman :: Tue, 10/08/2013 - 4:09pm

Aaron, youngster, you're showing your youth. I ran my car on dried mammoth dung and played football with Neandertal skulls.

by Crunch (not verified) :: Tue, 10/08/2013 - 10:24am

I'm not sure that the choice to exclude 4th and 1 and 4th and 2 plays when the attempting team was down by 14 or more is a justifiable scrubbing of the stats. It's seems like a distortion of the sample.

by panthersnbraves :: Tue, 10/08/2013 - 10:26am

With the advent of more mobile QB's I wonder if "Line up 5 wide and call a QB draw" will become its own category soon.

by Aaron Brooks Go... :: Tue, 10/08/2013 - 10:39am

I swear Brees runs that play occasionally. I think I've seen Philly run it w/ Vick, too.

by Jake (not verified) :: Tue, 10/08/2013 - 11:17am

As far as "engineering" the qb sneak goes, I think we don't see much of that because it isn't always the prescribed play call. Since only the qb and center really need to know it's a sneak, I know a lot of teams give the qb the option to audible by pinching the center's butt on the side he wants to sneak to. You can call something else then switch to it if you like how the DTs are lined up. This might also explain how unstoppable the play is; if you mainly call it based on a defensive look that can't stop it, you'll pick up a high percentage on it.

That being said, I really don't like lining up power-I to sneak it. You're making the defense stack up against extra backs and TEs that won't even help you on the play. It sure seems to make more sense to line up in 11 personnel and force them to cover more ground with smaller bodies.

by dan s (not verified) :: Tue, 10/08/2013 - 11:23am

I'm not sure the data backs up the "engineering" point...doesn't mean it's wrong, I just didn't see it in the data. It feels intuitively right, though.

Great article overall...fantastic work. I wish there were a few more football writers out there of this caliber. It's rare to see great stats work, but even rarer to see great stats combined with a writer with enough wisdom to interpret the stats right. For example, the line about not slicing/dicing the sample up too much because of limited sample size? Awesome. I loved the asterisks by the stats with limited sample sizes, too. I like the idea of signaling to the reader, right next to the number, that you need to be careful.

And the idea of not just stopping at the numbers (you convert 2/3 of the time), but then going further and asking which plays work best? Lovely. Stats + thoughtful Xs and Os = the best.

by Scott Crowder (not verified) :: Tue, 10/08/2013 - 12:02pm

Great research, but was kinda disappointed it didn't go further and take the Risk/Reward factor considering field position into account. Obviously, going for it on 4th down deep in your own territory on a QB sneak, even with 80% success rate, is a bad idea if 20% of the time it's likely to put points on the opponents board. But at what field position does going for it begin to break even or get the offense ahead? Since this factor is the defining factor in making the decision, adjusting success rate for the Risk/Reward factor of field position seems an integral part of the study and I was hoping it was part of what I was about to read.

by Pat (filler) (not verified) :: Tue, 10/08/2013 - 1:10pm

I don't think the article was attempting to look at the decision to go for it on fourth down at all. It was more a breakdown of what the results look like when you've already decided that you should go for it.

Honestly, the decision to go for it is more complicated than most advanced football stats guys think it is, because it's not just "0.8*(win probability if you make it) + 0.2*(win probability if you don't)" versus the equivalent for punting. That's the *mean* win probability, but it doesn't even have to be the *most probable* win probability (and it probably isn't). When you're already likely to win, adding any risk doesn't make much sense, for instance.

by Mr Shush :: Thu, 10/10/2013 - 7:24pm

That being said, I imagine there are very, very few circumstances where you should not go for it if you have an 80% chance of converting. Late and two or more scores up, maybe.