Writers of Pro Football Prospectus 2008

22 Jul 2010

The Steelers Struggled In Short-Yardage?

Mark Kaboly of the McKeesport Daily News has an interesting story on the Steelers' running game up that reader DrObviousso pointed us to.

The piece begins by quoting Bruce Arians in noting that the problem with the Steelers' offense isn't that they don't run the ball frequently enough, but that they weren't running the ball effectively when they needed to. It isn't exactly a refutation of run-to-win, but it's better than the alternative.

"When they needed to", unsurprisingly, is in short-yardage situations. There, Kaboly quotes statistics that say that the Steelers were 25th in the league in short-yardage situations, converting only 60 percent of the time. However, our "Power" statistic notes that the Steelers on 72 percent of power runs, which was fifth-best in the NFL a year ago. What gives?

Well, the way the two statistics are put together, for one. Kaboly defines short-yardage to be "third down and three yards or less", which seems to be a very strange way to define such a statistic. Third-and-3 isn't exactly a short distance to go, and Kaboly notes that the Steelers didn't run the ball once in that situation all year. NFL teams ran a pass play nearly 81 percent of the time on third-and-3 last year, so it's strange to include that distance in the analysis.

So on third down with two yards or fewer to go last year, the Steelers converted 15 of their 25 carries for first downs, which yields that 60 percent figure. Kaboly's short-yardage figure doesn't include other obvious short-yardage situations, though, that our figures do. And there, the Steelers aren't so bad.

On fourth down with two yards to go or less, the Steelers ran the ball seven times, with all seven chances coming from a yard out. They converted six of them. If you include these fourth down plays for all teams, the Steelers improve their short-yardage conversion rate from 25th in the league to 12th.

Now, throw in those other situations we consider to be short-yardage plays -- when the Steelers are running from two yards out of the opponent's end zone or less on first and second down. The Steelers had six such carries in 2009; they scored on every one of them. After we include carries in this situation for every NFL team, the Steelers have a conversion rate of just under 72 percent (28-of-39). That's fifth-best in the NFL.

In short? Any assertion that the Steelers were poor in short-yardage last year involves some serious statistical gerrymandering.

Posted by: Bill Barnwell on 22 Jul 2010

17 comments, Last at 25 Jul 2010, 5:56pm by Hurt Bones


by Raiderjoe :: Thu, 07/22/2010 - 7:38pm

1. is Dr. Obvuiousso related to mr. Roboto?

2. fidn't get FOA 2010 yet (9will place order in nexyt day or other day) so didnt see stats but going from memory of 09 season and stats B. Barnwelll post now streelers didnt seem like crsppy short yardage team

3. Fianl sentnece make wonder if Jerry mathers ever did gerrymanderiing .

by Vincent Verhei :: Thu, 07/22/2010 - 9:19pm

This is my favorite Raiderjoe post ever. For many reasons, but mostly "crsppy."

by dbostedo :: Thu, 07/22/2010 - 10:39pm

The whole "3rd and 3" thing seems like a bit of a throwback choice. I'd guess if you asked long time Steelers' fans (being one myself) a lot of them would just assume that you should always run on 3rd and 3, and should be able to convert most of them. I'm also guessing a lot of fans would be very surprised at how often all NFL teams throw on 3rd and 3.

by Joseph :: Fri, 07/23/2010 - 12:49am

I would bet that the teams/offenses who run best from that situation like to spread out the D with 4 wide, then run a draw/shotgun handoff to a RB. Tackles push DE's out wide to give good lanes--back picks his hole. HOWEVER, the short pass to a Wes Welker-type receiver must be a viable threat. Seeing the Bills, Browns, et al line up 4 wide won't scare most defenses--but the Saints, Packers, Pats, Warner-led Cards, Chargers--now those teams can cause coverage problems, thereby giving a very credible pass threat, allowing a 3rd & 3 run to be effective. Do it about once per game, and teams have to prepare for it. That 81% pass/19% run sounds about right--1 of 5, or about once per game. I'd love to see some stats on 3rd & 3/4--NFL overall percentage (run/pass, conversion rate & DVOA for each), as well as, say, top & bottom 5 teams. Sounds like an interesting article.

by Nathan :: Fri, 07/23/2010 - 2:56am

i dunno, i'm of the mindset that if you have a decent qb you should run on 3rd and 3 only enough to keep the defense honest and even then it should be a draw. i have no data to back this up but 3 yards on a run seems 40%. if you can complete a pass 3 yards is cake. and most good qbs are around 63-65%. no brainer.

by Brendan Scolari :: Sat, 07/24/2010 - 9:29am

IIRC the median run in the modern NFL goes for 3 yards (sorry, no link, but I remember reading it at Advanced NFL Stats), so I think there must be at least a 50% chance of converting without adjusting for the situation (which is obviously a big factor).

Regardless, if you read this study:


It seems obvious that teams should be running the ball a heck of a lot more on 3rd down. As you can see passing has a lot lower average payoffs than running currently in 3rd and 2/3 situations. I think if coaches we're using a more rational 4th down strategy (ie: going for it a lot more often) it would only become more obvious. Keeping a team from gaining 3 yards on 2 running plays is very hard to do.

by Jerry :: Fri, 07/23/2010 - 7:05am

3rd and 3 is a passing situation in today's NFL. Coming out in an empty set on 3rd and 1, though, still bothers me eight months later.

by Israel P. (not verified) :: Sat, 07/24/2010 - 4:14pm

Doesn't TMQ have a standard line about the average running play in the NFL's gaining whatever, so why don't people do it? Third and three should fit that mold, at least enough to be a threat.

by Bobman :: Sat, 07/24/2010 - 6:48pm

Been a while since I read TMQ regularly, but IIRC it was along the lines of "the average pass is 7 yards and the average run is 4 yards, so you chould pass in this situation."

Then he turns off his computer and goes home to his wife of 13.7 years and his 2.4 children who get exactly 75 on every test they have ever taken in school.... Amazing how averages work, Mr. Easterbrook.

by Joe Don Looney (not verified) :: Fri, 07/23/2010 - 9:53am

I don't have the numbers for goal to go from the 2 or less and on early downs. For the others:

The Steelers didn't run at all on 4th and 2; 0 attempts.
On 4th and 1, their RB didn't get the ball. All 7 runs were QB sneaks.
Overall: 3rd and 1 or 2, 4th and 1): QB converted 10 of 11; RB converted 11 of 21.

This does not describe a powerful run game.

by Eddo :: Fri, 07/23/2010 - 10:10am

It might describe a bad power run game, but it might not. How does the rest of the league do in those situations? Do running backs, on average, convert less than half their runs on 3rd and 1 or 2 and 4th and 1?

by ChaosOnion (not verified) :: Fri, 07/23/2010 - 10:03am

On 3rd and 3 or less, when PIT opted to run the ball, they only succeeded 60% of the time.

by theprophecy :: Fri, 07/23/2010 - 10:51am

I think the conclusion ("serious statistical gerrymandering") is a bit overstated. Realistically this is laziness (picking one stat and running with it) and a general lack of knowledge of the growing number of ways football can be measured. To assume "serious statistical gerrymandering" would be giving the author just way too much credit. If he was being overspecific with a broad range of stats while ignoring more obvious ones, that would say sleazy intent a lot more than this, which is just a lazy beat writer grabbing one easy stat and spinning an article out of it to get paid.

by dmb :: Fri, 07/23/2010 - 12:19pm


by Brendan Scolari :: Sat, 07/24/2010 - 9:41am

Agreed as well.

by Bobman :: Sat, 07/24/2010 - 6:51pm

I always think of gerrymandering as changing election districts, so the 3rd and short plays would be voting along with the 1st and ten plays next year, while they voted with the 2nd and six plays a year ago. And now the 1st and ten plays vote with the ever-so-rare quick kicks on 3rd down.

by Hurt Bones :: Sun, 07/25/2010 - 5:56pm

Or maybe in football gerrymandering the line of scrimmage looks an ekg. So it would only be 3rd and three straight up the middle, but 3rd and inches off left tackle and 3rd and 8 off right tackle.