Stat Analysis
Advanced analytics on player and team performance

2011 Formation Analysis: Number of RBs Part I

2011 Formation Analysis: Number of RBs Part I
Photo: USA Today Sports Images

by Rivers McCown

Continuing our look at game-charting statistics from the 2011 season that'll be featured in Football Outsiders Almanac 2012, this time we'll examine the tendencies and performances of running offenses based on the number of running backs they used. We'll hit defense in Part 2.

One of the biggest changes that happened last season occurred in Indianapolis. And for once, we aren't talking about losing Peyton Manning! No, as Mike Tanier documented in an earlier Walkthrough, the Colts (*gasp*) managed to actually break out the I-Formation from time-to-time as their season spiraled into the abyss. After using formations with multiple running backs on just five percent of snaps in 2010, they were all the way up to 25 percent in 2011. Despite that, the league continued to move more towards single-back sets, and for the first time since we started accumulating charting data, more teams ran out of single-back formations (53 percent) than with two or more backs (47 percent).

Of course, just because they used the I-Formation doesn't actually mean they performed better. In fact, the Colts offense was 25 percent less effective according to DVOA when using two or more backs. That was the fourth largest drop in effectiveness, and two of the teams ahead of them used two or more backs less than 10 percent of the time.

Speaking of ... the new kings of the single-back offense are the Detroit Lions. After two years of mixing in the two-back set at a 25 percent rate, a combination of a breakout Matthew Stafford season and a dearth of talent at running back sent the Lions spiraling down to using multiple running backs in just six percent of snaps. As noted, Jim Schwartz was basically using passes to Brandon Pettigrew as a de facto running game last year.

It only makes sense that you'd use two backs if you only had two good ones, and naturally, the teams that ran with just a single back the most often were generally much worse when they brought another on the field. The lone exception to that is Buffalo, which actually saw its DVOA improve by 25.3% with multiple backs as opposed to one. Corey McIntyre earned his extension prior to the 2010 season -- the Bills also had a better DVOA in multiple-back sets in 2010.

The most effective team out of the single-back? Would you believe the New Orleans Saints? As we noted last year, it's fairly rare for a team to average six or more yards per carry out of a single-back set. The Saints joined the 2010 Texans, 2009 Titans, and 2008 Giants as the only teams to break that mark since we began charting. Replacing Reggie Bush with Darren Sproles seems to be working out pretty well in the bayou, even if nothing else currently is.

There were a select few teams that actually ran more out of multiple-back sets than they did in 2010, but the one that really stands out is the Baltimore Ravens. After spending big on free agent fullback Vonta Leach, 80 percent of their runs came out of multiple-back sets. That was ten percent higher than any other team, and 15 percent more than they used them last year. Dallas came in second, and fellow Harbaugh brother Jim put the 49ers in third place by boosting their multiple-back run usage 12 percent. On the other hand, without Leach, the Texans' usage of multiple-backs on runs declined 12 percent.

In 2010, eight different teams ran with multiple-back sets 65 percent of the time. In 2011, that number was down to just three. The Jaguars, Cowboys, and Ravens are the only three teams to crack the top six in both years.

Below are all the running statistics we keep for single- and multi-back formations, sorted by DVOA from single-back sets. This is about formation, not personnel: if a receiver is in the backfield, for the purposes of this study, he is counted as a back. Also, no Wildcat-style runs were counted, meaning we did not include (among a few other rare things) direct snaps to running backs or receivers.

2011 Rushing Offense by Number of Backs
Offense DVOA 1 RB DVOA 2+RB DVOA Difference 1 RB Pct. 2+RB Pct.
NO 38.0% 3.4% 34.6% 38% 62%
CAR 25.4% 7.3% 18.1% 51% 49%
NE 14.0% -31.6% 45.6% 91% 9%
PHI 12.0% 7.0% 5.0% 81% 19%
PIT 9.6% -9.0% 18.6% 65% 35%
JAC 7.5% -11.7% 19.2% 36% 64%
DEN 5.7% -15.8% 21.5% 57% 43%
NYG 0.6% -8.2% 8.8% 42% 58%
SEA -0.4% -2.8% 2.5% 58% 42%
BUF -2.2% 21.3% -23.5% 79% 21%
HOU -2.8% 5.6% -8.4% 47% 53%
Offense DVOA 1 RB DVOA 2+RB DVOA Difference 1 RB Pct. 2+RB Pct.
CHI -3.1% -19.7% 16.6% 53% 47%
IND -3.2% -27.7% 24.6% 75% 25%
BAL -3.2% 2.4% -5.6% 20% 80%
DET -4.7% -41.9% 37.2% 94% 6%
ATL -5.1% -17.1% 11.9% 46% 54%
STL -5.7% -20.3% 14.6% 73% 27%
CIN -7.8% -17.3% 9.5% 48% 52%
NYJ -8.3% -9.7% 1.5% 48% 52%
KC -8.3% -24.7% 16.3% 52% 48%
GB -8.4% 1.2% -9.6% 40% 60%
MIN -8.8% 7.8% -16.6% 47% 53%
Offense DVOA 1 RB DVOA 2+RB DVOA Difference 1 RB Pct. 2+RB Pct.
SD -8.8% 4.7% -13.6% 47% 53%
TB -9.1% -10.3% 1.3% 55% 45%
ARI -9.5% -3.9% -5.5% 50% 50%
OAK -11.2% 7.5% -18.7% 39% 61%
MIA -13.2% -17.1% 3.9% 60% 40%
CLE -13.2% -19.7% 6.5% 56% 44%
WAS -13.9% -1.4% -12.6% 50% 50%
SF -16.2% -10.0% -6.2% 31% 69%
DAL -16.7% -8.8% -7.9% 30% 70%
TEN -17.5% -22.3% 4.7% 41% 59%
NFL -1.8% -7.1% 6.1% 53% 47%

Comments

31 comments, Last at 09 Jul 2012, 10:47pm

1 Re: 2011 Formation Analysis: Number of RBs Part I

"After two years of mixing in the two-back set at a 25 percent rate, a combination of a breakout Matthew Stafford season and a dearth of talent at running back sent the Lions spiraling down to using multiple running backs in just six percent of snaps."

More like a dearth of health. Kevin Smith (a street FA) played with a high-ankle sprain, which made him their healthiest experienced RB. Best was out for the season with a concussion, and Jerome Harrison had a brain tumor. By the playoffs, the Lions RBs consisted of two switch FB/RBs -- one playing injured (Morris) and the other a practice squad reject -- and one undrafted rookie. Williams was their #5 RB. No team has quality depth that far down -- ask the 2010 Saints. The Lions didn't run because they literally had no competent warm bodies remaining to carry the ball after Thanksgiving.

The Lions never ran a two-back set because they never had two healthy RBs at the same time!

24 Re: 2011 Formation Analysis: Number of RBs Part I

All three won a playoff game, in fact, despite none of them having an above average quarterback (once Schaub went down). If you have Brees, Brady, Rodgers or a Manning under centre, an old-fashioned lead blocker in the backfield may indeed not be much of an asset. If you don't, I think it can still have real value.

26 Re: 2011 Formation Analysis: Number of RBs Part I

DEN, PIT, and NO were notably worse with 2 RBs than with one. SF, HOU, and BAL were marginally better. ATL was worse with 2 RBs. GB was better with 2 RBs, but their running game was pretty poor overall, almost as bad as DET's, who didn't carry a FB at all. They all got into the playoffs, but I'm not convinced it was because they carried a FB on the roster for run-blocking.

28 Re: 2011 Formation Analysis: Number of RBs Part I

I'd be interested to see Houston split out into Schaub and Yates incarnations.

It looks to me like forcing defenders out of the box with a strong passing game is more valuable than an extra blocker, but if your passing game is questionable, an extra blocker can help. It should also be noted that early in the season the Texans' primary fullback was James Casey, an average at best blocker but an excellent receiver, while later on they used the more smashmouth Lawrence Vickers with greater frequency.

Edit: A QB who's a strong threat to run probably also makes a fullback superfluous, hence the Broncos.

2 Re: 2011 Formation Analysis: Number of RBs Part I

"No Wildcat-style runs were counted, only plays where the ball was taken by a back or receiver lined up in the backfield."

Which means the numbers for Denver are definitely crap, and the numbers for NYJ likely crap. Denver didn't run the Wildcat; they lined a TE up under center.

7 Re: 2011 Formation Analysis: Number of RBs Part I

I think the Packers have more variety in formations than any other NFL team, but I think they're as likely to pass as run out of any formation. I don't recall any 3 RB sets (with a TE lined up as a second FB) last year, but I probably just missed it. I wouldn't be surprised if they tried the center-eligible play from the movie M*A*S*H.

5 Re: 2011 Formation Analysis: Number of RBs Part I

Some sortable tables would be nice, it would make it easier to sort through the data. If you are going to the effort to provide the numbers why not make it easier to digest?

And do the 49ers 2 back sets include the assortment of stupid formations where defensive and offensive linemen line up in the backfield?

14 Re: 2011 Formation Analysis: Number of RBs Part I

The niners did it soooooo much last year and I don't think it was particularly effective. Nary a 3rd and short came along where Sopoaga, Justin Smith, Alex Boone or Chilo Rachal in the backfield. Some of it made sense, Sopoaga is pretty useful as a lead blocker and Staley makes a decent option at tight end having played there in college. However, Justin Smith might be a really fine defensive end but he's a poor blocker, something he has in common with Chilo Rachal while Bruce Miller is shaping up as a decent full back. Taking all of the receiving threats off the field also invites the defense to stack against the run so that it becomes increasingly difficult to run the ball at all.

9 Re: 2011 Formation Analysis: Number of RBs Part I

I'm surprised that the NE 2 back number is so bad. When do the Pats line up in 2-back except for short yardage? Were they really that bad at short yardage last year? I wonder if there's sample size trouble going on there, after all, how many plays does 9 percent represent?

12 Re: 2011 Formation Analysis: Number of RBs Part I

I find myself more fascinsted by the teams that called runs counter to their DVOA strength: NO, JAC, NYG, BUF, ATL, CIN, and TEN. (NYJ, ARI, and WAS were marginally deviant). Why would these teams call run plays that were counter to their strengths?

29 Re: 2011 Formation Analysis: Number of RBs Part I

There are reasons for this. Like why would NE ever run? (or 80% of the teams).

There are many reasons.
Some are good reasons --
A play that has less value on average can increase the value of other types of plays by adding unpredictability. Play calling makes the value of one side of the coin intertwined with the other.
A play type might be worse overall, but better for some situations (short yardage, for example).

And other reasons are not --
Coaches simply may not know it is less effective.
Coaches might be forcing scheme onto personnel that don't match it.
etc.

17 Re: 2011 Formation Analysis: Number of RBs Part I

Are passing plays mixed into the data? I'd want to see that also, as DVOA on passing plays in those same sets may provide clues about why they were in those sets in the first place.

For instance, New England probably passed a TON out of 1RB sets and this may have made their actual runs much more effective. However, when the 2RB set comes on the field, the pass/run balance is more even and makes the run less effective.

Need more data here.

19 Re: 2011 Formation Analysis: Number of RBs Part I

In this table/article, the runs are compared against the league average (DVOA); not if it made the other situations better or not.
.
What I am curious about though, if the effectiveness is measured against the same personel the league used.
So comparing all 2 backs vs 2 backs in a situation. Not 2 backs vs whatever-the-rest-of-the-league-did.

22 Re: 2011 Formation Analysis: Number of RBs Part I

I wonder if BAL near double the average percentage of runs in 2 back sets serves as a small indictment of what they think of the ravens passing offense? It feels like years they keep saying they want to make this switch from the stoneage style running game they developed in 08 to the wide style passing offenses of today and yet they still haven't done it.

Which brings me to the question of how to rate joe flacco? People tend to feel hes overrated, but the ravens routinely field a pretty poor set of receivers(cosell seemed to content in calling them incapable of winning in man coverage), so would flacco's stats resemble romo, eli, or matt ryan if he didn't play in baltimore?