September 30, 2016

Itty Bitty Running Back by Committee

By Bill Barnwell, Football Outsiders – special to BSMW Patriots Game Day

Week 3 brings the Patriot-vanquishing Denver Broncos to Massachusetts on Sunday night. While I’d love to provide some numbers correlating playoff performance to ref-pocketing for the BSMW readers, one playoff game is exactly that: one game. Had the Patriots played more like… the Patriots in that game, they easily could have won it. But I’m sure you already knew that.

What the Broncos mean to me is something entirely different: fantasy football agony. Anyone who plays knows what I’m talking about — the Broncos have essentially made running backs entirely fungible as part of their offense, mixing and matching everyone from Clinton Portis to Ron Dayne into their schemes over the last few years. This is often referred to in the media as a running back by committee, which is a misnomer in much the same sense that the Red Sox didn’t operate (or intend to operate, I would imagine) a “closer by committee” in 2003. The Broncos don’t use a multitude of backs each year for a significant number of carries each; they do, however, not worry about replacing a previously successful back with another one, placing their trust in their scouting abilities and that the pieces surrounding the running back — his offensive line and the quarterback — will remain solid enough to make his life that much easier.

While the Broncos don’t actually run a committee at tailback very often, the concept intrigues me, and I don’t think I’m the only one. Almost every time a running back committee rolls around, they usually get a nickname (most recently “Thunder and Lightning” for either Tiki Barber and Ron Dayne or Warrick Dunn and TJ Duckett; most recent running back combinations seem to be based on “Of Mice and Men”, with the heavier back often ending up like Lenny) and media attention. What I wanted to determine was relatively simple: do teams that use multiple running backs perform better than teams that don’t?

To do this, I employed an old methodology from the Bill James’ Abstract days. Using the criteria I’ll list in a second, I located 41 teams that used what I considered to be a running back committee. For each of those 41 teams, I gave them a buddy. Those “buddy” teams had the most similar rushing statistics to the team employing the running back by committee in the year, but did not operate the running back by committee themselves. For example:

RBmatch.jpg

As you can see, these teams have eerily similar numbers — they’re the closest match between two teams I was able to find. The 49ers employed Roger Craig and Wendell Tyler at halfback that year, each of whom carried the ball 176 times; St. Louis, meanwhile, had Ottis Anderson. Ottis had 296 carries, while no one else on the Rams had more than 75. As you can see, both teams did reasonably well for themselves.

To determine which teams used a running back by committee, I found those 41 teams from 1978 (the advent of the 16 game season) through 2005 that fit these rules:

  • Had two backs who played more than 12 games (to eliminate the strike season as well as situations where a back got hurt and another took over, since that’s not what is being analyzed here)
  • Had those two backs carry over 100 times each
  • Had the second back carry the ball 90% (or more) as often as the starter did

So then, the Dunn and Duckett committee doesn’t make it (since Duckett never had the required percentage of carries to qualify, outside of the one year Dunn was injured and missed several games), but the Kevin Faulk and Antowain Smith pairing from 2003 does.

The results were a little surprising.

results.jpg

Whew, that’s a big ol’ chart. I highlighted that 1986 Patriots’ season solely in amazement as to how awful it was — it’s unfathomable to me that a team couldn’t even average three yards (and the requisite cloud of dust) a carry over an entire season, let alone a team that won eleven games! It wasn’t just one back who held the team down, either; Craig James averaged 2.8 yards per carry, Tony Collins 2.6, Mosi Tatupu 2.4… it was a mess. Really, Stanley Morgan was the 2005 Steve Smith of that team. Eleven games!

The research found that teams who employ a running back committee perform very slightly, but not significantly, better than those teams who get the bulk of their yards from one back. What this means to me, then, is that as long as you can get your rushing yards from somewhere, it doesn’t really matter whether it’s from one running back or two; that’s a very useful data point when determining whether to trade for or sign a star running back (I’m looking in your direction, Arizona).

The next question, then, is whether the running back by committee approach is more cost-effective than going single-back. The USA Today Salary Database has information ranging back to 2000, which means we can look at the compensation paid by the final six teams in the study. I included all the running backs for each team, so we can determine what the overall spending on the position is to achieve the (similar) rushing results.

pricetag.jpg

That really surprised me. Of course, there are a million and one accounting tricks when it comes to NFL teams and the salary cap, but when you judge the players’ compensation based on the total salary they earned in that season (as opposed to their cap hit), the committee was more expensive every year except for one — when the Pats out produced Eddie George & company for over $1.3 million less. While a larger sample might show different results, this is a small data point in favor of foregoing the committee route in order to commit to one back and paying his backups significantly less than, say, the second and third backs in a committee might make.

Unfortunately, this study doesn’t provide any quick and dirty conclusions as to what the better model for building a staff of running backs is. What it does show, though, is that the difference between the two, when it comes to winning, is minute and dependent upon the other factors that make up a team.

Leave a Reply