Eight weeks ago, the Vikings were undefeated and ranked No. 1 in DVOA. Now they're 6-6 and probably done. What went wrong?
09 Jul 2012
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 run defenses based on the number of running backs they faced. (We did offenses two weeks ago.)
The big difference between the offensive stats and the defensive stats against multiple-back sets is the lack of definitive conclusions to be drawn from them.
Take last season. The three teams that saw the most multiple-back sets used against them were St. Louis, Indianapolis, and Cleveland. Simple narrative, right? Bad teams faced more multiple-back sets as teams tried to run the clock out against them. A little over halfway through the season, we ran a chart about the teams that spent the most game time in the lead, and those three teams spent the most time trailing. The only other teams that saw multiple-back sets 55 percent or more of the time, Seattle and Carolina, were also losing teams.
However, go back to 2010, and you'll find the teams that faced the most multiple-back sets were New Orleans, Seattle, and Cleveland. That's a pair of playoff teams, though one of them of course had a losing record in the regular season. The interesting thing is that St. Louis and Indianapolis finish above 60 percent again in 2010, along with a couple of teams you probably wouldn't expect in Atlanta and Philadelphia. On a per-play statistical basis, there doesn't seem to be much reason to target those teams heavily. Nor is there much precedent for this as a purely divisional thing -- if it were, we'd expect to see Tennessee in the mix because of Houston and Baltimore, and we wouldn't expect to see St. Louis at all since Seattle and Arizona haven't shown much of a bias and the 49ers only truly embraced multiple-back sets last season. My anecdotal guess would be that it has to do with perceived personnel weaknesses at linebacker. Most of these teams have had a solid man in the middle (D'Qwell Jackson, James Laurinaitis, Jonathan Vilma, Gary Brackett), but the outside positions have been in flux for the past few seasons.
It certainly appeared that offenses were targeting these teams smartly, as four of the five teams that saw the most multiple-back sets in 2011 saw a decline in DVOA when facing them. Seattle actually was almost 15 percent better against multiple-back sets. However, they were slightly worse against it (~4 percent) in 2010. Perhaps old information was relied on?
The Eagles, weak linebackers and all, were the worst team in the league against single-back sets. That shouldn't be a surprise to their fans and they'll move forward with DeMeco Ryans and Mychal Kendricks hoping they've addressed that problem. The second-worst team against single-back sets, Tennessee, dealt with some rookie mistakes from Akeem Ayers and a half-season of a turnstile middle linebacker in Barrett Ruud. A full season of Colin McCarthy should help remedy half of that problem. The funny thing is that both Philadelphia and Tennessee were both much better against multiple-back sets. In fact, last year no team improved their run defense as much as those two did with multiple-back sets on the field. Perhaps that says something about the value of having an every-down linebacker who can tackle?
The teams that saw the fewest multiple-back sets also tend to clump together a bit. In 2011, the teams that saw under 40 percent of runs against them come from multiple-back sets included Dallas, Miami, Kansas City, and Oakland. In 2010? Miami was the only one, but if we scale back a little bit to hit 45 percent, we find Washington, Kansas City, New England, the Jets, and Baltimore. Again, this seems to tie into a trend we pointed out in our play-action defense post: the 3-4 defense. Every one of those teams besides Oakland is at least a hybrid defense, if not a pure 3-4.
Below are all the running statistics we keep for single- and multi-back formations, sorted by the difference in DVOA allowed between single-back and multi-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 Defense by Number of Backs|
|1 RB Pct.||2+RB Pct.|
|1 RB Pct.||2+RB Pct.|
|1 RB Pct.||2+RB Pct.|
6 comments, Last at 16 Jul 2012, 12:26pm by chemical burn