clock menu more-arrow no yes mobile

Filed under:

Examining Buffalo's Heavy-Front Run Defense

In a Week 5 loss to the Jacksonville Jaguars, the Buffalo Bills unveiled a new "heavy" run defensive front designed to help shore up the league's worst run defense. Though the Jags still ran for 216 yards in a 36-26 victory, the heavy package will apparently remain a part of coordinator George Edwards' defensive arsenal moving forward.

After breaking down the numbers from two of Buffalo's worst run-defending performances of the 2010 season, we're not precisely sure why the package will continue to be utilized.

In order to properly express our argument here, we decided to break down film of the Jaguars game, as well as the Bills' Week 3 loss to the New England Patriots, when the defense allowed 200 rushing yards playing the vast majority of the game with a four-man front. Sure, Jacksonville and New England play different styles of offense, but they're still among the league's best rushing offenses, as Jacksonville currently rates No. 6 averaging 135 yards per game, while New England is No. 12 at 123.2 yards per game.

In short: going from a "light" four-lineman look in New England to a "heavy" four-man look against Jacksonville really didn't change anything at all.

Week 3 in New England
As the Patriots operate out of multiple wide receiver sets so frequently (especially then, when Randy Moss was still on the team), Buffalo spent much of Week 3 playing out of defensive packages featuring four down linemen. It was in this game that Aaron Maybin saw his most extensive playing time, getting over 30 snaps as a defensive end along a four-man line.

Chris Kelsay also spent the vast majority of his time at defensive end. Dwan Edwards saw a handful of reps at end, as well - particularly on early downs - but he, along with Kyle Williams and Marcus Stroud, spent the vast majority of their time inside at defensive tackle positions. Torell Troup? He barely saw the field. Alex Carrington? He was inactive. At linebacker, the Bills were without Paul Posluszny, so they spent most of their time with Keith Ellison and Bryan Scott on the field in traditional linebacker roles (a factor in the team's lack of physicality against the run, to be sure).

New England ran the ball 38 times in that contest; that total includes three Tom Brady kneel-downs to end the game, so really, they ran the ball 35 times. Of those 35 rushes, 21 - yes, that's 60% of New England's runs on the day - gained four or more yards. Clearly, after that type of performance - and particularly after the one that followed, a miserable Week 4 effort against the Jets - some sort of tweak needed to be implemented. Ideally, that tweak would lower the number of rushes that gained four or more yards. Right?

Hence, the heavy set against Jacksonville.

Week 5 against Jacksonville
The Bills said so long to their basic sub-package four-man look, and instead unveiled a new four-man look featuring Troup and Williams at defensive tackle, with Edwards and Stroud at defensive end. Carrington and Spencer Johnson rotated in at end and tackle, respectively. Kelsay, who spent most of Week 3 with his hand in the dirt, saw extensive action as a stand-up outside linebacker in this 4-3 look, with Andra Davis the "middle" linebacker, and Posluszny an "outside" linebacker (though, from a terminology and technique standpoint, he was still an inside linebacker).

Nothing major changed from an expectations standpoint. The ends, and Kelsay, were expected to set the edge. The tackles (with the exception of the one-gap Williams) and an end were to try to keep Davis and Posluszny clean. The defensive scheme didn't change much; it's just the personnel that saw a rather radical shift philosophically.

Not much changed from a run-stuffing standpoint, either. In fact, you could argue things got worse.

Buffalo did improve in terms of the number of runs allowed that gained four or more yards, as 20 of the Jaguars' 40 rushes fit that description. That's a 10% improvement in one attempt, but really, should the team be rewarding a defense that allows 50% of runs to gain four or more yards with more game action? Won't teams figure out how to beat this defense, too?

Here's another fact worth considering: Jaguars rushes by non-running backs (meaning Mike Thomas and David Garrard) gained 62 yards on six carries. Go figure - when you get bigger and slower in your front seven, teams will have a lot more success getting to the edge and ripping off big chunks of yardage on the ground. In fairness, the Bills weren't great in that area in New England, either, as tight end Aaron Hernandez (!) ripped off a 13-yard run to spearhead a seven-rush, 25-yard effort from New England's non-running backs - a group that included Brady, Hernandez, Brandon Tate and Julian Edelman.

It's fair, we suppose, to give the defense another shot against a team purporting to be a physical, run-first outfit like Baltimore - if only to see if they can shave another 10% off of their over-four-yard total. But at what point does trying, and failing with, an experimental defense become detrimental to the implementation of the defensive scheme you actually want to run?

We talked about the fact that the scheme didn't really change, but let's face it - Spencer Johnson isn't going to be setting the edge in an OLB role for us in any sort of future worth watching. If the Bills want to continue to use their heavy set early in games, so be it - NFL coaches are, after all, still trying to win football games. But if a contest is out of hand, it would be ridiculous if the team did not revert back to its base 3-4, with actual linebackers playing outside linebacker, and with young players playing the techniques they'll be playing over the long haul. When a team is re-building, particularly in year one of a new regime, the first year should be a "scheme" year - get your systems in and run them into the ground. Get your players deeply immersed into the terminology, into the execution, and to improving each week within that scheme. Experimentation is fine, but it seems rather pointless when it's not working.

But maybe that's just us.