Five Decisions That Shaped The Cowboys Season, Pt. III: Missing Out On FA Tight Ends

This time last year, it was patently clear that tight end Martellus Bennett's time in Valley Ranch was done. Not only had he been a liability in the passing game over the course of his time at Valley Ranch but, by all accounts, he wasn't the kind of RKG that Jason Garrett or his brother John, then the Dallas TE coach, wanted aboard. So, as Cowboys fans gnashed their teeth over other potential free agent losses, they seemed universally to adopt a "don't let the door hit you in the backside" attitude insofar as Bennett was concerned

In the subsequent free agency period, the organization sought to fill the hole created by his departure by going after a couple of available tight ends with similar CVs: the Bears' Kellen Davis and Seattle's John Carlson. The leading bullet point on these men's resumes? They were plus blockers. Indeed, Bennett never presented much of a receiving threat. In his four years in Dallas, he never caught more than 33 passes and had only one game in which he amassed more than 45 receiving yards.

But that's not the entire story. Although Bennett never became much of a receiving threat, he did develop into a fearsome blocker. Let's track Bennett's (and the teams TE's) progress, using Pro Football Focus' run blocking numbers. Note that the third column is where Bennett ranks among all Cowboys run blockers in a given year:

As a blocker, ‘Tellus clearly improved each year relative both to his team and to other NFL tight ends. In his final year in Dallas (due in part to a substantial drop-off in Witten's blocking grade) the fine gents at PFF found him to be the team's best run blocker - better than Witten or Tyron Smith, better than any Cowboy.

The PFFers are not alone in this. One very instructive way to assess a team's running game is to look at how successful a team runs in each of the traditional running lanes. In the following table, we'll look at the average in each lane in the years since Bennett joined the team (the numbers in parentheses are the league rank):

While the Cowboys have been in the middle of the pack running to the "strong side" in recent years, they had some excellent seasons running to the left. In the three seasons when both Witten and Bennett were at the top of their run blocking games, earning high PFF grades for their blocking (recall that Witten's blocking fell off markedly in 2011), they finished in the top six in the league running to the "weak side," where Bennett often lined up. In effect, Bennett became a third offensive tackle; like most third OTs, who come into the game in short-yardage jumbo sets, he didn't contribute much in the passing game, but was a key ingredient of whatever success Dallas enjoyed on the ground.

Because of this, Jason Garrett frequently trotted out offensive formations with two or more tight ends. The Cowboys used two or more tight ends 582 times in 2009 (57.1% of all offensive snaps), used multiple TEs on 462 occasions (44.6%) in 2010, and lined up with two or more tight ends a whopping 611 times (60.1%) in 2011. Bob Sturm, who has spent the last few seasons "decoding" Garrett's offense on his blog, has repeatedly demonstrated that, from 2008-2011, "12" personnel (one RB and two TEs) has been Dallas' most successful offensive personnel package - even without Bennett presenting much of a threat in the passing game.

But this changed radically in 2012. Losing out on Davis and/ or Carlson meant that Dallas had to elevate third-team TE John Phillips to Bennett's position. While Phillips is arguably a better receiver, he can't compare to Bennett as a run blocker. The numbers bear this out; his 1.1 PFF run blocking grade (a fall off of 10.6 points from Bennett's 2011 mark) placed him sixth on the team and 20th among NFL tight ends. To make matters worse, the draft offered a slew of receiving TEs, none of whom had much sand in his pants. Sixth round draftee James Hannah struggled inline, earning a moribund -1.0 grade. This falls in line with Sturm's data, which began to show that "12" personnel, so long the Cowboys' go-to, was substantially less effective in 2012.

To get a clearer sense of the falloff, lets take a look at where the Cowboys measured up last season:

With the possible exception of Tyron Smith, the Cowboys' running game was bad in 2012, finishing 31st in the NFL - no surprise when you look at these numbers. They ranked fifth lowest in rushing plays of 4 or more yards with 150. The running game improved incrementally late in the season, but, if we take away one flukey DeMarco Murray carry - the 48-yard run in week one at New York - he'd be at 3.8 yards per carry, which is pretty close to Felix Jones 3.6 per. As the above numbers suggest, the Cowboys were particularly poor on the perimeter; as one example, the PFF guys maintain that Dallas averaged 3.0 YPC over left end, a number that pales in comparison to the 5.4, 6.1 and 6.8 YPC gained over left end in 2010, 2008 and 2009, respectively. Left end, of course, is where 'Tellus plied his trade...

In part due to this dropoff, Jason Garrett significantly reduced the frequency with which he deployed "12" personnel. Take a look at the number of snaps Bennett played in his four years in Dallas: 2008 (14 games): 389; 2009: 513; 2010: 495; 2011 (14 games): 428. By comparison, Phillips played only 342 offensive snaps. The Cowboys, who had deployed multiple tight ends no less than 43.1 % of the time from 2008-'11 used it only 33% of the time in 2012.

To be clear: the moment that negatively impacted Dallas' 2012 season was not the decision to let Bennett walk, Rather, it was the team's inability (in an otherwise stellar free agency period) to secure a blocking tight end in free agency, especially in a year in which the draft offered few, if any, answers. Now, as we prepare to bid adieu to Phillips, the Cowboys have another chance to fill this sizeable roster hole. If they don't, I'd expect some more poor rushing numbers in 2013, particularly on the perimeter.