Who’s Behind Joseph Randle in the Dallas Backfield?
This won’t come as a surprise to those reading this website, but the Dallas Cowboys have a hole to fill at the running back position. The Cowboys chose not to resign DeMarco Murray during the offseason, and that decision led to the loss of the NFL’s leading rusher in 2014 and 392 of the team’s 472 carries from a running back last season.
Joseph Randle has been the assumed starter and using our Net Expected Points (NEP) metric, that doesn’t seem like a truly terrible idea. NEP factors in on-field variables such as down-and-distance in order to compare a team or player’s production to historical expectation levels./p>
Over his 51 carries, Randle had a Rushing NEP per attempt of 0.11, which tied for fourth among running backs with at least 40 carries. Randle, obviously, is and was not the fourth best running back in the league, but there were moments of value when he has given the ball. But even if Randle is expected to be the starter, his previous career workload and numerous off-the-field make it unlikely he sees the same type of carry split Murray had in 2014.
That potentially leaves a chunk of attempts available for non-Randle Cowboys running backs, almost all of whom are not being talked about in regards to the 2015 Dallas running game.
Many would agree Dallas features one of the best offensive lines in the league, and the replaceability of running backs has been debated over the past few years. Many would also agree they would not expect a team run by Jerry Jones to have that experiment play out on the field. But this is where we are, wondering if these other backs can help replace last year’s star.
If there’s such a thing as an incumbent backup, that’s what Lance Dunbar is for the Cowboys. Entering his age-25 season, Dunbar will be playing his fourth year in Dallas. Over his first three seasons in the league, Dunbar has not carried the ball more than 30 times or fewer than 21 times in a season. Those types of small samples allow for the perspectives on Dunbar to vary greatly. In his rookie season, he averaged only 3.6 yards per carry. That bumped up to 5.0 yards per carry in 2013 before falling back to 3.4 last season.
Judging running backs by yards per carry in a given season, especially over such a small amount of carries can lead to some false conclusions since yards per carry takes a really long time to stabilize. That’s also something to keep in mind when citing Randle’s 6.7 yards per carry last season.
Luckily, though, this type of small sample makes it an undaunting task to go back and watch how Dunbar performed on the field. Dunbar appears to be a player who is going to take what the defense and his offensive line give him. Take the following plays from this past season.
In the play above, there’s a well-blocked hole from the offensive line, but the blocking of his fullback fails him. Tyler Clutts (44) couldn’t contain his defender, and Dunbar was dropped for only a three-yard gain. Dunbar had 11 of his 29 carries during this Week 2 game against Tennessee, before both the Cowboys and the Titans became would they would be during the season. The day was not kind to Dunbar, who picked up just 27 yards on those 11 carries and was on the wrong end of plays like this.
Then take plays like below, when he easily gains 13 yards on a wide open hole. This is from Week 7 against the Giants.
There are also some times when Dunbar could make a slightly better read given what the offensive line has given. This play is from Week 15 against the Philadelphia Eagles. Instead of making the cut to a wider portion of the field, Dunbar bounced off his blocker and was tackled by Malcolm Jenkins. Jenkins was drifting towards the more open side, so it wasn’t an awful read, but a cut still could have given Dunbar a better chance to make a move with more room. Still, the play went for a seven-yard gain -- certainly not bad.
Dunbar did not fare well by NEP, totaling -0.20 Rushing NEP per attempt last season, though his Success Rate -- the percentage of plays that contribute positively to NEP -- was higher than that of Randle’s last season. However, neither Randle’s 35.29 percent or Dunbar’s 39.29 percent Success Rates are anything to be proud of. With more carries, Dunbar could be just as close to his 2013 production as 2014. He’ll be likely to get at least a few more this upcoming season.
Darren McFadden and the Rest
We mentioned Jerry Jones going into the season with unproven players at running back to be very un-Jones-like. Signing Darren McFadden during the offseason was a perfect Jerry Jones move, just a decade late. Instead of getting McFadden in the 2008 Draft, Jones settled for teammate Felix Jones 18 picks later.
McFadden was pegged as a dynamic runner coming out of Arkansas but never lived up to the hype in the NFL. Even Jones and the Cowboys only gave McFadden $3 million over two years as a free agent, only $200,00 of which is guaranteed. There is some reason for that. Last season, McFadden had a Rushing NEP per attempt of -0.15, better than only nine other running backs out of the 79 with at least 40 attempts. It’s unlikely at age-27 MacFadden will improve greatly, even after getting out from behind the black hole of an offensive line in Oakland.
Ryan Williams spent all of 2014 on the Cowboys practice squad after three years in Arizona. He also could be the odd man out for a roster spot in training camp, according to some reports, as he’s been experiencing some setbacks with a surgically repaired knee.
An interesting player could be Lache Seastrunk, who was signed by Dallas in June. The former Baylor Bear could be one of the most talented backs on the roster as a pure runner, but a mix of maturity and coachability concerns caused Seastrunk to spend his rookie season on three different practice squads in Washington, Carolina and Tennessee.
While there was so much hype surround Murray’s accomplishments last season, Dallas only finished ninth in Adjusted Rushing NEP per play. That’s still good, but not the best, and it’s realistically repeatable. None of these players may be the single answer, but a rotation of these backs could be just enough of an answer to make realistically repeatable certainly attainable.