In fantasy football, like most endeavors involving chance, narratives, simple stereotyping, and recency bias reign supreme. While all three help the cluttered human mind simplify and digest myriad stats, opinions, blog-posts, etc, they also confine our thinking to a narrow range of possibilities that are often only a subset of what is truly possible.
Andre Ellington has been cast in cohort of smallish, incomplete running backs (e.g. CJ Spiller and Reggie Bush, both being drafted within 5 spots of Ellington in the late third round of recent MFL10s) that thrive out in space, depend on infrequent, long plays for success, and can’t be effective in goal-line situations. The reason for Ellington’s classification as a small back is superficially obvious: he weighs a little under 200 lbs (i.e he isn’t big enough to carry a full workload), he’s dealt with ankle, toe and knee injuries in college and his short pro career, and he broke off a lot of long plays last year while receiving only a handful of goal-line looks. This small-back narrative is being further inflated with fantasy group-think that lets insignificant data sets reinforce accepted themes like, “Ellington struggles in goal-line situations because he failed to score on three carries within the five yard-line last year.”
I’ve seen this stat a few times on #FFTwitter and I think it’s an example of anecdotes masquerading as analysis, combining an insignificant sample size and an arbitrary 5 yard-line cutoff. (It should be noted that Ellington broke multiple tackles and scored from the 6 yard-line against the Rams last year). We need more robust data to determine if Ellington is capable of being a complete back, and should be involved in goal-line situations. If so, his PPR ADP shouldn’t be hovering in the same range as backs like Spiller and Bush who have shown over multiple pro seasons that they aren’t going to excel on the goal-line. Instead, Ellington should be priced more like Gio Bernard (in the 2nd round of PPR drafts). Or as I will show later, perhaps he should be drafted even higher.
Is Ellington Truly Ineffective In Goal-line Situations?
Rather than basing our assessment of Ellington short-yardage ability on 3 rushing attempts, let’s look back to more extensive data, and compare him to Spiller, Bush, and Gio Bernard (a player going over a round ahead of Ellington) in terms of college non-QB, rushing touchdown-share (nQBTD). Matthew Freedman pioneered the concept of non-QB dominator rating for running backs to control against quarterback production deflating production share, and I think isolating non-QB rushing touchdowns is a useful concept to determine if a running back was capable of being a substantial touchdown producer, which likely involves goal-line work.
|% of nQBTDs||71%||69%||50%||63%|
|% of nQBTDs||33%||35%||52%||40%|
|% of nQBTDs||25%||36%||32%|
|% of nQBTDs||62%||48%||54%|
*nQBTDs excludes TDs scored in games the player missed
Although none of the four were dominant touchdown scorers in college, clearly Ellington and Bernard are in a different category from Spiller and Bush when it comes to TD-share. Many will assume that Ellington’s touchdowns were primarily longer scores, but I was surprised how much the opposite is true. Of the 19 touchdowns Ellington scored in his junior and senior years, 10 (53%) were from within 5 yard-line, and 13 (68%) were from within the 10 yard-line. His college coaches obviously felt comfortable making him the go-to, goal-line option despite his small size. The prevailing narrative has let Ellington’s limited use during one NFL season trump a much more significant history of goal-line effectiveness in college. Bernard seems to have some goal-line potential priced into in his current ADP in the second round, while Ellington is right in line with Bush and Spiller, in the tier that should lose all goal-line carries.
Football outsiders uses a number of advanced statistics to evaluate the performance of running backs. Two I’m going to focus on are defensive-adjusted value over average (DVOA, or value per over an average running back in the same game situation) and “success rate” (SR, a statistic that could be used to measure the consistency of a running back by treating all runs as either hits or misses). As Football Outsiders says, “a player with higher DVOA and a low success rate mixes long runs with downs getting stuffed at the line of scrimmage. A player with lower DVOA and a high success rate generally gets the yards needed, but doesn’t often get more.”
Let’s look at the runners above and their backfield counterparts performed on these metrics last year:
On first blush, Ellington’s is what the narrative says: a value-adding rusher who uses big plays, but isn’t particularly good at ensuring successful runs. However, you have to assume that the quality of your blocking has an effect on success rate, and Arizona’s offensive line ranked near the bottom of the NFL as graded by Pro Football Focus in terms of rushing grade (-27.5). On a side note, Arizona’s offensive line should be much improved this year with the signing of offensive tackle Jared Veldheer and the 2012 number seven overall pick returning from injury. Buffalo’s line was also bad (-19.7), partially explaining why Spiller’s success rate was so low. Bush and Bernard enjoyed running behind good offensive lines (+22.9 and +32.6, respectively). I included the respective big backs in the table to see how effective each runner was in comparison to the other backs on the same team, running behind the same line. Not only was the differential between Ellington and Mendenhall enormous in terms of DVOA, he also was the only runner to have a higher success rate than his big back counterpart. Spiller and Bush look particularly bad with success rates and DVOAs than Jackson and Bell.
*Height and weight information from nfldraftscout.com
The common perception is that Ellington is not physically capable of sustaining a workload of 250+ touches and has sustained lots of injuries primarily because he doesn’t have the build to make an effective every-down runner. Chase Stewart in the Pro Football Reference Blog found that BMI correlated to RB success. It makes sense that players with a lower center of gravity and bulk can more easily withstand the punishment of the position. You can see that Ellington and Bernard’s BMIs again are on a different plane than Spiller and Bush’s. It’s probably not merely coincidence that they had larger shares of nQBTDs in the college and performed relatively well in comparison to their respective big-back mates in terms of success rate. I also included top-3 ranked PPR backs to show that Ellington, in fact, has a higher BMI than all three. All of top-3 are are involved in goal-line situations and have been productive on heavy workloads, despite past questions about Charles and Forte’s goal-line effectiveness. It is true that Ellington has been injured a few times in the past, but it’s difficult to determine to what extent injuries are random or a function of a player’s build or running style. Plus, is it really logical to assume that his toe, ankle, and knee injuries were cause definitively by a lack of weight? Do these injuries not afflict heavier running backs? As long as Ellington is healthy heading into the season, I see no reason to avoid him since the accumulation of these injuries didn’t adversely affect his play late last season.
Organizational Faith and Competing Talent (Or, lack thereof)
We know that running back success is largely a function of opportunity, and opportunity is derived from a combination of what I will call organization faith (i.e. how much the organization, in particular coaches, is willing to give a particular player the first crack at dominating a backfield and will overlook poor performances) and competing talent – better yet the lack thereof. We saw last year how volume superstars like Zac Stacy and Le’Veon Bell, while obviously talented and thoroughly loved by Rotoviz, were hoisted into workhorse roles due to a lack of organization faith in the current options, and the fact that the other backs on those rosters lacked talent. Last year also taught us that an assessment on talent alone, without fully considering organization faith, can lead to overvaluing a talented back like Lamar Miller, who was mired in a committee all season. The fantasy committee almost completely ignored numerous statements by the Miami coaches and front office that Miller wasn’t necessarily going to be the dominant back. At the same time, the fantasy community fully brought into the declarations of faith for CJ Spiller last summer, and discounted the potential impact of Fred Jackson, a talented runner and former top fantasy producer, whose absence due to injury, not poor play, gave Spiller the opportunity to shine the prior year. You can see above that Jackson ranked highly in DVOA and success rate last year, and should still be considered a major competing factor this year.
I would argue that Andre Ellington enjoys not only declarations of faith from his Bruce Arians’ – whose many calls to get Ellington 25-30 touches a game have largely been mocked and equated with Doug Marrone’s head-fake last summer that Spiller would get the ball “until he throws up.” – and encouraging actions like the team not spending any draft capital at the position, but also has competing talent that redefines the word lacking. The organization did use a 5th round pick on Stepfan Taylor last year, and seems somewhat eager to use him, but Rotoviz readers know well that his measurables of 4.76 40-yard dash and agility score north of 11.5 at only 214 lbs fits him completely in the JAG camp. In limited attempts last year, he was predictably underwhelming, with a DVOA of negative 14.7%, and there’s no reason to assume he’s going to be a guy that demands touches through his stellar play. Johnathan Dwyer has decent athleticism, and was an okay producer when given the opportunity. But, his value on the goal-line is questionable with 2 TDs in 230 career carries. I believe the threat to Ellington from both Taylor and Dwyer is overblown, and fully priced into his 3rd round ADP, leaving massive upside if he can truly grab the reins and dominate the backfield. It’s assumed that these guys will come in and fill the Mendenhall role, but perhaps we should give more credit to Mendenhall for producing as much as he did behind a horrible Arizona offensive line. Mendenhall was a former 1st-round pick and enjoyed multiple years of top-25 fantasy RB finishes earlier in his career. Neither Taylor nor Dwyer has the pedigree or historical performance to match Mendenhall.
Spiller, Bush and Gio have much more considerable talent to compete with in their backfields. Buffalo traded for by Bryce Brown, all but eliminating the possibility that Spiller can avoid a committee if Jackson goes down. The Lions gave big guaranteed money to Joique Bell, raising the possibility that the organization may have more faith in him than Bush, and that Bell could take the lead in the backfield this year. Bernard continues to be priced well above the other three, while his backfield became more competitive with the additional of 2nd round pick Jeremy Hill, and obvious upgrade over BJGE and someone who may be the number one rookie in 2014.
Gio Bernard With Upside At Spiller Prices
Ellington’s baseline ranking for this year assumes that he isn’t effective on the goal-line and that he will likely cede material early-down work to either Taylor or Dwyer. Our analysis shows that the perception of goal-line ineffectiveness is false, based on limited usage in one NFL season. We have also shown that the organizational faith and lack of competing talent in Arizona gives Ellington a chance to carve out a dominate role, much like other top running backs (Charles and Forte) who are of a similar build and have overcome durability concerns. Considering Ellington’s effectiveness last year on limited touches, even with a regression of fantasy points per touch (an impressive .0771 last year), Ellington’s current PPR ADP in the mid/late 3rd round should give you nothing but upside, barring injury. He’s equally priced to Spiller and Bush, whom we have shown did not profile to goal-line roles in college and weren’t even the most effective backs on their respective teams last year. Plus, he’s cheaper than Gio Bernard, who has a similar profile with much more competing talent. Ellington is an easy value play in the 3rd round, and one of the only options at that price with a 250 carry, 50 reception, 10 TD season within his possible range of outcomes.