BackCAST 2019

by Nathan Forster

Advanced statistics and Internet draftniks often agree more than they disagree about the best prospects in a given draft. For example, take Football Outsiders' metric for projecting the likelihood of success for running backs, "BackCAST." Last year, BackCAST agreed with virtually everyone who followed the NFL draft that Saquon Barkley was (by far) the best running back prospect.

Every once in a while, however, advanced statistics will break strongly from conventional wisdom. For BackCAST, this is one of those drafts. This year, up is down and down is up, as a third-round prospect rockets to the top of BackCAST's projections.

BackCAST projects NFL running back success based on statistics that have correlated with success in the past. Historically, a college running back with a good size-speed combination, a high average yards per carry, and who represented a large percentage of his college team's running attack is more likely to succeed at the NFL level. BackCAST considers these factors and projects the degree to which the running back will exceed the NFL production of an "average" drafted running back during his first five years in the NFL. For example, a running back with a +50% BackCAST is projected to gain 50 percent more yards than the "average" drafted running back. BackCAST also projects whether each running back is likely to be heavily involved in the receiving game or is more of a "ground-and-pound" back.

BackCAST is based on a statistical analysis of all of the Division I halfbacks drafted between 1998 and 2017, and measures the following:

• the prospect's weight at the NFL combine;
• the prospect's 40-yard dash at the NFL combine -- if he did not run at the combine, BackCAST uses his pro day time;
• the prospect's yards per attempt, with an adjustment for running backs who had fewer career carries than an average drafted running back;
• the prospect's "AOEPS," which measures how much, on average, the prospect's team used him in the running game during his career relative to the usage of an average drafted running back during the same year of eligibility; and
• the prospect's receiving yards per game in his college career.

What follows are some of the most notable BackCAST projections for the running back prospects available in the NFL draft.

Darrell Henderson, Memphis
BackCAST Score: +116.1%
RecIndex: +0.35
Similar Historical Prospects: Melvin Gordon, Felix Jones

Darrell Henderson is by far BackCAST's favorite prospect in this class, and it likes him far more than conventional wisdom does. Scouts, Inc. rates Henderson as only the sixth best running back prospect and the 95th prospect overall. So what does BackCAST see in Henderson that others do not?

Henderson's best data point is his fantastic yards per attempt number. Henderson averaged an absolutely bonkers 8.2 yards per attempt in three years at Memphis. BackCAST adjusts that down slightly because Henderson only has 431 career attempts, but Henderson still comes out of the adjustment with a great 7.74 adjusted yards per attempt, better than all prospects in the BackCAST database except for Melvin Gordon. Henderson may have had good circumstances at Memphis, but he should also get credit for capitalizing on those circumstances in a big way.

That said, the rest of Henderson's BackCAST metrics are just average. He ran a 4.49-second 40-yard dash at 208 pounds, which is not bad, but not exactly head-turning, either. Moreover, despite Henderson's historic efficiency, Henderson's coaches ceded a significant number of carries to other players on the Tigers. Part of the reason could be simply that Memphis's scheme was to run the ball a lot, but Henderson only had 431 career attempts, so he certainly could have carried more of the load.

On the other hand, Henderson also had strong receiving numbers, suggesting that he would also have value catching passes out of the backfield.

Trayveon Williams, Texas A&M
BackCAST Score: +57.3%
RecIndex: +0.16
Similar Historical Prospects: Jonathan Stewart, Laurence Maroney

Williams is the second-best running back prospect in this class, at least according to BackCAST. However, there is a huge gap between Henderson and Williams. The reason has less to do with Henderson's strength as a prospect and more to do with what BackCAST sees as a weak class. Williams is a fine prospect -- especially for a prospect sometimes rated as a fourth-round pick -- but he is an unusually weak No. 2 prospect. For example, BackCAST rated 2018 draftees Royce Freeman, Derrius Guice, Rashaad Penny, Nick Chubb, and Ronald Jones higher than it rates Williams.

Williams is similar to Henderson in most ways, except that he shouldered a heavier load of his team's carries and he lacks Henderson's freakish yards per attempt numbers.

Justice Hill, Oklahoma State
BackCAST Score: +46.5%
RecIndex: -0.03
Similar Historical Prospects: C.J. Spiller, Jahvid Best

Justice Hill is this class's speedster, clocking in at 4.40 seconds on his 40-yard dash at the combine. However, Hill is relatively small at only 198 pounds, one of only two running backs invited to the combine who weighed in at less than 200 pounds. Of course, if you looked only at Hill's stat line, you might guess that he was a ground-and-pound back. Hill shared a large load of the Cowboys' carries while a senior but had a relatively low 5.60 average yards per attempt.

Bryce Love, Stanford
BackCAST Score: +39.3%
RecIndex: 0.00
Similar Historical Prospects: Mikel Leshoure, C.J. Prosise

Bryce Love has had a somewhat unusual college career. Love was stuck on the depth chart behind Christian McCaffrey at Stanford until his junior year, and then he burst out in a big way. As a junior, Love exploded for 2,118 yards and averaged an insane 8.05 yards per carry. Love made the (perhaps unwise) decision to return to the Cardinals for his senior year, and it was a disaster. Love averaged only 4.45 yards per carry as a senior and tore his ACL, which kept him from running the 40-yard dash at the combine. Overall, Love averaged 6.79 yards per carry over his college career, which is excellent.

You could make a fairly compelling argument that BackCAST underrates Love's talent. He had a strong average yards per attempt in his freshman and sophomore years, which does suggest that his senior year was a bit of a fluke. Love's share of the backfield at Stanford was also affected by two circumstances outside of his control -- Christian McCaffrey's presence on the roster and then Love's ACL tear. He also is reportedly very fast -- some reports say that he could run a sub-4.40 40-yard dash -- but could not show off his speed in pre-draft workouts due to his injury. These kind of reports should be taken with a grain of salt, of course -- this type of hype is often exposed at the combine, where "sub-4.4" players ultimately prove to be "sub-4.7" players. But a team could certainly justify spending a mid-round pick on Love if he checks out medically.

Benny Snell, Kentucky
BackCAST Score: +35.3%
RecIndex: -0.37
Similar Historical Prospects: Alfred Morris, Samaje Perine

Benny Snell carried the rock at Kentucky. Snell had more college career rushing attempts than any running back invited to the combine this year, save for Washington's Myles Gaskin. As a result, Snell's AOEPS (BackCAST's measure of running back usage) is the best in this class, and it provides a major boost to his projection.

However, Snell is not particularly good in any of BackCAST's other metrics. Snell carried the ball a lot, but he averaged only 5.26 yards per carry. Snell is also unusually slow -- he ran the 40-yard dash in 4.66 seconds, which is bad for even a larger back. Snell's upside is probably Alfred Morris, who is the most successful running back in BackCAST's database who ran a 40-yard dash at 4.66 seconds or slower.

David Montgomery, Iowa State
BackCAST Score: +24.0%
RecIndex: +0.06
Similar Historical Prospects: Javorius Allen, Correll Buckhalter

Montgomery is similar to Snell in a lot of ways. Montgomery excels in AOEPS but falls short otherwise. Montgomery was even less efficient than Snell -- he averaged only 4.7 yards per carry. Montgomery is also slow for a running back prospect, as he ran a 4.63-second 40-yard dash at the combine. His time is a little faster than Snell's, but Montgomery is a few pounds lighter than Snell too. Montgomery's one edge over Snell is in the receiving game. Montgomery averaged 15.7 receiving yards per game, which is fairly typical for a drafted running back.

Damien Harris, Alabama
BackCAST Score: -8.4%
RecIndex: -0.22
Similar Historical Prospects: Carlos Hyde, Montee Ball

Josh Jacobs, Alabama
BackCAST Score: -39.3%
RecIndex: +0.02
Similar Historical Prospects: Chris Perry, Julius Jones

If you have been following the 2019 draft prospects closely, you might have thought that we had forgotten Josh Jacobs, who many consider to be the best running back in this class. Jacobs' BackCAST is unusually low for a running back who could go in the first two rounds.

Below is a chart of the running back prospects selected in the first two rounds of the draft who have the lowest BackCAST projections:

That is not a great list. So why is BackCAST so down on Jacobs?

When a running back fails to even lead his own team in rushing attempts, it is typically a major red flag. Here, Jacobs had fewer carries than his teammate, Damien Harris, whom BackCAST does not particularly like either. In terms of yards, Jacobs was not even the second-leading rusher on his own team. That distinction goes to sophomore Najee Harris, who had three fewer carries but 140 more rushing yards. Of the top 50 most productive running backs in BackCAST's database, only one running back -- Alvin Kamara -- had a worse AOEPS than Jacobs.

That leads to another issue with Jacobs' prospects -- his relative inefficiency. Last year, Jacobs averaged only 5.33 yards per carry, which was a full half-yard per carry less than Damien Harris and nearly a yard and a half per carry less than Najee Harris. Indeed, Damien Harris is a similar prospect to Jacobs, but he edges Jacobs out in BackCAST because he had a larger share of Alabama's carries and a higher average yards per carry.

To top it off, Jacobs appears to be rather slow. Jacobs did not run at the combine, but reports from his pro day suggest that he, at the fastest, ran the 40-yard dash in 4.60 seconds. That is not a time that will doom a running back to failure, but slow college running backs who succeed at the NFL level typically have more college production than Jacobs. Unfortunately for Jacobs, his BackCAST numbers are simply poor overall.

What could BackCAST be missing about Jacobs? There is something: his ability to break tackles. That's a metric that can't be included in BackCAST because it's only been counted in recent years and is not publicly available data. But according to the Sports Info Solutions 2019 SIS Football Rookie Handbook, Jacobs and James Williams of Washington State are tied for the lead among this year's running back prospects with 38 broken tackles per 100 touches over their college careers. (Montgomery and Devin Singletary of Florida Atlantic are right behind with 37.)

Portions of this article originally appeared on ESPN+.