Draft grades can be funny things.
Looking at each team’s roster of selections immediately after the draft ends is a common – yet perhaps overvalued – way of assessing how each organization fared in the draft. In almost every single case, the passage of time is going to illuminate how each team executed their draft plan. Revisiting the selections a year or two down the road to get a better picture of how each team performed in the draft is probably a better exercise.
But that does not stop us from trying in the moment.
A team that was given high marks in most quarters for how they fared this draft season is the Baltimore Ravens. Eric DeCosta worked his draft board the way he wanted to, and saw linebacker Patrick Queen fall to him at the end of the first round. Then on Day Two of the draft they added five more players, including defensive tackle Justin Madubuike, wide receiver Devin Duvernay, linebacker Malik Harrison, and offensive tackle Tyre Phillips, all in the third round. They also added running back J.K. Dobbins in the second round, at the 55th selection overall. For many, that Day Two haul – coupled with the selection of Queen on the first night – made this draft a smashing success.
But there were also doubters, and they had their reasons. The Ravens are known as one of the league’s more analytically-minded teams. Daniel Stern, a behavioral economics major in college, sits right next to offensive coordinator Greg Roman in the booth on game days. Stern’s job? To provide a reasoned decision-making process on fourth down decisions. As a result of his influence, the Ravens are one of the league’s most aggressive teams on fourth down, and the results speak for themselves. The Ravens have also built out a strong analytics department, and general manager Eric DeCosta relies on the department to make the team better. “Analytics is a way that I see of organizing information,” DeCosta said. “We have all these different pieces of information – bullet points and different things. How do we organize that information effectively? And, how do we use that information to help us make decisions? So, is it a growing field? Yeah, I think it is. Is it something that we’ll just rely on strictly, ever? No, I don’t think that’s the case. Is it something that will help us make decisions? I think it can be. We would be foolish as an organization to not look at that and consider that as a way of helping us be better.”
But the selection of Dobbins flies in the face of much of the evidence put forth from the analytics world about the running back position, which has led to some questioning the Ravens’ draft – and their strategy – overall. So let us wade into the debate a bit, looking at it from both sides of the ledger.
The Case for Positional Value
The main crux of the argument against the Dobbins selection is rooted in the idea of positional value. Proponents of this position argue that in today’s NFL, where passing is king, a different position at that point in the draft (second round, 55th pick overall) would provide more of a benefit to a team than selecting a running back. Selecting a running back at this point is a poor expenditure of the draft capital, when compared with what the return on a wide receiver or even an offensive lineman might bring. Proponents of this position have a ton of evidence to back them up.
Now, the ideas behind this general position have often been boiled down to a simple argument: “Running backs don’t matter.” That has become almost a bumper sticker, political slogan of an idea that obviously has more behind it than just those four words. But as President Josiah Bartlett told us once, life is more complicated than ten – or in this case four – word answers. The main crux of this idea is that in the modern NFL, the passing game is more critical to an offense’s success than the running game, and there is data to back this up. For example, in this meaty piece from Josh Hermsmeyer at FiveThirtyEight.com, the author made the case that for a passing league, the NFL is still not passing enough. Hermsmeyer studied first down situations, which he termed the “most traditional run situation.” As he phrased it, “[i]t’s here where NFL coaches are consistently missing an opportunity to pass, particularly against defense that have stacked the box or are playing at least seven defenders close to the line of scrimmage.” Hermsmeyer called these situations “FANS,” or “First (down) Against Neutral or Stacked (boxes.)”
What did he find?
With seven to nine men near the line of scrimmage and the subsequent dearth of extra defenders in the secondary, we’d expect passing to be effective in these situations. That’s just what we found. Last season, 30 of 32 teams were more successful passing than running on FANS as measured by success rate.2 And passing wasn’t just a little more successful than running. The difference in passing success was large: 27 teams had a success rate that was at least 10 percentage points higher when passing on FANS than running; 14 teams were more than 20 points better. The league average difference of 19.3 leaned wildly toward passing.
Hermsmeyer also looked at Expected Points Added per Play (EPA/P) on these situations. Specifically, he examined NFL teams’ EPA/P and success rate when running vs. dropping back for a pass on first and 10 facing seven to nine men in the box, 2017. What did he find? Only three teams had a positive EPA/P when running in this situation. The New Orleans Saints, the Green Bay Packers, and the New England Patriots. How successful were they? In terms of EPA/P, the Saints added a whopping +0.03, the Packers a stout +0.02, and the Patriots a league-high number of +0.04.
But throwing the football? Only six teams saw a negative EPA/P when throwing the football in this situation. Every other team was positive, with some teams adding as much as +0.37 (Tampa Bay and Atlanta) or even the Los Angeles Chargers, who added an EPA/P of +0.60 when throwing the football in this situation.
The whole point of the game is to score more points than your opponent, and through the lens of EPA/P we know that the passing game adds more value than the running game. Hermsmeyer’s study is not alone in this regard.
Now the genesis of the “running backs don’t matter” argument can be traced, in all likelihood, to the unlikely MVP candidacy of Todd Gurley back in the 2018-2019 season. When the Los Angeles Rams were taking the football world by storm, and everyone was enthralled by the wizardry of Sean McVay, the running back became a common name associated with a potential MVP selection.
But people much smarter than me started to point out that his success rushing the football was due more to factors outside of what he brought to the table, and not his particular skillset as a running back. For example, it was pointed out that the Rams were a heavy 11 personnel team, and back in that season the Rams used that package 70% of the time, or on 586 of their offensive snaps, according to SharpFootballStats.com. More critical to this analysis is the fact that the Rams ran the football mor than they threw the football out of their 11 personnel package. In fact, they ran it 58% of the time when lined up with three wide receivers, above the league average of 46%. That meant Gurley was going up against lighter boxes for the most part, making his job easier.
In addition, it was also pointed out that by putting the defense in a light box, the offensive linemen were put in an advantageous situation. That gave the guys up front some good blocking angles to execute and some lighter personnel packages to beat up. This led to the overall point: Gurley’s success is due more to the impact of forces around him, such as offensive scheme, defensive personnel, and offensive line execution. Other running backs would be just as, or at least almost as, effective in this situation.
This fact was crystallized when Gurley suffered a knee injury down the stretch, limiting his effectiveness. In his place stepped journeyman running back C.J. Anderson. Now, nobody would call Anderson a generational talent – as Gurley was often described – but thanks to EPA/P we know that if that tag was going to be placed on someone in the Rams’ backfield, it was not Gurley. As Steven Ruiz wrote in a piece for For The Win:
The numbers back this up, too. We’re obviously dealing with a smaller sample size here, but Anderson was a far more efficient runner than Gurley was in 2018. On Anderson runs, the Rams averaged 0.31 Expected Points Added, according to Sports Info Solutions. On Gurley’s runs, the average dropped to 0.12. That was barely higher than the average for Malcolm Brown runs (0.10). In fact, Brown finished ahead of Gurley in percentage of run attempts that ended with a positive EPA — 60% to 53%. Anderson led the group with 76% of his handoffs resulting in a positive EPA. If traditional stats are more your speed, Gurley and Brown both finished the season averaging 4.9 yards per carry while Anderson averaged 7.0.
Ruiz encapsulated the argument as follows: “In other words, the Rams’ run game was effective no matter who had the ball in their hands. Behind that line and with that coach calling plays, a highly-paid running back is more of a luxury than a necessity. In a league with a hard salary cap, teams can’t afford luxuries.”
So while the bumper sticker, political slogan argument is drilled down to “running backs don’t matter,” the true position reads more as follows: Yes, running backs matter, but there are bigger factors at work that can lead to the success of a running back on a play-to-play basis, such as offensive line, defensive personnel, down-and-distance, game situation, and more. As such, the wise expenditure of either salary cap resources or draft capital would be in positions other than running back, if you want to bolster your ground game. But remember the overall picture here, passing is still king, so perhaps an even wiser investment would be in the passing game.”
But that does not fit on a bumper sticker — or in a tweet — so here we are.
Now that leads us to the Dobbins selection. Given what we know about the game right now, and given the value of the passing game juxtaposed with the value of the running game, is a second-round pick at the running back position a wise investment? After all, if passing is still the best way to achieve the ultimate goal of scoring points, wouldn’t a smarter investment be a receiver, or someone who can aid the passing game? Even if you want to still run the football, as we have seen there are other ways to aid your ground game besides using cap space or draft capital on a RB. The Ravens, after all, need to replace a retiring Marshal Yanda. If they wanted to help their ground game, they could have selected someone like Damien Lewis or Jonah Jackson, some highly-regarded guards who were available at that spot.
Furthermore, let’s look at it this way. Last year, the Tennessee Titans made a run to the AFC Championship Game, and their ground attack led by Derrick Henry was viewed as a huge reason for their success. But in terms of EPA/P at least, their passing game was usually more valuable. Ben Baldwin, who covers the league for The Athletic – Seattle, has created a box score tool at his website rbsdm.com (yes, short for running backs don’t matter) that allows us to look at EPA/P on a per game basis. Take Tennessee’s Week 17 victory that clinched a playoff berth for them. Henry had 32 carries for 211 yards and three touchdowns. Ryan Tannehill completed 13 of 20 passes for 198 yards and two touchdowns. Henry’s rushing attempts had an EPA/P of 0.31 per attempt. Tannehill’s dropbacks? They had an EPA/P of 0.78. MyCole Pruitt had two receptions for 13 yards, and that accounted for an EPA/P of 0.97 per target. A.J. Brown, who had four receptions on eight targets for 124 yards and a touchdown added an EPA/P of 0.70, more than double Henry’s output in the game.
The point? Even a great ground game might not contribute as much on an EPA/P basis as an average passing game. So when you factor that into the equation, it is one more arrow in the quiver against adding a running back. In fact, when you got to the playoffs, look at this. Tennessee upset the New England Patriots on Wild Card Weekend. Tannehill completed 8 of 15 passes for 72 yards, one touchdown and one interception. Henry carried the ball 34 times for 182 yards and a touchdown. And yet?
Now, maybe there is an argument to be made against EPA/P, but that is best left for another time. For now, there is a big body of evidence to be made that in terms of value, the passing game offers more than the running game. Furthermore, individual running back success is often predicated more on the factors outside of the individual player, so in the second round, even if you want to help the ground game, you can invest that draft capital elsewhere.
The Case for Dobbins
Now, let’s examine the case for picking Dobbins in the second round. As with many counter-arguments to data driven position, the case here begins with one word: “context.” In support of picking a running back in the second round, proponents of the selection argue that the Baltimore Ravens are a team that places a huge emphasis on the ground game, and as such, a running back selection makes good sense.
Putting the relative importance of the running game vis-a-vis the passing game to the side for a moment, let’s dive a bit deeper into the 2019 Ravens. Specifically, personnel usage. An argument in favor of addressing a different position – say, wide receiver – can also be made on personnel package grounds. Most teams rely on 11 personnel, with three wide receivers on the field. Given that, and the relative terms of what a wide receiver could bring in EPA/P versus a running back, even a third wide receiver might be more valuable than a top running back. After all, teams used 11 offensive personnel on 55% of snaps league-wide a year ago.
But there are often exceptions to every rule, and the 2019 Ravens are one such exception. They used 11 offensive personnel on just 44% of their offensive snaps a year ago, and only six teams used that package less as a percentage of their offensive plays. First up, the Arizona Cardinals, who used 11 offensive personnel just 37% of the time. The Cardinals by contract used 10 offensive personnel, with four wide receivers, 26% of the time, far and a way the league leaders in that personnel grouping. The Denver Broncos were one of those six teams, who used 11 offensive personnel 43% of the time. The Minnesota Vikings used 11 personnel a league-fewest 18% of the time, as they were a heavy 21/12 personnel team, similar to the San Francisco 49ers, who used 11 personnel 30% of the time but relied on both 12/21 personnel. The Philadelphia Eagles used 11 personnel just 38% of the time, as opposed to their 54% usage of 12 personnel. Finally there were the Titans, who used 11 personnel on 41% of their offensive snaps.
The Ravens, outside of 11 personnel, use 12 offensive personnel (two tight ends, two running backs) on 17% of their offensive plays. They used 21 offensive personnel (two running backs, two wide receivers, one tight end) on 12% of their plays. Finally, they used 22 personnel (two running backs, two tight ends and just one wide receiver) on 16% of their passing plays. That was a league-high, and it accounted for 148 of their total plays. In fact, when you combine 21 and 22 personnel, that accounts for 258 offensive plays, or 38% of their offense. 38% of the time they had two running backs on the field. Then add in 12 personnel or another 161 plays. That gets you to 45% of their offense. If you really want to get crazy, the Ravens used 13 offensive personnel – three tight ends – on 7% of their plays. Meaning that on the majority of their offensive plays, they had two or fewer wide receivers on the field.
So if they did draft a wide receiver in the second round, and added that player to their crop of WRs, that player might not see the field as much as a player at running back or tight end does. Especially when you consider that a season ago the Ravens drafted two receivers in the first three rounds of the draft: Marquise “Hollywood” Brown, their first-round pick, and Myles Boykin, their third-round selection.
Furthermore, it is clear that the Ravens did have a “type” this season when drafting wide receivers, as they ended up doubling down on the position after the Dobbins pick. In the third round they drafted Devin Duvernay, a wide receiver from Texas. Then on the third day of the draft they added SMU wide receiver James Proche. What do these two wideouts have in common? They are both slot receivers. Duvernay was described in the Pro Football Focus Draft Guide as “one of the top slot receivers in the class,” and his 104 receptions from the slot a year ago were second in college football. Proche is viewed in a similar manner. PFF described him as a player “…tailor-made to play slot in NFL. Quickness and route running are pluses.” PFF went on to say that Proche is “[d]eadly with a two-way go. Perfected hesitation move to freeze defenders on in/out routes.”
That leads us to a potential discussion of slot receivers. Because they have that two-way go opportunity, the job is a bit easier for them. Without the sideline near them (and the corresponding ability of the defensive back to potentially pin them along the boundary) slot receivers with that two-way go opportunity have it easier when it comes to getting off the line of scrimmage and into their route. As such, receivers with a proven ability to beat coverage on the outside are more highly coveted, while slot receivers fall in the draft.
So here we have the Ravens, a team that relies on three receiver looks less than the majority of the teams in the league, a team that drafted two receivers early last year, and a team that looked to maximize value in slot receivers later in the draft after picking a running back in the second round. Perhaps they did account for positional value after all. Perhaps their view of the situation was: “We rely on running backs and tight ends more in our offense than we do wide receivers. We drafted two wide receivers early last year, and if we want to use more 11 personnel in the season ahead we really need a slot type, but since we can get those guys a bit later in the draft because of how they are valued, we can go with a different position here in the second.”
But even still, as we just saw, there is a case to be made that if they wanted to help the ground game, offensive line is perhaps a more valuable asset. With the Yanda retirement, perhaps a more effective use of draft capital. After all, you already have Mark Ingram, right?
Looking at that argument, an offshoot of the – again, incredibly oversimplified – “running backs don’t matter” argument is that you should not give running backs big money deals, because the return on the investment is never worth the expenditure. In draft capital terms, you can wait on these players and get running backs to contribute in the later rounds. This year alone players like Antonio Gibson, Zack Moss, Joshua Kelly and Anthony McFarland Jr. came off the board after Dobbins. In salary cap terms, spending money on these guys is a poor financial choice.
When the Ravens signed Mark Ingram last year in free agency, they signed him to a three-year, $15 million dollar deal. For the 2020 season, he is due a base salary of $4 million with a $1.3 million signing bonus. In 2021? Ingram has a base salary of $5 million. But if the team were to move on from him prior to next year, he would account for a dead cap number of $1.3 million.
Dobbins’ second-round contract? According to OverTheCap.com that will com with a base salary of $610,000 in 2020.
If the idea here is to move on from Ingram prior to 2021 with a cost-controlled running back in Dobbins, that might be a wise expenditure of cap resources after all. Now, if they end up giving Ingram a big money deal – and then doing the same with Dobbins a few years later – that would be ripe for criticism. But right now it does not seem like the Ravens are going to go down that road.
So ultimately, the case for Dobbins might also be a positional value one, in a sense. While the majority of teams may value a third wide receiver at that spot in the draft, the Ravens rely more on running backs and tight ends in terms of position groupings, and if they are going to implement more 11 personnel in 2020, they needed more of a slot receiver. A position that, in terms of positional value, is not coveted as highly, allowing them to double-dip at that position later in the draft with two guys viewed as among the best the draft had to offer. So sure, for the bulk of NFL teams a guy who might be WR3 is more valuable than RB2 or even RB1B, but in Baltimore, the rules are different.
The Verdict
Look, I was a bad lawyer. My attempts here at carving out arguments on both sides of this debate probably brings that point home. So it stands that I would make an even worse judge, but here we are…
There are compelling points on both sides of this debate. In the end, the “verdict” might read something like this: “In a vacuum, the positional value of a running back in the second round is largely outweighed by the positional value of another position, particularly one like wide receiver, that can help you in the passing game. Years of evidence points to the fact that the passing game is critical to scoring points, which is after all the goal of any offense, and even the most successful ground games still are outperformed on an EPA/P basis by average passing games. The 2019 Tennessee Titans are a testament to that idea. Even if you want to help the ground game, perhaps a position along the offensive line is a wiser use of draft capital at that spot, because we have seen that running back success is often due more to the offensive line play in front of them. Oh, and an offensive lineman can help protect your quarterback too.
However, in this limited case, the Baltimore Ravens might serve as an exception to that general rule of positional value. Unlike most teams, the Ravens rely more on two- or even one-WR packages, so adding a receiver at that spot – a year after drafting two receivers early – might not be the best use of draft capital. If indeed the Ravens are going to use more 11 personnel in the year ahead, their true need was a guy to play in the slot, a position where you can find options later in the draft. Which is what they did, in picking Duvernay and Proche later on Day Two and even later on Day Three. Ultimately, what the Ravens did was select a player who can be a cost-controlled option in their running game, allowing them to move on from Ingram prior to next year, preventing them from paying a running back a base salary of $5 million.”
But that doesn’t fit on a bumper sticker. Or in a tweet.
