≡ Menu

After the Jaguars drafted Leonard Fournette with the 4th pick in the 2017 Draft, NFLResearch tweeted the following:

That was, at least for me, a surprise. And it is true: there have been nine running backs drafted in the top 5 since 2000, and those teams have improved by 43 wins. There is some natural regression to the mean built in to any analysis like this, along with two big outliers: the 2016 Cowboys and 2006 Saints used a top five pick on a running back, but also added Dak Prescott and Drew Brees, producing two of the greatest improvements in passing efficiency in NFL history. Those two teams produced 16 of those 43 wins; without those two teams, the average increase drops to a still-impressive 3.9 wins.

The year 2000 does make top-5 running backs look good. 1999 brings in Edgerrin James (+10 win improvement!) but also Ricky Williams (-3); Curtis Enis (0) was drafted the year before. If you go back to 1967, the first year of the common draft, the average improvement (pro-rated non-16 game seasons to 16 games) was +3.2. The average team jumps from 4.1 wins per 16 games in Year N-1 to 7.3 wins; that’s substantial, and more than you would expect simply due to regression to the mean.1

In the table below, I’ve listed every running selected in the top five since ’67. I’ve included his team’s Year N-1 winning percentage and Year N winning percentage, along with the (pro-rated) improvement in total wins.  The final column shows the (pro-rated) rushing yards produced by that player as a rookie. As you can see, there’s a lot of big jumps:

PlayerYearTeamYr N-1 Win%Yr N Win %Win ImpRush Yds
Ezekiel Elliott2016DAL0.2500.81391631
Trent Richardson2012CLE0.2500.3131950
Darren McFadden2008OAK0.2500.3131499
Reggie Bush2006NOR0.1880.6257565
Cedric Benson2005CHI0.3130.6886272
Cadillac Williams2005TAM0.3130.68861178
Ronnie Brown2005MIA0.2500.5635907
LaDainian Tomlinson2001SDG0.0630.31341236
Jamal Lewis2000BAL0.5000.75041364
Edgerrin James1999IND0.1880.813101553
Ricky Williams1999NOR0.3750.188-3884
Curtis Enis1998CHI0.2500.2500497
Ki-Jana Carter1995CIN0.1880.43840
Marshall Faulk1994IND0.2500.50041282
Garrison Hearst1993PHO0.2500.4383264
Blair Thomas1990NYJ0.2500.3752620
Barry Sanders1989DET0.2500.43831470
Alonzo Highsmith1987HOU0.3130.6005141
Brent Fullwood1987GNB0.2500.3672365
Bo Jackson1986TAM0.1250.12500
Eric Dickerson1983RAM0.2220.56351808
Curt Warner1983SEA0.4440.56321449
Freeman McNeil1981NYJ0.2500.6567623
George Rogers1981NOR0.0630.25031674
Billy Sims1980DET0.1250.56371303
Curtis Dickey1980BAL0.3130.4382800
Terry Miller1978BUF0.2140.31321060
Earl Campbell1978HOU0.5710.62511450
Ricky Bell1977TAM0.0000.1432498
Tony Dorsett1977DAL0.7860.85711151
Joe Washington1976SDG0.1430.42950
Chuck Muncie1976NOR0.1430.2862753
Walter Payton1975CHI0.2860.2860776
Bo Matthews1974SDG0.1790.3573375
O.J. Simpson1969BUF0.1070.2863797
Clint Jones1967MIN0.3210.321026

Five teams improved by at least 7 wins per 16 games, including the ’16 Cowboys and ’06 Saints. The 1999 Colts made huge strides, too, but that also had a lot more to do with Peyton Manning than the arguably downgrade the team made at running back. The 1980 Lions had Gary Danielson return from a knee injury that caused him to miss all of 1979: Detroit ranked 3rd to last in ANY/A in 1979, but jumped back up to 8th in ’80. The other team to to make a huge leap in wins was the ’81 Jets, but McNeil didn’t have a huge rookie season: instead, the team’s defense was the big driver.

So what do you think? On one hand, the big improvement is real; on the other, with such a small sample, it’s hard to know how much of that is actually due to the running back. There is one thing does feel clear: the Jaguars went 3-13 with the 27th-ranked passing game in 2016. Regardless of how good Fournette is, Jacksonville isn’t going to post a winning record unless the team’s passing game improves significantly. The real question, though, is whether Fournette — solely due to his rushing ability — can help drive that improvement.

  1. Whether the following is noteworthy I leave to you: remove the 2016 Cowboys and 2006 Saints, and those numbers become 4.1 wins and 7.1 wins. []
  • Tim Truemper

    Well, a visual perusal suggests that many improvements in any gain in wins by a particular team had little to do with a RB picked in the 1st round, given the low yardage by some of the RB picks who either did not play or were not a starter. Not sure what correlation formula would allow for assessing the effect of RB rookie yardage to win improvement. Might need some type of ranking system of yardage gained to improvement and use the Mann-Whitney U but that is an old computational method and not sure if it is in use. But the chart tells me that for every team that showed some improvement, there are other specific case (aka , team) factors that could account for its magnitude.

  • Four Touchdowns

    I think it’s similar to the idea of the “teams who kneel the ball down win 100% of games” stat — it’s an interesting coincidence but the idea that a great runningback will automatically and substantially improve wins on his own is laughable.

    If I were going to invent a theory off the top of my head, I’d guess that teams drafting a runningback in the top five have made enough roster improvements in previous drafts and FA feel that they’re a player away from competing for a title.

    Dallas is a god example — they have a great offensive line, they have great receivers, they have good a quarterback (and were lucky to pick up a good rookie that same season), and most importantly, they had such an awful record in the previous season due to QB injuries. The year prior, they won the division with a 12-4 record.

    While they probably needed more help on defense than offense, you could argue that a true play-maker at runningback was all the offense needed to really explode.

    It’s a weird idea that NFL Research is trying to put out there, IMO.

    • Sure. The real question is what happens with Bortles this year, and how much — if any — does Fournette help. What do you think?

      • Four Touchdowns

        They might improve in wins because of all the talent they got on defense. I can’t really say as I don’t follow the Jags — if they have a great offensive line and can run the ball, they might keep Bortles out of tough situations and he can have better metrics (though I doubt he’ll ever really develop into a true franchise QB from everything I’ve seen and read).

        I think if wins come, it will be from improvements to the coaching staff and roster overall, not a single runningback.

    • Quinton White

      I’ll second this. Running backs are more likely to be selected by teams with more complete rosters (or at least they perceive themselves to be more more complete). The Colts with James and the Cowboys with Elliot are nice examples. Somewhat oddly given the Jags recent performances, I think this fits them as well (at least given the talent for this draft. They would have loved a LT). Their defense, either via young players or free agency, looks good on paper. They have decent talent at WR. Thus, they talk themselves into Fournette. If the team improves this season, my guess is it is due more to defensive improvement than improvement on offense. I would be curious if you could decompose the improvement of teams who drafted a RB in the top 5 in the prior draft into run game improvement, pass game improvement and defensive improvement.

  • “the average improvement (pro-rated non-16 game seasons to 16 games) was +3.2. The average team jumps from 4.1 wins per 16 games in Year N-1 to 7.3 wins; that’s substantial, and more than you would expect simply due to regression to the mean.”

    Is it? I wonder what the average improvement from a top-5 drafting team is overall. I’d guess it’s in the range of 2.5-3 wins. Just taking the average of the positions in that tweet/article (not the right way to do it, just a quick calculation) gives you 2.95. That would make drafting a RB worth about on the order of 0.5 better than average. A bit of a bump, but not that major.

    • That seems about right. With a 70% regression to the mean calculation, you get 6.8 wins.

  • Frank Yi

    I think a few things need to be parsed out to figure this out clearly.

    1. What is the average and standard deviation for increase in wins for a team drafting in the top 5? 4.8 (RBs) might be the highest, but it also may not be statistically significantly higher than the average. Additionally, what is the volume of each position? Relatively few players at a position could skew the mean due to high variance.

    2. Passing efficiency has a much stronger correlation to win% than running efficiency. What was the average gain in passing efficiency in the teams who drafted a RB vs other positions? As stated before, it is likely that a team drafting a RB that high has already improved the other facets of the team, and RB was kind of a luxury selection (ie Edgerrin James in 1999 after drafting Peyton Manning in 1998).

    3. How much did the running game actually improve due to the RB? PFF, DVOA, or AV could give us an idea of the surrounding talent on the RB

    • Agreed; food for future thought.

  • Richie

    My hunch is that there is generally more to it than just plugging in a top-5 RB.

    BUT, don’t RB’s generally have more impact as rookies than most (all?) other positions? It’s a position that relies heavily on skill and instinct, as opposed to many other positions that require players to learn schemes and working with teammates.

  • Four Touchdowns

    Ha ha, and by sheer coincidence…

    “Running Backs Are Finally Getting Paid What They’re Worth

    […] But it may also reflect a growing recognition that, for all their talent, traditionally great running backs probably don’t actually contribute that much to their teams’ chances of winning.”

    https://fivethirtyeight.com/features/running-backs-are-finally-getting-paid-what-theyre-worth/

  • sacramento gold miners

    Interesting to note 16 of the 36 backs listed above had disappointing careers, from Trent Richardson to Clint Jones. The failure of Jones was one reason the Vikings selected Chuck Foreman in 1973.

  • Dave B

    I think a lot of this MIGHT have to due with the fact that RB’s skill is pretty much the best right when they come into the league. The learning curve is pretty small and the physical punishment of the NFL has yet to take it’s toll. Also you are getting a RB on a cheap-ish contract vs a likely overpaid veteran RB.

    Many of the other positions take more time to develop or have less singular impact.

    • Frank Yi

      I agree with the skill of a RB is at his peak as a rookie (hasn’t had the beating, position is a lot of instinct, etc), but I can’t agree with the cheap. Fournette is now the 4th highest paid RB by contract average, so he’s now being paid like an elite RB. McCaffrey is 14th, which means he’s still paid like an upper-tier RB.

      For comparison, #1 pick Garrett is the 21st highest paid Edge Defender, and #2 pick Trubisky is the 23rd highest paid QB. Maybe, as you said, it’s ok since you’re getting immediate production at the prime of a player’s career, but Dalvin Cook is the 36th highest paid RB, and I can’t imagine he’s going to be so far behind Fournette in production to warrant that kind of pay differential.