Elite Running Backs and Value in the Passing Game


Only one running back was selected in the first round in this year’s draft, the lowest total since the merger. As Brian Burke at Advanced NFL Stats points out, the position has been declining in importance over time as the passing games have become paramount. While I agree that passing has become more important, and teams should probably throw the ball more on first down and in some other situations, I am not one who thinks that Fungible running backs are any more fungible than most other positions. I just think that we have more detailed ways to measure running backs (when we are really, like all individual football statistics, measuring lots of teammate contributions) and so we see that “BenJarvus Green-Ellis” puts up good efficiency numbers and we say things like backs are fungible. I think the same would be true if we measured guard play or looked at tackles for outside linebackers. But that doesn’t mean that Clay Matthews or a random guard from New Orleans or Adrian Peterson is fungible. Elite play matters at running back just like it does other positions.

So what I did is look at all quarterbacks since 1978 who threw 300 or more passes in a season at least 7 times between age 25 and 34, and then look at the leading rushers during each of those seasons. I’m not looking at the individual running back numbers for those seasons, but rather total career longevity and career numbers. If you started at running back in the NFL for 9 years, you were probably a pretty good player during most of those years, regardless of what the specific numbers say. I then compared the era-adjusted ANYA+ numbers available at pro-football-reference for those quarterback seasons, and sorted by the running back career.

Here is a quick summary of the quarterback’s ANYA+, based on the career rushing yards of his leading rusher in a given season:

starting RB career ydsQB ANYA+ score

less than 1500 109.1
1500 to 2999 107.9
3000 to 4499 105.4
4500 to 5999 103.4
6000 to 7499 111.9
7500 to 8999 114.8
9000 or more 111.9

Those “less than 1500” numbers are higher, with a small sample size, based on Ickey Woods with Boomer Esiason in 1988 (Woods blew out a knee and was not your typical one year starter), plus two seasons of Dan Fouts with Clarence Williams and Lionel James, and three seasons of Joe Montana with Ricky Patton, Moore, and Dexter Carter. In other words, not a representative sample of the quarterback group, as it featured some of the best offenses in seasons they happened to have short term running backs.

In fact, looking at Fouts’ numbers, his numbers were very good when he had some combination of Clarence Williams, Lionel James and Earnest Jackson at running back (ANYA+ of 119). When he had Lydell Mitchell and Chuck Muncie, it was other-worldly (ANYA+ of 130). As a quick explanation of those numbers, a ANYA+ score of 100 would represent league average. Every 15 points is one standard deviation better than average, so the 130 score that the Air Coryell offense put up with Lydell Mitchell and Chuck Muncie is two standard deviations better than average (Aaron Rodgers was a 120 last year, for comparison).

The same story is true for Montana. The 49ers were pretty good even with replacement level backs (ANYA+ of 118 for 3 years) but were even better with a combination of Wendell Tyler and Roger Craig (ANYA+ of 125). In fact, going down the list we see this over and over. Not every top quarterback from age 25 to 34 had variation; Aikman for example had the advantage of playing all ten years with Emmitt Smith; Kelly spent most of his career with Thurman Thomas.

  • Drew Bledsoe peaked at age 25 (ANYA+ of 111), with by far the best running back he played with, Curtis Martin;
  • Mark Brunell’s only below average season through age 34, besides his first year as starter, came the year Fred Taylor was injured all year and Stacey Mack was the leading rusher;
  • Kerry Collins was signficantly better with Tiki Barber (ANYA+ 105) than with Fred Lane, Joe Montgomery, Amos Zereoue, and Lamont Jordan (ANYA+ of 94);
  • John Elway’s worst seasons were with Gaston Green, who had the lowest career rushing yards, at running back;
  • Jim Everett’s only below average season until his last one as a starter came at age 28, the only year that Robert Delpino started at running back in his career;
  • Jim Harbaugh’s career revival and by far best season in 1995 came with a young Marshall Faulk at running back;
  • Ron Jaworski’s best seasons came with Wilbert Montgomery at running back (ANYA+ of 109) and his numbers were below average after Montgomery declined;
  • Peyton Manning’s worst season came in 2001, the year that Edgerrin James blew out his knee and Dominic Rhodes started;
  • Donovan McNabb emerged with his best passing seasons with Brian Westbrook at running back;
  • Steve McNair’s numbers fell off with Chris Brown at running back after a lengthy run with Eddie George;
  • Ken O’Brien fell off a cliff and went from an above average quarterback (ANYA+ of 108) with Freeman McNeil to a below average quarterback with Johnny Hector and Blair Thomas at running back (ANYA+ of 98);
  • Vinny Testaverde was an interception machine with Lars Tate at running back; when he played with guys like Curtis Martin and Earnest Byner late in his career, he experienced a career revival at an advanced age.

You can also look at others who weren’t in this data set, where a star running back got injured or traded and was replaced, like the Edgerrin James/Dominic Rhodes situation. In most cases, the backup put up decent numbers, but it was the passing game that struggled. Brett Favre’s worst season until last year came when running back injuries forced Samkon Gado into the lineup. Steve Young was much better with Ricky Watters from 1992-1994 than with Derek Loville in 1995. Dave Krieg dropped from 8.8 yards per attempt with a stud Curt Warner in 1983 to a still very good 7.6 without him when he blew out his knee in 1984.

It’s a complex issue, because sometimes the quarterback influences the running game and opens lanes because of the coverages, and sometimes the running back influences the type of defense the quarterback sees. When we look at it on a global level, we often see that rushing yards per carry and passing yards per attempt have little correlation, because of the effect of big plays and that we have several different causes running different directions. But I think that star quarterbacks do benefit from having star running backs versus replacement level backs, and they would probably tell you the same thing.

[photo via Getty]