Possibly the closest Super Bowl matchup in history also poses the question: how much does it mean when certain aspects of an NFL team improve dramatically in the second half of the season?
02 Aug 2004
by Michael David Smith
When we say that Jamal Lewis gained 2,066 yards rushing in 2003, that is not precisely what we mean. What we really mean is that the Ravens gained 2,066 yards when Lewis ran the ball. Lewis doesn't deserve all the credit, but it's hard to determine how much of it he does deserve. If we could find the exact number of yards to award to the offensive linemen, we'd have found the holy grail of football research.
I certainly haven't found that, but I would like to propose a simple measurement of how much credit a running back should get for the yards he gains. Let's just examine a running back's yards per carry average and then take the same team's No. 2 running back and look at his yards per carry average. In the case of Lewis, he averaged 5.3 yards a carry, while the Ravens' No. 2 running back, Chester Taylor, averaged 4.4 yards a carry behind the same offensive line. So we could say that Lewis is 0.9 yards per carry better than his replacement.
Let me acknowledge here that backup running backs often face different situations than starters. They often run the ball when their team has a big lead (and the defense is expecting a run) or when their team has a big deficit (and the defense doesn't care if a run gains a few yards). Other times a backup running back faces specific situations, like short-yardage or long-yardage downs. But I'm going to examine entire careers, and I hope it's a large enough sample size that these things will even out.
I've taken a look at some of the best running backs in NFL history (based on my own subjective judgment) and analyzed how they performed compared to a teammate playing behind the same offensive line. This should give us some idea of how much credit the runners deserve and how much credit the linemen deserve. I also tossed in three recent running backs who are on a lot of people's minds at the moment.
The tables below show each of the star running backs and how their backups did in each season. An asterisk indicates that the backup had a higher yards per carry average than the star.
The NFL Record & Fact Book says only this of Bobby Mitchell's career highlights: "91 touchdowns, including eight on kickoff and punt returns. 14,078 combined yards." He led the NFL in receiving yards as a Washington Redskins wideout in 1962 and 1963, but before that he was a running back in Cleveland, and a great one. How great? He was Jim Brown's backup for four seasons and had a better average per carry than Brown in three of those seasons. Mitchell and Brown are both in the Hall of Fame, but none of their linemen are. Why? If these linemen could not only open the holes for Brown but also for a series of backups who collectively averaged 4.8 yards a carry, they must have been doing something right.
Payton wasn't an every-down back as a rookie or in his final season; he actually had more carries but fewer yards than Anderson in 1987. But in his other 11 years his backups had a better average than he did only once. Considering what a workhorse Payton was, it's amazing that Roland Harper once came only eight yards short of 1,000 as Payton's backup. I recently heard Mike Greenberg on ESPN say, "Eddie George is the all-time workhorse in NFL history with eight consecutive seasons of 300 carries or more." I'm not sure why consecutive seasons of 300 carries or more is the standard -- is a guy who gets 300 carries every year more impressive than a guy who gets 350 one year, 299 the next, then 322 and 294? But anyway, if consecutive 300-carry seasons were the standard, I think we'd still have to give the crown to Walter Payton, who had 10 300-carry seasons in an 11-year stretch, missing the mark only in the strike-shortened 1982 season. George led the league in attempts once; Payton led the league four times.
In 1977 and 1979 Simpson (who was slowed by injuries) wasn't his team's primary ballcarrier. I find it amazing that in a 14-game season in 1973, the Bills had O.J. Simpson rush for more than 2,000 yards, Jim Braxton rush for nearly 500 and Larry Watkins, the third-stringer, rush for 400. That sounds like a team with an impressive offensive line, but make no mistake: Simpson was special. He averaged a little more than six yards a carry that year, about a yard and a half better than his backups.
If you're like me your first question is, "Who the hell is Barry Redden?" His career numbers are nine seasons, 396 carries, and 1,735 yards, but he was an effective change of pace to Dickerson with the Rams. In 1993 Dickerson only played in four games for the Falcons and gained less than 100 yards total, so we'll just disregard that year entirely; in all of his other seasons he led his team in both attempts and yards. For 1987, a strange year that includes both the Dickerson mega-trade and the three games with scabs filling in for the striking players, we're looking only at his numbers with the Colts. Dickerson clearly played past his prime, as you can see when you notice that in 1991 Ken Clark averaged 3.2 yards a carry and still gained at a better rate than Dickerson. Aaron's Miami essay from Brassey's Pro Football Forecast, which ran on the site last week, shows how the greatness of Dickerson was his ability to keep running effectively after many, many carries.
Sanders' backups only out-averaged him twice, and those were in his two injury-plagued seasons -- 1993, the only year he ever missed a game because of injury, and 1998, his last season, when he was great early in the year but struggled through rib injuries in his last six games. Not too many people describe Sanders as a workhorse, but if a workhorse is a back who gets the bulk of his team's carries, there aren't too many who deserve that title more than Sanders. Only once (1993 again) did the Lions' No. 2 running back average more than two carries a game. In 1990, Wilder's 11 carries represent the only rushes by any running back on the Lions other than Sanders.
Smith's wasn't the Cardinals' top back last year, his first season not being his team's primary ballcarrier. I find Smith's numbers to be the most interesting of all. One way of looking at it is that on a typical carry, he was only two-tenths of a yard better than his backup would have been. It's interesting that from 1992 to 1995 he put up huge numbers behind a well-regarded offensive line, and the other backs running behind the same line didn't fare so well. But clearly, the Cowboys kept Smith along well past his prime. In his last eight seasons he hasn't averaged better than 4.2 yards a carry, and in six of those seasons his team's other runner was more effective. And yet Dennis Green says Smith is the Cardinals' starter this year. I think Peter King said it best when he wrote that what Dennis Green was really saying was that he doesn't have a starting running back.
Finally, just for fun, I'm going to include a few of the running backs who have been on the minds of all football fans recently.
And now we've come across our first back who has a lower average than his backups. I said before that Eric Dickerson's greatness came from his ability to keep performing effectively after a lot of pounding. I'd say the same thing about Eddie George except that "greatness" really doesn't describe George's career. George got to 10,000 yards in the same way that Fred McGriff might get to 500 home runs: by hanging around after he ceased to be effective.
Once again we have a player whose backups gained slightly more per carry than he did, although it's close enough that if you remove Deuce McAllister's longest run in 2001 Williams moves ahead of his backups. The question is how much the Dolphins will miss Williams, and when you consider both Minor's rushing numbers and the fact that he has never fumbled, I don't think they'll miss Williams much at all.
Davis is different from the other backs we've discussed because he wasn't a starter in his first three seasons; Terry Allen was Washington's starter until Davis took over in 1999. This subject has been discussed a great deal on this site already, but I really don't get why so many people seem to think the Panthers should take the ball out of Stephen Davis's hands and get DeShaun Foster the ball more often. The Panthers have a good coaching staff, and they did exactly the right thing last year by giving Davis almost three times as many carries as Foster. In fact, last year was only the second time that Davis outperformed his backup.
Conclusions? It's no surprise that Barry Sanders out-gained his backups by more than any of our other backs. But it's quite surprising that the Lions' backups had a higher average than the Cowboys' backups, when all through the '90s it was taken on faith that Emmitt Smith had a huge advantage by running behind a superior offensive line.
For modern players, we don't really need this information. It doesn't tell us as much as points above replacement, the statistic Football Outsiders uses to rank running backs. But points above replacement relies on play-by-play information that isn't available for most of the NFL's history. So this system will give us some ideas about past running backs. We ought to have some way to look at the great running backs in the game while acknowledging that they didn't play behind equal offensive lines. As Bill James wrote in his first Historical Baseball Abstract, I'm not arguing for my system but for a system. I look forward to hearing from those who have other ideas.