SITE NEWS: We are moving all of our site and company news into a single blog for We'll tag all PFR content, so you can quickly and easily find the content you want.

Also, our existing PFR blog rss feed will be redirected to the new site's feed. » Sports Reference

For more from Chase and Jason, check out their work at Football Perspective and The Big Lead.

Most Dominant RB Ever: Methodology Discussion

Posted by Chase Stuart on April 6, 2009

Here at PFR, we've looked at the best Wide Receivers, the top Quarterbacks, and even the strongest defenses in NFL history. While long overdue, let's finally take a look at analyzing which RBs have produced the most dominant statistics of all time. I apologize in advance for the length of this post - one of the reasons I've put this project off for so long was how complicated it is to accurately rank running backs. Running back fans, rejoice: PFR's devoting the full week to looking at the greatest RBs in NFL history.

I've already mentioned two caveats. One, this will just be a statistical look. Opinions about the offensive lines of Emmitt Smith, Jim Brown, Jim Taylor and Priest Holmes are excluded from this study; feel free to move those players up or down based on your views of their supporting casts. Second, we're focusing on dominance, not necessarily value. A RB who runs for 1,200 yards in ten straight seasons is probably more valuable to his team than one with 1500 yards in five seasons and nothing else, but the latter RB is more dominant. We're focused on peak production and sustained success (although from time to time we'll change gears).

Tackling the question of which RB has separated himself most from his peers is a difficult question. How does Barry Sanders' 1883 rushing yards and 7 rushing TDs compare to Terry Allen's 1353 rushing yards and 21 scores on the ground? What about Earl Campbell's 1934 yard season with fewer than 2,000 yards from scrimmage versus Roger Craig's 1050/1016 season in 1985?

There is no simple or obvious solution. Because of that, unfortunately, this will get pretty statistics-heavy, and will surely bore most of you. But before going down that road, let me start with the biggest difference I had to keep in mind when ranking RBs as opposed to QBs.

You've probably never thought about this before, but how many yards do you think the average QB gets on his median pass attempt? The answer is zero, and for most of NFL history, it was less than that. 2008 was the greatest passing season of all time (by adjusted net yards per attempt), but even this past season, the median pass attempt probably went for only one or two yards.

The average completion percentage was 61% while the sack rate was 5.9%; this means that on every 1,000 dropbacks, 59 times the QB was sacked. On the remaining pass plays, 574 times (61% of 941) of the time the QB completed a pass. So only 57.4% of all pass plays were completed, and surely a bunch of those completions went for negative yards or no gain.

In 1998, the completion percentage was 56.6% and the sack rate was 7.2%; this means only 4.8% of all completions would need to go for no gain (or worse) to make the median pass attempt be zero (or negative). In '88, the numbers were 54.3% and 6.8%; only 1.2% of completions would need to go for no gain (or worse) to make the median pass attempt be zero (or negative). In '78? A leaguewide completion percentage of 53.1% coupled with a sack rate of 7.9% meant that 51% of all pass plays did not gain yardage even ignoring all completed passes for negative or zero yards.

Passing is high risk, high reward. The large gains offset the risk, which is why teams average more yards per pass than yards per rush. For the passers, frequency of success isn't nearly as important as quality of the success. What about rushing? Just the opposite. In modern times, most RBs have a median carry length of three yards. I suspect that's been the case for the majority of RBs for a long time. LenDale White and his 3.9 YPC last season? Median rush of 3 yards. Adrian Peterson and his 4.8 YPC? Median rush of 3 yards.

We measure passing by adjusted net yards per attempt because raw totals are misleading (good passing teams stop passing, bad passing teams pass more often) and because the average is much more important than the median for passers. For RBs, it's reversed. We care about RB consistency because rushing isn't supposed to be high risk, high reward. For runners, we want them to move the chains. Points come out of the passing game, but moving the chains and killing the clock is the domain of runners.

With that said, I decided to break RB statistics down into three categories -- rushing, scoring and catching. Then each RB will be compared to the top rushers, scorers and catchers at his position, and only get credit for his above average work. I like this because a RB with 10 receptions for 50 yards will get a zero in the receiving category while a RB with 25 catches for 200 yards will also get a zero. It's not hard to find a RB that can catch 200 yards worth of passes; however, some teams don't use their RB that way. Michael Turner shouldn't be penalized for not getting a few more receiving yards last season -- that doesn't change the fact that he was a dominant rusher. Similarly, by comparing to the league average, 24 TDs is significantly more valuable than 12 TDs, and not just twice as valuable. Let me explain in detail.

Rushing: You might expect me to give you some convoluted stat like Rushing Yards over three yards per carry or rushing yards minus league average times carries, or who knows what. But believe it or not, I think simpler is better. I spent a lot of time deciding how to weigh yards per carry, and eventually I decided pure rushing yards alone is all we need.

Is 270/1100 better than 330/1100? One argument that I've certainly used before is "they got the same number of yards, but the first guy's team had an extra 60 plays with which to gain more yards!" The natural response to that is "why would a coach give a RB 330 carries if he was only getting 3.6 per carry?" From there, we have a two different answers. Either: a) the RB wasn't that good but either the coach was dumb or the backups were really bad, or b) the RB was good and his YPC is misleading.

Once again, with rushing, I think median carry is a more telling number than average carry. Yards per carry is not a very good measure of central tendency. On the other hand, we can infer that if a RB is getting a high number of carries, he's doing something right. Carries themselves are highly correlated with greatness.

Terrell Davis, Edgerrin James, Curtis Martin, LaDainian Tomlinson, Eric Dickerson, Clinton Portis, Eddie George, Walter Payton, Barry Sanders, Jim Brown, Emmitt Smith, Ricky Williams and Earl Campbell. Those are the RBs with over 19 carries per game for their careers. A RB that gets carry after carry is doing something right. Maybe he's consistently getting gains, maybe he's running hard despite a bad OL, or maybe he's able to kill the clock without fumbling. All of those things are good. A list of the top RBs by yards per carry? Jim Brown, Barry Sanders, Paul Lowe, Robert Smith, Joe Perry, Wendell Tyler, Greg Pruitt, James Brooks, Tiki Barber, Hugh McElhenny, O.J. Simpson, Fred Taylor and Charlie Garner all have career averages over 4.6 YPC. Ignoring the overlap, I'd prefer the first list.

Remember, teams can choose to pass instead of run. So if a RB is getting 350 carries, it can't be just because his RB teammates are bad. It's got to be because the team's QB is bad, too. And if a team's QB is bad, its other RBs are bad and one guy keeps getting carry after carry, then he's pretty valuable to his team. And if year after year he gets these carries, he's definitely doing something right. A RB with 1600 yards on 400 carries may be just as valuable or dominant as one with 1600 yards on 325 carries. After all, the obvious question for the latter RB is -- why didn't he get more carries? Perhaps the former RB was getting three and four yards on every carry, a very valuable trait.

So I'm going to go with pure rushing yards, with just one addition. At PFR we have fumbles and fumbles recovered information, but not fumbles lost data. The value of a lost fumble is about -50 yards; not all fumbles are lost -- a good number are recovered by a teammate or go out of bounds. Therefore, I decided to give -25 yards for all fumbles and also +25 yards for all fumbles recovered (and not +50, since someone else on his team could have recovered the ball. Note: I should have derived the exact fumble recovery rates, but I took the easy way out and just said a fumble had a 50/50 chance of being recovered by either team). So a 1600 yard rushing season with 6 fumbles and 2 fumbles recovered (presumed two fumbles lost to the other team) is equivalent to a 1500 yard season with zero fumbles.

Scoring: This one's pretty simple -- rushing and receiving touchdowns combined. There's no reason to separate them out -- a receiving TD is just as valuable as a rushing one. This is important because a RB like James Brooks in 1988 had 6 receiving TDs but under 300 receiving yards. If we combined all receiving numbers together, Brooks might not get any receiving credit because his receiving yardage was lower than average. But his TDs were very valuable, and this way he'll get full credit for them.

Catching: Another simple formula here; adjusted catch yards are simply receiving yards plus 1.5 yards for each reception. This gives a small bonus for having a bunch of receptions but not too much.

Each RB will receive a grade in rushing, scoring and catching. Their grade will show how dominant they were relative to their peers. How? By comparing their production in that category to the league average. The league average is defined as the average production of the top N RBs in the league (in that specific category), where N is the number of teams in the league. Additionally, the RB in question will have his production removed from the league average, so for 2008, all the top RBs get compared to the other top 31 RBs in the league. Finally, per game production was used for the league average, and a RB "counted" as long as he played in at least 50% of the league's games that season.

So in 2008, the top 32 RBs by adjusted rushing yards averaged 66.5 adjusted rushing yards per game (with the adjustment being just for fumbles). The top 32 RBs in scoring per game averaged 0.70 TD/game. The top 32 RBs in adjusted catch yards averaged 27.7 ACY/G. So a RB that played in 16 games would need to have more than 1,064 adjusted rushing yards, 11.2 total TDs or more than 443 adjusted receiving yards to get credit for being "above average". Because we're looking to find the most dominant RBs ever, only above average performance is going to be rewarded. There's nothing dominant about 1,000 rushing yards.

One more thing -- what do we do with a RB that misses some games? A 2008 RB that has 1,200 adjusted yards in 12 games is more valuable than one who has 1,200 adjusted yards in 16 games, and should be rewarded as such. The latter RB would get credit for being 136 yards over average - he was 8.5 yards over average for 16 games. What about the former? He was 33.5 yards over average for 12 games, and then did not play for four games. So he'll get credit for being +402 when he played; what about when he didn't? Some arbitrary penalty must be given -- we can't ignore that his team received below average production while the RB was out. I decided to give his team 80% of average production while he was out -- this means an assumption that the backup RB(s) get 53.2 adjusted rushing yards per game (in '08). The converse of that is the starter gets a penalty of -13.3 yards per game he was out; for four games, that's -53 yards.

It's not as complicated as it sounds. It gives a RB in 2008 who had 1200 rushing yards for 12 games a value of +349. That was a mouthful and it sounds really complicated, but it makes 1200 yards in 12 games (100 yards per game) equal to 1413 yards in 16 games (88.3 yards per game). That looks like a good balance to me; the first RB was more dominant but some weight must be given for staying healthy.

Finally, how do we add it all together? A RB gets his score in the rushing category, whatever it is. He then gets his score in the TDs or the receiving category but only if his score is positive. This way, a RB that isn't a goal line threat (or plays with Peyton Manning) or a great pass catcher isn't severely penalized.

Let's use an example. LaDainian Tomlinson in 2006. He played in every game and had 1815 rushing yards, 2 fumbles and 1 fumble recovery. That's an average of 111.9 adjusted rushing yards per game. The other top 31 RBs that season averaged 71.4 ARY/G; therefore, LT gets credit for being 647 yards above average that season (the difference in yards per game times 16 games). What about scoring? He had 31 TDs, an incredible 1.9 TD/G; the other top 31 RBs averaged 0.62 TD/game, meaning Tomlinson scored 21.0 more touchdowns than league average that season. Finally, he also chipped in with 56 receptions for 508 yards, a total of 592 adjusted catch yards and 37 ACY/G. The other top 31 RBs in the league averaged 27.9 ACY/G, meaning LT added 146 yards in receiving value over the average back.

Now we just add it all up; LT had 647 adjusted rushing yards over average, 146 adjusted receiving yards over average and 21 TDs over average. His total score is therefore 647 + 146 + 20*21, giving 20 yards for each TD. That total of 1,213 adjusted yards over average is one of the five greatest seasons ever. Tune in tomorrow to see just where it ranks.

I'm sure this is going to all sound pretty math heavy and complicated, but unfortunately, analyzing RBs takes a lot of work. I really like the results, though. Here are the most dominant RBs, statistically, for each year since 1932:

2008	atl	Michael Turner
2007	phi	Brian Westbrook
2006	sdg	LaDainian Tomlinson
2005	sea	Shaun Alexander
2004	sea	Shaun Alexander
2003	sdg	LaDainian Tomlinson
2002	kan	Priest Holmes
2001	ram	Marshall Faulk
2000	ram	Marshall Faulk
1999	ram	Marshall Faulk
1998	den	Terrell Davis
1997	det	Barry Sanders
1996	den	Terrell Davis
1995	dal	Emmitt Smith
1994	det	Barry Sanders
1993	dal	Emmitt Smith
1992	dal	Emmitt Smith
1991	det	Barry Sanders
1990	det	Barry Sanders
1989	buf	Thurman Thomas
1988	clt	Eric Dickerson
1987	2tm	Eric Dickerson
1986	ram	Eric Dickerson
1985	rai	Marcus Allen
1984	ram	Eric Dickerson
1983	ram	Eric Dickerson
1982	rai	Marcus Allen
1981	det	Billy Sims
1980	oti	Earl Campbell
1979	oti	Earl Campbell
1978	chi	Walter Payton
1977	chi	Walter Payton
1976	buf	O.J. Simpson
1975	buf	O.J. Simpson
1974	den	Otis Armstrong
1973	buf	O.J. Simpson
1972	was	Larry Brown
1971	den	Floyd Little
1970	nyg	Ron A. Johnson
1969	clt	Tom Matte
1968	cle	Leroy Kelly
1967	oti	Hoyle Granger
1966	nwe	Jim Nance
1965	cle	Jim Brown
1964	cle	Jim Brown
1963	cle	Jim Brown
1962	gnb	Jim Taylor
1961	gnb	Jim Taylor
1960	cle	Jim Brown
1959	cle	Jim Brown
1958	cle	Jim Brown
1957	cle	Jim Brown
1956	chi	Rick Casares
1955	clt	Alan Ameche
1954	sfo	Joe Perry
1953	sfo	Joe Perry
1952	ram	Dan Towler
1951	nyg	Eddie Price
1950	cle	Marion Motley
1949	phi	Steve Van Buren
1948	phi	Steve Van Buren
1947	phi	Steve Van Buren
1946	pit	Bill Dudley
1945	phi	Steve Van Buren
1944	nyg	Bill Paschal
1943	chi	Harry Clarke
1942	pit	Bill Dudley
1941	chi	George McAfee
1940	was	Dick Todd
1939	was	Andy Farkas
1938	pit	Whizzer White
1937	was	Cliff Battles
1936	det	Ace Gutowsky
1935	crd	Doug Russell
1934	chi	Beattie Feathers
1933	was	Jim Musick
1932	was	Cliff Battles

And the number of times each RB led the league:

Jim Brown	        7
Eric Dickerson	        5
Barry Sanders	        4
Steve Van Buren        	4
Marshall Faulk	        3
Emmitt Smith	        3
O.J. Simpson	        3
LaDainian Tomlinson	2
Shaun Alexander	        2
Terrell Davis	        2
Marcus Allen	        2
Earl Campbell	        2
Walter Payton	        2
Jim Taylor	        2
Joe Perry	        2
Bill Dudley	        2
Cliff Battles	        2
Michael Turner	        1
Brian Westbrook	        1
Priest Holmes	        1
Thurman Thomas	        1
Billy Sims	        1
Otis Armstrong	        1
Larry Brown             1
Floyd Little	        1
Ron A. Johnson	        1
Tom Matte	        1
Leroy Kelly	        1
Hoyle Granger        	1
Jim Nance	        1
Rick Casares	        1
Alan Ameche	        1
Dan Towler	        1
Eddie Price	        1
Marion Motley	        1
Bill Paschal	        1
Harry Clarke	        1
George McAfee	        1
Dick Todd	        1
Andy Farkas	        1
Whizzer White	        1
Ace Gutowsky	        1
Doug Russell	        1
Beattie Feathers	1
Jim Musick	        1

Tomorrow, we'll look at the best 50 RB seasons of all time along with some of the best performances in team history. On Wednesday, we'll check out the career list. Because all of this data focuses just on regular season work, Thursday we'll check out the top post-season runners of all time. Friday we'll put it all together again, matching post-season data with regular season data and look at new single season and career rankings.

This entry was posted on Monday, April 6th, 2009 at 7:23 am and is filed under History. You can follow any responses to this entry through the RSS 2.0 feed. Both comments and pings are currently closed.