Writers of Pro Football Prospectus 2008

09 Aug 2010

Was Emmitt Smith Just a Product of His Line?

A few weeks after Emmitt Smith was voted into the Hall of Fame, I found myself at a poker table in the middle of a discussion with nine strangers about his merits as a player. In these sort of situations, I like hearing what other people think far more than I enjoy spouting off my own opinions, so I neglected to mention what I do for a living.

I was surprised at the level of vitriol towards Smith. While there was an argument about Barry Sanders vs. Jim Brown with supporters on either side, everyone at the table agreed: Emmitt Smith was supremely overrated, a product of his offensive line. "I could've ran for 1,200 yards a year behind those guys," one player noted. I suggested that I could separate my shoulder behind them. The discussion got me thinking: Was Emmitt Smith really just a cog in a great offense?

With Smith's induction into the Hall of Fame this weekend, now seems as good a time as any to revisit the issue. Of course, we can't just take Smith out of the Cowboys offense in his prime and insert him in a mediocre one to get reliable new statistics. So I'm going to employ a reasonably simple methodology to compare Smith to the performance of other running backs: I'm going to compare his yards per carry to the yards per carry gained by every other running back on every team he was on. In this case, that means comparing Smith to every running back carry made by a Cowboys back from 1990-2002, as well as every Cardinals back in 2003 and 2004. I'll do this for the 19 other retired running backs with the most yardage in NFL history, and see how Smith measures up.

Of course, this is far from a foolproof methodology. Yards per carry isn't by any means a perfect measure of performance, and the data could be skewed by the nature of the other backs on the roster. As an example, let's say Smith finished his career with three years as a very part-time back, getting 10 mediocre carries a year as a backup for an elite running back. The closest example to this in real life as part of the study would might be Marshall Faulk getting replaced by Steven Jackson, although Faulk had more than just a handful of carries.

For players that changed teams in mid-season (Eric Dickerson and Ottis Anderson), I measured the performance for the other backs on their teams while Dickerson and Anderson were still on the teams in question.

You might not be surprised at who ranks at the top of the list. You'll probably be surprised at who ranks directly below him.

Table 1: Elite Starters Vs. Backups
Player Star YPC Other RB YPC Difference
Barry Sanders 4.99 3.97 1.02
Tiki Barber 4.71 3.70 1.01
Jim Brown 5.22 4.33 0.89
O.J. Simpson 4.67 3.94 0.73
Tony Dorsett 4.34 3.89 0.45
Walter Payton 4.36 3.93 0.43
Thurman Thomas 4.20 3.83 0.37
Marshall Faulk 4.33 3.97 0.36
Corey Dillon 4.30 3.95 0.35
Emmitt Smith 4.24 4.00 0.24
Player Star YPC Other RB YPC Difference
Eric Dickerson 4.42 4.25 0.17
Franco Harris 4.11 4.09 0.02
Warrick Dunn 4.11 4.10 0.01
Curtis Martin 4.01 4.05 -0.04
John Riggins 3.89 3.95 -0.06
Ricky Watters 4.06 4.13 -0.07
Marcus Allen 4.05 4.15 -0.10
Ottis Anderson 4.01 4.11 -0.10
Jerome Bettis 3.93 4.30 -0.37
Eddie George 3.64 4.09 -0.45

Having sat through the Rodney Hampton years, I, more than anyone else, should remember how exciting Tiki Barber was. But I was surprised to see him towards the top of this list. Note the large dropoff from Simpson to Dorsett -- Dorsett is closer to 11th than fourth.

To an extent, the guys at the bottom of the list are underrated by this measure -- there's something to be said for the bruiser abilities possessed by John Riggins, Ottis Anderson, Jerome Bettis, and Eddie George. Ricky Watters, well, he doesn't belong in that group.

As for Smith? He rates as comfortably in the middle of the pack as compared to his backups. The difference between Smith and the YPC of the backs around him was .24 yards per carry; the average for all 20 backs, coincidentally, was also .24 yards per carry.

If Smith really was a run-of-the-mill back made into a star by his line, this measure doesn't seem to pick up on it.

Posted by: Bill Barnwell on 09 Aug 2010

146 comments, Last at 15 Sep 2010, 7:02pm by pbrane

Comments

1
by alexbond :: Mon, 08/09/2010 - 2:58pm

I don't think a sane person could dispute that Emmitt Smith was really good, I think the question is, is he one of the greatest ever? He's certainly one of the most durable players of all time, but this chart for one doesn't suggest he should be in the same category as Brown/Sanders/Payton. In the end, it's tough to argue with the scoreboard - you run for that many yards, I don't care who blocks for you, you deserve to at least be in the conversation.

2
by SammyG (not verified) :: Mon, 08/09/2010 - 3:11pm

Why not look at DVOA? A big argument for Emmitt being great is that he was able to consistently produce successful runs and move the chains, which wouldn't necessarily show up in yards per carry but would show up in DVOA.

8
by tuluse :: Mon, 08/09/2010 - 4:33pm

Because you can't compare him to backs before 1993 then.

46
by t.d. :: Tue, 08/10/2010 - 3:38am

Emmitt was as good as Barry, and there was genuine debate, until he landed on his head on a touchdown dive play in week 1 of the 1996 season. I believe he had to be carted off the field on a stretcher after the play, and he was never quite the same. To that point he was neck-and-neck with Sanders for career yards, in fewer seasons

50
by jebmak :: Tue, 08/10/2010 - 8:25am

...on the Cowboys instead of the Lions.

72
by JIPanick :: Tue, 08/10/2010 - 12:46pm

...and had smashed Sanders by *230* DYAR a year over 1993-1995.

73
by sn0mm1s :: Tue, 08/10/2010 - 12:59pm

DYAR is a pretty flawed measurement though as I mentioned in response to your previous post - unless you truly believe a scrub replacement RB would rush for nearly 1700 yards with the Titans last year.

84
by Eddo :: Tue, 08/10/2010 - 3:32pm

That's not really what DYAR is saying. Johnson's "DY" (not above replacement) isn't 2006; rather, it's less, as there will be diminishing returns on every rush over 40 yards. His 90+ yard touchdowns are as much as function of field position as they are of Johnson's skill. Had the Titans had the ball at the 50 instead of their own 10, those would have been "only" 50-yard touchdowns.

85
by sn0mm1s :: Tue, 08/10/2010 - 3:51pm

Unfortunately, that is *exactly* what DYAR is saying. I have written Aaron before regarding this and my interpretation is correct. Actual Yards - DYAR is approximately the amount of yardage a replacement level player would gain over a season if he took the carries instead of the RB that actually took the carry. Which is why the stat just isn't that great. Also, the player starts experiencing diminishing returns after the 10th yard gained. In fact, here is the email I sent him and the response:

> Hi Aaron,
>
> I was hoping to get a little clarification on DYAR.
>
> If a player rushes for 1000 yards, 10 TD, 2 fumbles on 250 carries and has a DYAR of 100. Does this mean that a replacement level RB would roughly be expected to rush for 900 yards, 10 TD, 2 fumbles on those same 250 carries?
>
> Thanks.

His response:

Sorry it took me a bit of time to get back to you.

The measurements aren't really as exact as that, but yes, if those 250
carries came in the same situations against the same defenses, that
would be the expectation.

Aaron Schatz

97
by tuluse :: Tue, 08/10/2010 - 7:20pm

It's actually more complicated than that because Chris Johnson creates good situations for himself. For example, lets say he gains 8 yards on first down, then only manages 1 yard on 2nd and 2. He did well on first down, but doesn't look that great on 2nd down because of a situation he created for himself. So a replacement level back wouldn't have got 8 yards on first down, and would have never created the 2nd and 2 situation.

98
by sn0mm1s :: Tue, 08/10/2010 - 7:46pm

No, it isn't. DYAR doesn't chain runs together the replacement runner benefits from whatever good or bad situations Chris puts himself in. Every run is an individual measurement that is not dependent on the previous run. Here is how DYAR works.

1st and 10 on the Titans' 20.
Chris Johnson runs for 8 yards.
The replacement RB is expected to run for 6 yards.

The next run is from where Chris Johnson ended his run. What the theoretical replacement runner did makes no difference.

2nd and 2 from the 28
Chris Johnson runs for 72 yards.
Replacement back expected to run for 55 yards.

This is what causes the problem. If a game breaking RB breaks off a long run then the replacement is expected to break off a long run as well.

103
by tuluse :: Tue, 08/10/2010 - 8:45pm

The next run is from where Chris Johnson ended his run. What the theoretical replacement runner did makes no difference.

It does for the comparison that was made above. If you gave a replacement level running back 358 carries, he would gain 1700 yard. The answer is no because Chris Johnson created situations that made his yards look easier to get than they were.

106
by sn0mm1s :: Tue, 08/10/2010 - 9:13pm

That isn't what DYAR states - which is my point. Chris Johnson makes those yards look easy. Why? Because he is good. DYAR doesn't care about that. I have it from the guy who developed the stat. Actual Yards - DYAR = expected yards for replacement RB.

110
by tuluse :: Tue, 08/10/2010 - 9:53pm

Actual Yards - DYAR = expected yards for replacement RB.

Which is correct. If they were in the exact same situations, but the situations the Titans were in only existed because of Chris Johnson.

The only analogy I can think of is to basketball, where shot creation is a very difficult to measure and players like Allen Iverson get underrated by advanced stats because they don't account for creating shots.

113
by sn0mm1s :: Tue, 08/10/2010 - 10:08pm

That doesn't matter. It doesn't matter at all.

118
by tuluse :: Wed, 08/11/2010 - 3:29am

It absolutely does matter. You can't just say, "if plug in a replacement level runner, he'll get 1700 yards with the Titans" because he wouldn't get to 358 carries because he would fail to get first downs and the team would punt or pass more.

DYAR doesn't measure this, and really how could it?

123
by Aaron Schatz :: Wed, 08/11/2010 - 12:09pm

Hey. You do seem to be ignoring the idea of "it's not as exact as that." The more extreme the situation, the less literal the idea of DYAR is going to be. It's meant to measure the totality of a player's performance by folding everything into one number: first downs, touchdowns, etc. If you try to split it out into single carries, of course it will look sort of weird, because that was never the point.

137
by sn0mm1s :: Thu, 08/12/2010 - 8:59pm

I understand it isn't as exact as that and is supposed to summarize the performance of an RB over a season. However, it is a measurement derived from single carries and a larger sample of carries really doesn't "even out" issues with measuring a single carry by itself. It isn't like you are taking an average in regards to success points - it is more like you are ignoring or penalizing outliers (outliers being runs greater than 10 yards because that is when the diminishing returns kick in).

Obviously, the public doesn't have access to your stats or your formula so we can only go off of what you print and the final results. DYAR penalizes big plays and players that routinely make big plays. If an 80 yard run (no TD) by RB A is worth less than 3 consecutive 10 yard runs for 1st downs by RB B then you are effectively ignoring the 50 yards "extra" that RB A gained.

94
by JIPanick :: Tue, 08/10/2010 - 6:06pm

Agree that DYAR is far from perfect, but I think it is better than raw yardage. How do they compare in success rate?

3
by The Original Omar :: Mon, 08/09/2010 - 3:36pm

Derrick Lassic says no.

4
by spenczar :: Mon, 08/09/2010 - 3:44pm

Doug addressed exactly this question in a PFR post in 2006: http://www.pro-football-reference.com/blog/?p=8

Also for anyone else wondering, here is the list of other running backs on Emmitt Smith teams:

Tommie Agee FB 1990-1994
Alonzo Highsmith RB 1990-1991
Daryl Johnston FB 1990-1999
Timmy Smith RB/FB 1990
Ricky Blake RB 1991
Curvin Richards RB 1991-1993
Derrick Gainer RB 1992-1993
Derrick Lassic RB 1993
Lincoln Coleman RB 1994
Blair Thomas RB 1994
Robert Wilson FB 1994
David Lang RB 1995
Dominique Ross RB 1995-1996
Sherman Williams RB 1995-2000
Herschel Walker RB 1996-1997
Nicky Sualua FB 1997-1998
Robert Thomas FB 1998-2000
Chris Warren RB 1998-2000
Robert Chancey FB 1999
Tim Lester FB 1999
Troy Hambrick RB 2000-2002
Michael Wiley RB 2000-2002
Tony Taylor RB 2001
Terry Witherspoon RB 2001
Woody Dantzler RB 2002
Jason McKie FB 2002

And in Arizona:

Damien Anderson RB 2003
James Hodgins RB 2003
Josh Scobey RB 2003-2004
Marcel Shipp RB 2003-2004
Obafemi Ayanbadejo FB 2004
Larry Croom RB 2004
Troy Hambrick RB 2004 (who was actually also on the cowboys from 2000-2002)

In other words: Daryl Johnston, lots of backup fullbacks, and Herschel Walker's corpse. Not a very impressive bunch.

5
by Shattenjager :: Mon, 08/09/2010 - 3:51pm

Also in 2008: http://www.pro-football-reference.com/blog/?p=569

I think that one does a really good job (and parts I and III are good, too--there are links to them on that post).

7
by Dean :: Mon, 08/09/2010 - 4:29pm

So what we've learned is that all the other backs were essentailly flotsam, yet Emmitt was only 1/4 yard per carry better than them.

15
by Temo :: Mon, 08/09/2010 - 6:16pm

Can you name all the non-flotsam RBs that ran behind these guys?

20
by spenczar :: Mon, 08/09/2010 - 8:01pm

Sure. These are just the first four I did, I promise I'm not cherrypicking. I am just listing players that I think are subjectively above flotsam level. Seems like Emmitt does have an unusually bad list of backup RBs. Tony Dorsett gives him a run for his money, though.

Ricky Watters:
Nobody special from 1992-1994, but then
Charlie Garner 1995-1997
Ahman Green 1998-1999
Shaun Alexander 2000-2001

Marcus Allen:
Nobody special from 1982-1986, but then
Bo Jackson 1987-1990
Roger Craig 1991
A 32-year old Eric Dickerson 1992
Kimble Anders is sort of notable 1993-1997

Corey Dillon:
Ki-Jana... oh wait. Nobody special 1997-2001
Rudi Johnson 2002-2003
Kevin "Notarunningback" Faulk 2004-2006
Laurence Maroney 2006

Tony Dorsett:
Nobody I've heard of 1977-1985
Herschel Walker 1986-1987
Sammy Winder 1988

88
by Jacob Stevens (not verified) :: Tue, 08/10/2010 - 5:28pm

Which undermines the idea that the line made Emmitt great. Why couldn't that great line make flotsam good?

13
by Danish Denver-Fan :: Mon, 08/09/2010 - 6:08pm

Not that I read 'em carefully, but are there more than one Josh Scobey? I believe he is/was a kicker.

18
by spenczar :: Mon, 08/09/2010 - 7:30pm
6
by Dean :: Mon, 08/09/2010 - 4:23pm

I've always considered him to be among the most overrated players in the history of the game.

We have YAC for WRs, but do we have any sort of Yards After Contact for RBs? Or, more accurately in Emmitt's case, Yards BEFORE Contact. The impression created by Emmitt and his line - an impression I'm inclined to believe barring evidence to the contrary - is that he consistantly ran through big holes and was several yards downfield prior to contact. Barring, of course, two concecutive fairly well known 4th and 1 failures.

There's no question that Emmitt is a Hall of Famer. However, there are probably 20 or so backs I'd rather have playing for me. That sounds like a lot, but it still puts him as one of the greats. Just not NEARLY as high as he's made out to be.

Another anecdotal offering as to his overrated nature: one of the cornerstones of this website. Correllation vs. causation. The very example we use when explaining correllation vs. causation is "when RB X gets Y carries..." That stat came into prominence because John Madden used it to laud Emmitt and his teams record when he gets 20 carries. We know how worthless the stat is, but that never stopped the network boys from treating it as somehow valuable insight. For a decade, that number was rammed down our throat, both as evidence that running lead to winning, and somehow as evidence of Emmitts alleged greatness.

He played on the most heavily marketed team of his era. John Madden loved him. His QB is now a broadcaster and frequently trumpets his achievements. But at the end of the day, I never believed the hype matched the ability.

If anything, this post re-enforces my opinion that he's overrated. Emmitt is presented as one of the greatest of all time. Yet here are some very good and some great backs, and he doesn't stand out at all among them. Based on the hype he had throughout his career, he should have topped the list, but he doesn't.

That doesn't mean he shouldn't be in Canton - he absolutely deserves that honor - but it does point strongly towards "overrated."

12
by Theo :: Mon, 08/09/2010 - 5:16pm

To be honest, I'm only watching this game since the '99 (Rams) season. So no, I haven't seen Smith run.
In the meanwhile, I've seen a lot, I've seen old games. Smith was a runner that could run into a 2 yard hole and make it a 3 yard hole by just pushing forward without the punishment, that's the first thing I noticed.
If there was just more than a hole, Smith could run this into a 4 or 5 yard gain.
This is someone who reads lines and defenses extremely well and makes the most out of every run.

40
by socctty :: Tue, 08/10/2010 - 2:09am

In the spirit of putting a number on things, and of being devil's advocate, how do you measure how "rated" (over-rated, under-rated) someone is? Rhetorical question. :-D

9
by vinyltoupee (not verified) :: Mon, 08/09/2010 - 4:50pm

What separated Emmitt was not his speed or his elusiveness. if there's one traditional attribute that he had in spades it was vision. But more than any of these things, what made Emmitt great was his will to win, which gave birth to his consistency. On almost any given year there were other backs more thrilling than Emmitt, and he would not top any all-time list of individual attributes, but he was a workhorse who produced at a high level for such a long time that we took it for granted. He was the Lou Gehrig of running backs.

10
by Joseph :: Mon, 08/09/2010 - 4:57pm

Better baseball comparison might be Cal Ripken--Gehrig was really great, too.

89
by Jacob Stevens (not verified) :: Tue, 08/10/2010 - 5:30pm

Yes. Maximizing the blocking in front of you is a skill. Not all backs have that skill in equal proportion. He was an extremely good player.

108
by Xeynon (not verified) :: Tue, 08/10/2010 - 9:29pm

As an Eagles fan who suffered a lot of heartache at the hands of Emmitt Smith, I wanna put the kibosh on any kind of Freakonomics-style counterintuitive statistical analysis that argues he was overrated. The guy was great - maybe not the best ever (but then I don't know many people who claim that), but indisputably great. He didn't have the flash or big play ability of guys like Sanders or Chris Johnson, but he was tough, relentless, and almost never went down behind the line of scrimmage - he had an uncanny ability to know when there was nothing there and just take the 2 or 3 yards he could get rather than dancing in the backfield. He rarely fumbled, almost never got hurt, and had tremendous stamina, getting better and better as games wore on. Those are all important traits for a running back to have and Smith was among the best I've ever seen at them, even if he wasn't the biggest, fastest, or most agile.

All those backups he had gained a lot of their yards after he'd gone out of the game in the 4th quarter having run for 155 yards on 31 carries or whatever and thoroughly beaten down and worn out the defense. So I agree that this is a *very* flawed measure.

128
by chemical burn :: Wed, 08/11/2010 - 2:08pm

Agreed - as an Eagles fan, Smith deserves all the respect in the world. I still have nightmares of him chugging through Mark McMillian in the 4th quarter and falling forward for another very meaningful 2 yards.

11
by James (not verified) :: Mon, 08/09/2010 - 5:13pm

Did anyone consider that goalline carries would seriously decrease a player's average YPC? I think you need to do the analysis over again after taking out all carries within the 5 or possibly the 10 yardline to make sure we are measuring apples to apples. After all, if a player only gets 20 carries a game and three of those are goalline carries from the 2, that's enough to drop your YPC from 4.0 to 3.7 just because you cannot get more than two yards.

Otherwise, you also have to consider any team that has a specialized goalline back, or even just another back that takes some of those carries (including other short yardage carries). That could seriously affect any RB's YPC, particularly when we are only considering a few tenths of a yard in most cases. Think about how this would affect Warrick Dunn, who spent a significant amount of time with Alstott handling short yardage situations, or Jerome Bettis who became a goalline back almost exclusively towards the end of his career.

Along similar lines, you are punishing a team for having any sort of RB depth by comparing people to their backups. No one is surprised Sanders is at the top of that list because those Lions teams were terrible when he was there, so of course he was better than the players around him.

Then again, Sanders gets a huge benefit to his numbers because he quit during his prime and didn't drag down his numbers by playing past his best years like others did. Emmitt Smith on the other hand did the complete opposite and played as long as a team would let him, even going to the runningback graveyard of the Cardinals. That alone could easily make up 0.2 yards per carry and propel Smith to 5th on your current list. Why don't you compare everyone's best three or four year window to get a more reasonable comparison?

Then again, if you were really serious you would compare how all these runningbacks did in the same down situations such as 1st and 10, 2nd and 5, etc because situation affects stats as well. A better team would presumably be closer to the endzone more often than a bad team, making it more difficult to put up longer runs, etc. For instance, it's reasonable to think say 80% of Sanders' carries came between the 20s compared only 75% of Smith's runs, not because of the two running backs but just because the Cowboys teams were better so the defense gave them better field position, the passing game was more effective and moved the ball, etc.

That's all I have off the top of my head, but I'm sure there's more to say about how such an analysis just breaks the surface of how to compare running backs.

14
by Temo :: Mon, 08/09/2010 - 6:14pm

Tiki Barber and Jim Brown also quit early... ie, all of the top 3 RBs on the list.

21
by spenczar :: Mon, 08/09/2010 - 8:07pm

I think that may be a result of teams drafting/trading for young RB talent when their star back gets over 30. Raiders getting Bo Jackson while they have Marcus Allen, that sort of thing. So running backs who have long careers are penalized twice when this stat is averaged over the entire career - once by dragging down their success in youth with the slow decline of age, and second by increasing the quality of the backups as they age.

24
by Marko :: Mon, 08/09/2010 - 8:31pm

"Did anyone consider that goalline carries would seriously decrease a player's average YPC?"

I have thought about that, too, with respect to YPC for players like Jerome Bettis (as you mentioned) and John Riggins. I recall Bettis had a game towards the end of his career where he had something like 4 carries for 5 yards, but scored 3 TDs.

57
by James (not verified) :: Tue, 08/10/2010 - 10:35am

Found it! On Sept. 12, 2004, Jerome Bettis had 5 carries for a grand total of 1 yard (a 0.2 YPC) and three, yes THREE!, touchdowns. That must be some sort of record for TDs per total yards.

Bettis' carries all came in goalline situations because the Steelers had a new starter in Duce Staley (who?), and the three TD runs were all from the one, meaning he had 3 carries for 3 yards and 3 TDs, and two other carries for -2 yards. However, Staley was hurt midseason and Bettis became the starter again.

This provides two excellent examples of how this isn't a great system of comparison. Within one season Bettis went from an exclusive short yardage back, killing his YPC, to a full-time starter with a 3rd string backup who probably was terrible (thus providing an excellent YPC differential boost), considering the Steelers traded for Staley before that season specifically to be the starter while they had the 6th highest rusher in history on the roster.

Oh, in addition to my earlier problems above, there's the whole "3rd down back" problem to address. Think of RB situations like the Jets with Jones/Washington, Cowboys with Barber/Felix Jones, Giants with Jacobs/Bradshaw, and Vikings with Peterson/Taylor. The first is the guy who gets the majority of carries, the power back, the short yardage back. The second is the change of pace, speedy guy who plays much more often on 3rd and long. Think that might be unrepresentative of the YPC stat - either because the backup is faster and has more boom-or-bust type swing plays or because the defense is willing to allow a 5 yard carry on 3rd and 8?

Michael Turner averaged 4.5, 6.0 and 6.3 YPC as LT's backup on the Chargers. Perhaps the defense was focusing on the pass whenever LT was out of the game?

58
by sn0mm1s :: Tue, 08/10/2010 - 10:40am

One game doesn't define a career usage pattern though. For a good portion of his career Bettis was *the* guy. Plenty of those seasons with 300+ carries Bettis' rushing average was below 4.

82
by James (not verified) :: Tue, 08/10/2010 - 3:24pm

I understand that that one hyper extreme example didn't take down Bettis' average all by itself, but it's an example of how the YPC stat, particularly how it was used here, is not a sufficient metric to say Smith or anyone else was great or not great.

83
by sn0mm1s :: Tue, 08/10/2010 - 3:29pm

If you assumed that all of Smith's TDs were 1 yard plunges and they all took 4 carries (in other words stuffed for no gain 3 times and score on the 4th) and you removed all of those carries from his career totals his average would still be less than Barry's. Sure, Emmitt probably had more short yardage carries but he also had a TE and FB blocking for him and that doesn't come close to describing the difference between the two.

62
by zlionsfan :: Tue, 08/10/2010 - 11:02am

actually, no, the Lions weren't terrible when Sanders was there. They made the playoffs five times in a seven-year stretch, including an NFC Championship appearance in 1991. They were as high as 8th in DVOA on two occasions, 1995 and 1997. (No play-by-play for 1991 yet, so we don't know where they fall. I wouldn't be surprised if they were about 8th that season as well.)

Anyway, most of these arguments can be turned in the other direction. Emmitt's YPC was lowered by more goal-line attempts; Barry's was lowered by asking him to run short-yardage plays behind a run-and-shoot O-line. Emmitt's YPC was hurt by playing through (past?) the end of his career; Barry's was hurt by leaving behind prime years (and perhaps even an opportunity to play behind a better line). Emmitt's YPC difference was lower because his team had better depth; Barry's YPC difference was higher because he was that much better.

This is simply the '90s edition of the Debate That Cannot Be Named. Two outstanding RBs in two completely different situations ... both with excellent careers, but excelling in ways that appeal differently to different fans. All sorts of arguments can be made as to why one is "better" than the other.

It should be enough to agree that they were both special.

16
by ceolaf :: Mon, 08/09/2010 - 6:16pm

The key word here is "just."

Emmit Smith was at the very least a good back, and a durable back. The question is how GREAT a back he was. Was he ever the best back in the NFL -- as opposed to the one with the most yards.

He carried the ball a LOT. He got a lot of yards for those carries behind some excellent lines. There is no question about this.

So, the questions we need to ask:

* The old question of the value of durability, both in a season and in a career.

* What we expect of perhaps-great players when surrounded by other great players. That is, how much do credit their performance to others, and how much do we credit them for sacrificing their game for the team? How does this vary by sport and position?

Clearly, a great receiver is both benefited and hurt by a great reciever-teammate, but only benfited by a great QB-teammate. That's obvious. I wonder how many running backs were truly great on truly crappy teams? I wonder how many lousy running backs appeared very very good on excellent teams?

I've always thought Irvin, Aikman and Smith overrated, just as I thought Boomer and Icky to be overrated. Offensive skill players **always** benefit greatly from their lines, and we know that these folks had great great lines. We never credit Theisman, Rypien, Schraeder or Doug Williams, though they had excellent seasons with excellent offensive lines.

I think that it's hard to conclude that Smith was anything but overrated because even when we acknowledge that teams have great offensive lines, we still reward their skill-player teammates for their production (e.g. with pro-bowl appearances). Backs SHOULD give their lines watches, cars and houses, because they owe their greatness to their lines.

It is the nature of this sport we love that virtuall *all* great running backs are overrated. But that doesn't mean that they are *just* products of their lines. In football -- and here on football outsiders -- we recongized fit and team context. Smith produced great numbers with his line. On his plague, like that of all other great running backs, we should list his offensive linemen. They accomplished these things together.

17
by JSA (not verified) :: Mon, 08/09/2010 - 6:49pm

I wonder how much better Smith woudl fare if this just included his Cowboys years. I don't think his Cardinals years helped him get into the Hall of Fame at all.

19
by sn0mm1s :: Mon, 08/09/2010 - 7:38pm

I actually sent nearly this exact same research to FO and PFR several years ago (when PFR made their stats easily downloadable and in a single file). At the time Tiki was still number one Tomlinson was at the bottom (although he isn't included here). Also, was Bobby Mitchell's stats included? Most sites list him as a WR but he took a significant number of carries for the Browns (and was better than Jim for most of them). When I ran this query I included WRs as well because of players like Mitchell and Eric Metcalf etc. etc.

I also measured YPC vs. the league excluding the stats of the player I was analyzing. The guy who ran against the toughest NFL over his career (measured by YPC), believe it or not, was Barry Sanders.

Other random bits of info that I discovered running queries.

Tony Dorsett was a horrible fumbler. He is the only back with any significant carries (1800+) that has more fumbles than TDs. He makes Adrian Peterson look good. IIRC he fumbled every 30 or 35 carries.

Earl Campbell never caught a TD.

Curtis Martin really held on to the ball. IIRC he fumbled once every 120 carries or so.

The guy who took the highest percentage of his teams non-QB carries (just WR and RB) was Barry Sanders. I always thought this odd because people claim he was removed from the game all the time for short yardage when the stats don't back this up at all. IIRC he took about 85% of his team's carries.

Barry Sanders is the only player to score all of his team's non-QB rushing TDs 5 separate occasions. Which I again found odd since that meant that in 5 of his 10 seasons no one other than a QB scored a rushing TD for the Lions.

EDIT: In the way I ran the stats I think Barber would be ahead of Barry still. I think I looked at who had the highest percentage increase over his teammates rather than absolute increase.

23
by Spielman :: Mon, 08/09/2010 - 8:23pm

Lots of old backs make Peterson look good. Running back fumbles have declined dramatically over the years. Peterson's fumble numbers compare pretty well to Walter Payton's, and he did not have a reputation as a fumbler.

Meanwhile, QB fumbles have actually gone up a bit, as a function of the number of times they get hit.

26
by Marko :: Mon, 08/09/2010 - 8:36pm

There was a thread a few years ago that discussed this. Part of the reason is that botched QB-RB exchanges used to be counted as fumbles by the RB but now are counted as fumbles by the QB.

31
by Spielman :: Mon, 08/09/2010 - 10:03pm

Interesting. I must have missed that thread. I'll have to look for it.

32
by Eddo :: Mon, 08/09/2010 - 10:14pm

Here's a link to a set of comments from within the Quick Reads from this past year's conference championship round.

33
by Eddo :: Mon, 08/09/2010 - 10:18pm

And here's a link to a P-F-R blog post on historical fumble rates (referenced in the thread I linked).

29
by Marko :: Mon, 08/09/2010 - 8:48pm

"Earl Campbell never caught a TD."

A random fact that I remember from seeing on the highlights way back in 1978 is that the first TD of Campbell's career was originally credited as a 73-yard TD reception (it was a swing pass type-play), but was changed to a rushing TD the next day because it was ruled a lateral rather than a pass. That changed his rushing total for his debut from 64 yards to 137 yards. This obviously helped his YPC, too. Here is a link to the box score from the game.

http://www.pro-football-reference.com/boxscores/197809030atl.htm

22
by Thok :: Mon, 08/09/2010 - 8:22pm

Emmitt Smith tends to get hurt a bit when compared to Barry Sanders because they have slightly different aging patterns. Emmitt's peak was early in his career, while Barry's peak was late, and people remember the older versions of the players rather than the younger versions.

28
by sn0mm1s :: Mon, 08/09/2010 - 8:42pm

Well, Barry had a career year when he finally got to play in a standard offense - but Barry never really had a peak or decline in his career.

47
by t.d. :: Tue, 08/10/2010 - 3:44am

The 2000 yard season (his second-last year) is not in line with his average seasons

25
by MJK :: Mon, 08/09/2010 - 8:31pm

Another reason why the guys at the bottom of the list are a little unfairly treated...a lot of times, the "change of pace" backs that come in when the star comes out do so in obvious passing downs, or when the defense is especially guarding against the pass...and they tend to rack up good ypc when they run draws and such.

The classic case I know if is Kevin Faulk in New England...he almost always has a much higher ypc than any other back on the team, due mostly to his draws and 4 yard sweeps on 3rd and 5...but fails horribly when the Pats try to use him as a starting "featured" back.

I like this kind of analysis, though. Albeit with some error, it essentially tells us how much better, in the same offense, a given player is relative to a "replacement level player" on a per-play basis.

27
by JIPanick :: Mon, 08/09/2010 - 8:40pm

What's with all the love for "Boom-and-Bust" Sanders? I thought FO was an outpost of sanity against the YPC as a be all and end all mentality; Guess I was wrong.

I always figured that the "Greatest Ever" Cowboys line of the early 90s was quite a bit overrated; The best line ever doesn't let their HOF QB get concussed into retirement.

It's also worth noting that that line came and went in a hurry; the legendary "90s Cowboys" lines covered no more than half of Emmitt's productive seasons.

While running backs in general are overrated compared to the offensive line (the most overrated position in sports, IMO) I see no reason to credit the line with Emmitt's longevity.

As for his productivity:

"Of course, we can't just take Smith out of the Cowboys offense in his prime and insert him in a mediocre one to get reliable new statistics."

Even if that were possible it would be the wrong tack. How many of the guys on that list played behind bad lines during their best years? Or are you arguing the 1992 Cowboys had a line so good it makes the Electric Company, the Hogs, and the Greatest Show on Turf look mediocre?

Overall, yes, Emmitt Smith was a product of his line. So was Barry Sanders. So was Jim Brown. So was Walter Payton. So are ALL running backs, Emmitt no more than most.

30
by sn0mm1s :: Mon, 08/09/2010 - 9:08pm

FO is a little contradictory though when it comes to rating RBs. They devalue long runs - yet say long runs are almost exclusively due to the talent and skill of the RB. Their DYAR stats don't make sense when you apply them to players that break long runs.

Take this year for example. Chris Johnson led the league in DYAR but not by much. According to FO, a *replacement* level RB for the Titans would've been expected to rush for 1663 yards (2006 yards - 343 DYAR). So, a scrub replacing Chris Johnson would be expected to have a top 40 rushing season in the history of the NFL. That exact same RB replacing Ryan Grant (#2 in DYAR) would be expected to rush for exactly 1000 yards. This is consistently the case with Barry Sanders on FO too.

Now, someone like Terrell Davis (who was much more a product of his offense than Smith ever was) could gain *less* per carry than his offensive line's ALY has a huge DYAR. I find it hard to believe that if one of your stats says a player is rushing for *less* than his oline is giving him can be a DYAR monster.

Do the same comparison with Grant and Johnson. Grant gained .11 yards per carry above the ALY of the Packers meaning he generally got what his line gave him. Johnson gained 1.59 yards more than his ALY a pretty big discrepancy.

34
by Tim Wilson :: Mon, 08/09/2010 - 11:17pm

To use YPC as a chief measure here seems utterly insane to me.

Emmitt's greatness derived in large part from his ability to be so good over so long a period of time. He is the NFL's all-time rushing leader. His durability and longevity were freakish even by Pro Bowl running back standards.

Could detractors please just read the PFR post "Emmitt Smith: More Than Just Very Good," so we don't have restate all the arguments here?

http://www.pro-football-reference.com/blog/?p=5541

35
by sswoods (not verified) :: Tue, 08/10/2010 - 12:14am

The Barry Sanders angle is the intriguing area of this conversation to me. When I think of the greatest RB's of all time, I firmly believe it's a three horse race between Jim Brown, Walter Payton, and Eric Dickerson. (I vote Brown.) My perspective on Sanders is that he hurt his team an awful lot with so many carries that went for negative yardage, many times for significant loss. 2nd and 7 doesn't thrill anyone but it is certainly better than 2-12 or 2-15. To his credit he was so explosive that he would often make up for those losses on subsequent carries, but it still adds a significant burden to any offense. Granted, he did not have an elite line in front of him at any point of his career - though he did have at least one pro-bowler every year but his first and last seasons - but his style of play is as much to blame for that as the line play. Sanders was the most exciting back of his day, no question. His booms were certainly larger than Smith's; his busts were also larger than Smith's. For my money, Sanders was an elite back but not quite in the class of the top three (I don't believe Smith was either).
As for being relevant to the topic, I do not believe Smith is overrated. Aside from hardcore homers, who really argues that Smith was the greatest of all time? The conversation almost always starts "Emmitt was a great back, but ...". It appears to me - and granted, my circle of exposure may be limited - that the consensus is that he was a great back but not elite and benefited from superior line play (at least in his early years) and outstanding longevity. If I'm accurate on that view of the concensus, then no, he's not overrated at all.

37
by sn0mm1s :: Tue, 08/10/2010 - 12:48am

His negative runs aren't as prevalent as you would believe. He averaged 20 carries a game. 3 were for -2.5 YPC 17 for 6.3 YPC. His style didn't hurt the Lions. Since 1957 (the Lions' last championship) without Barry they make the playoffs about once a decade. One of those playoff appearances was with a losing record and one was with a .500 record. With Barry, they made the playoffs 5 of 10 years. If anything, blame lack of success on a poor D and inconsistent QBs. It also doesn't help that for most of Barry's career he didn't have a FB or TE to help block. Once he *did* get a standard offense installed he rushed for 2000 yards averaging over 6 yards per carry.

Brown and Glover both came into the league in 1985. Here are the rushing stats for the years before Barry got there

League rank.
Carries - Yards - TDs - YPC
1985 18-26-20-28
1986 16-15-16-19
1987 25-25-23-24
1988 26-27-28-28

So, Brown and Glover in their first 4 years couldn't get a single rushing stat in the top half of the league. They were dead last in YPC twice. Dead last in rushing TDs in 1988. Were in the bottom 5 in the league in rushing yards 3 out of those 4 seasons. Now, we add Barry into the mix.

1989 20-8-1-1

Lions jump from worst to first in rushing TDs and YPC. They jump up 19 places to become a top 10 rushing team even though they were still in the bottom 1/3 of the league in carries and that was with a rookie QB starting half the games. Barry made that line look good. No other back in the history of the NFL has made that much of an impact.

Brown played in such a different era you really can't compare the two. I mean, if an NFL Champion (1962 Packers) with 10 HOFers can get beat by a college senior all-star squad I have a hard time believing that the quality of play or competition is anywhere close to what it is now. Once the league was integrated, players were truly professional and PEDs had hit the scene the game became modern. Prior to that it was a 1/2 step above college. I don't see Tebow or Bradford leading a college team over the Saints - who have no sure HOFers.

48
by t.d. :: Tue, 08/10/2010 - 3:51am

No one considered Payton to be in Brown's league at the time he broke Jim Brown's record. It was very similar to the Smith-Sanders comp in that Brown was the electrifying back whereas Payton was considered to be the grinder. Then, the season after Payton broke the record, the Bears won the Super Bowl, and made a music video, and then Payton died tragically young, and now people forget anybody ever disparaged his credentials

49
by tuluse :: Tue, 08/10/2010 - 4:49am

I wasn't alive at the time, but I have seen video of Jim Brown saying he preferred Payton get the record.

Also, no one is debating styles here. Grinder vs electric is irrelevant. We're talking about how much a player was helped by his teammates.

I don't think anyone suggested that Walter Payton was unduly assisted by Bob Avelleni, James Scott, or the offensive line.

54
by t.d. :: Tue, 08/10/2010 - 9:19am

He preferred Payton get the record rather than Franco Harris, who he personally disliked

65
by Marko :: Tue, 08/10/2010 - 11:22am

I don't think that's accurate. Sure, many people considered Brown to be a better RUNNER than Payton. But I don't think many people thought Payton wasn't in Brown's league as a runner. And many people believed (and still believe) that Payton was the greatest RUNNING BACK in league history, considering his all-around ability not just as a runner, but also as a receiver, blocker and passer. As they say, he could do it all. No less an authority than John Madden said on many occasions that pound for pound, Payton was the greatest player in league history.

Also, Payton wasn't considered a grinder, although he could grind it out when needed; he had numerous spectacular and electrifying runs. And the idea that people don't disparage Payton's credentials because of the Super Bowl Shuffle is laughable. They weren't disparaging him in the first place. As mentioned above, the one being disparaged at the time was Franco Harris, not Payton.

68
by Dean :: Tue, 08/10/2010 - 11:34am

I agree with you on this. I was in my teenage years in the 80s, and saw Peyton in/after his prime.

The discussions at the time of who was the greatest RB of all time had 2, possibly 3 answers. Jim Brown, Walter Payton, and the occasional Eagles homer who would say Steve Van Buren. OJ would quickly enter the conversation as being right up near the top, but slightly behind those two.

As for me, in my lifetime, Payton is still the greatest RB I ever saw, with Barry Sanders and Bo Jackson close. Marshall Faulk was really special, too. Adrian Peterson has a chance to be special, too. Maybe Marcus Allen is in that group. Then there's everybody else.

Emmitt Smith is somewhere behind Curtis Martin, John Riggins, Earl Campbell, Terrell Davis, etc. but somewhere ahead of Priest Holmes, Shaun Alexander, etc.

79
by Tim Wilson :: Tue, 08/10/2010 - 2:35pm

To me, some of the arguments for Emmitt over Barry are similar. Emmitt was a better all-around back-- he averaged 55 receptions a year on the Cowboys, and he was a superb blocker. Sanders never excelled in either of those areas. And his boom-or-bust running style was less dependable in an offense. It led to some sensational long scoring plays, but also led to many failed drives due to losses on running plays.

87
by t.d. :: Tue, 08/10/2010 - 5:04pm

The argument I'm making is that being a member of one of the most celebrated Super Bowl championship teams and then dying tragically young enhanced Payton's reputation, and I think that's obvious. Furthermore, Brown was considered the best player on a championship team whereas the Bears' offense was never considered among the best in the game. The conversation for best player ever includes Jerry Rice, Jim Brown, and quarterbacks

91
by Marko :: Tue, 08/10/2010 - 5:49pm

I think your argument is severely flawed for several reasons:

1. Walter Payton's reputation was unquestionably (except by you obviously) established long before his team became "one of the most celebrated Super Bowl championship teams" and before he died tragically young. This point hardly seems worthy of serious debate.

2. Your point about Brown being "considered the best player on a championship team whereas the Bears' offense was never considered among the best in the game" makes no sense. I agree that Brown was considered the best player on a championship team. But what does that have to do with anything? Payton also was "considered the best player on a championship team." Again, what does that have to do with anything?

3. Your comment about the Bears' offense is similarly confusing because (a) it seems irrelevant to the discussion; (b) while the Bears' offense for much of Payton's career was not considered great, during the Bears' championship run it was a lot better than people give it credit for; and (c) even if we accept your comment about the Bears' offense as accurate, it seems to me that would serve to enhance Payton's reputation rather than detract from it because he was able to accomplish so much despite being keyed on by the defense (with 8 and 9 in the box) for virtually his entire career.

4. While you are certainly entitled to your opinion, others who are far more knowledgeable about the NFL than you or I (such as former coaches, executives and players) undoubtedly would include more players in the conversation for best player ever, including Payton. In addition to Payton, a serious conversation about the best player ever would have to include great defensive players such as Lawrence Taylor, as well as great offensive linemen. It is ridiculous to limit the list as you suggest.

93
by Eddo :: Tue, 08/10/2010 - 6:04pm

In support of your points #1 and #2, Marko, the '85 Bears were seen as being able to finally win a championship for Payton.

96
by t.d. :: Tue, 08/10/2010 - 7:14pm

I lived in Chicago for Payton's entire career. He was great and beloved. He was beloved as a local icon and a great human being, not as someone who anybody thought was the greatest player at the time or who ever lived. He won one MVP in 13 seasons (Jim Brown won three in nine seasons, his first two and last seasons of his career). If one were to compare MVP shares between the two, it wouldn't be close- the consensus of Brown's contemporaries was that he or Johnny Unitas was the best player in the game. No such consensus existed for Payton. If you think someone being a part of one of the most celebrated teams of his time, or his dying young, has no effect on his reputation, I don't know what to tell you (though I will say Mike Singletary, Richard Dent, and Dan Hampton would all have been in the conversation as 'best player on the 1985 Bears', whereas I don't think there was any question about who was the best player on the Browns).
As for not including defensive players in the discussion of being the best players, I'm reflecting the conversation (although I agree LT is often cited as the token defensive representative)

99
by Jeff Fogle :: Tue, 08/10/2010 - 8:16pm

t.d....it's hard to believe that Chicago was completely out of sync with the rest of the world in terms of how they viewed Walter Payton...but, I didn't live there so I'll have to take your word for it.

In terms of how his achievements PRE-1985 were catalogued...from wiki:

*"The Chicago Bears drafted Payton in the first round of the 1975 NFL Draft, as the fourth overall pick."

*"During the 1976 NFL season, Payton rushed for more than 1,000 yards and scored 17 touchdowns. After the season, he was selected to play in the 1977 Pro Bowl, where he was declared the Pro Bowl MVP."

*"The next year, he rushed for 1,852 yards and scored 16 touchdowns, becoming the league’s leading scorer for the season. He earned numerous awards that season, including the Associated Press and Pro Football Writers of America’s Most Valuable Player awards. His most memorable game of the 1977 NFL season was against the Minnesota Vikings on November 20. He rushed for a then-record 275 yards on 40 carries while combating the flu, breaking the previous record of 273 yards held by O.J. Simpson."

*"By the end of the decade, Payton had received additional accolades for his exploits as a blocker, receiver, emergency punter, and quarterback."

So...IN THE 70'S....he entered the league as a top five draft pick, he reached the Pro Bowl in his second season, in his third season he had a game where he rushed for 275 yards to break the league record, and he would subsequently made headlines for his versatility and his ability to impact the game in a variety of ways. He was an extremely celebrated player on the national level. Years before the Bears won the Super Bowl...

You said he was a "grinder" in an earlier post. From wiki:

"One of Payton's signature maneuvers was the "stutter-step," a high-stepping, irregularly paced run. He developed this as a way to distract his pursuers during long runs, saying that it startled them into thinking and gave him some advantage over players who were actually faster runners. In his autobiography, he likened the stutter step to a kind of "option play": when he was stutter-stepping, defenders would have to commit to a pursuit angle based upon whether they thought he would accelerate after the stutter-step, or cut—he would read this angle and do the opposite of what the defender had committed to. He reinvented the practice of stiff-arming his tacklers, which had gone out of favor among running backs in the 1970s. At times, he used his high school experience as a long jumper to leap over his opponents..."

It's a shame the city of Chicago missed out on this, and disparaged his credentials because he was a grinder who only later earned fame piggy-backing glory on a celebrity football team before an untimely early death. Could you please do what you can to inform them about everything they missed, so they can get up to speed with the rest of the football world that marveled at his many accomplishments, celebrated his induction into the Pro Football Hall of Fame in 1993, and was saddened by his untimely death in 1999.

102
by t.d. :: Tue, 08/10/2010 - 8:40pm

All I said is he was never mentioned as "the greatest running back of all time" at the time, because people still remembered Jim Brown. If you think being a great player is equal to the greatest player ever, you have a math problem. And compared to Jim Brown, he was a 'grinder'

104
by Jeff Fogle :: Tue, 08/10/2010 - 8:53pm

Among the other things you said:

"Payton was considered to be the grinder."

"the season after Payton broke the record, the Bears won the Super Bowl, and made a music video, and then Payton died tragically young, and now people forget anybody ever disparaged his credentials."

*The argument I'm making is that being a member of one of the most celebrated Super Bowl championship teams and then dying tragically young enhanced Payton's reputation, and I think that's obvious."

"He was beloved as a local icon and a great human being, not as someone who anybody thought was the greatest player at the time or who ever lived."

The Sporting News used the phrase "high stepping" instead of "grinder" in this interview just before he was inducted into the Hall of Fame. It's hard to imagine a high stepping grinder.

http://www.sportingnews.com/archives/payton/article1.html

His reputation was strong before the Super Bowl win, and he was a first ballot Hall of Famer long before he died. He was certainly seen AT THE TIME as one of the greatest players OF THE TIME, which is how you become a first ballot Hall of Famer. Nobody in Chicago thought he was the greatest player at the time? The other football players that went in with him that year were Larry Little and Dan Fouts.

Not going to dispute your comparisons between Brown and Payton. Tough to compare out of eras, and Brown certainly earned his place in history. Payton wasn't a grinder who's reputation was enhanced by his early death. He was an electric first ballot Hall of Famer who went in by acclimation six years before he died.

105
by t.d. :: Tue, 08/10/2010 - 8:59pm

His reputation was never "the greatest player of all time" until he became a sentimental favorite.

111
by Jeff Fogle :: Tue, 08/10/2010 - 9:57pm

Are you suggesting Steve Sabol didn't notice any of this stuff until after Payton had died?

"Greatest Running Back: Walter Payton
Jim Brown was the greatest ball carrier, but no one ever played the position of running back as completely as Payton. He was a crushing blocker. I saw him lift blitzers off their feet. When it was required, he was an effective decoy who followed through convincingly on all his fakes. He once led the Bears in kickoff returns. He's Chicago's all-time leading receiver. When he threw passes, he completed most for touchdowns. The Bears threw enough interceptions for Payton's skill as a tackler to be noticed and, in addition to all of that, he missed only one game in his entire career. And when he retired in 1987, he had carried the ball more times for more yards than any player in history."

115
by Marko :: Tue, 08/10/2010 - 11:25pm

He is suggesting exactly that. Oh, and that Steve Sabol also must have been unduly influenced by the cultural impact of the Super Bowl Shuffle. John Madden evidently was similarly influenced.

The awesomeness of the Super Bowl Shuffle also explains why "Speedy" Willie Gault is sometimes considered the greatest WR ever, and why Refrigerator Perry is universally acclaimed as the most devastating and best DT ever.

64
by bravehoptoad :: Tue, 08/10/2010 - 11:21am

When I think of the greatest RB's of all time, I firmly believe it's a three horse race between Jim Brown, Walter Payton, and Eric Dickerson.

O. J. Simpson was the most jaw-dropping runner I ever saw, and he definitely played on the weakest team of any great running back.

36
by NHPatsFan (not verified) :: Tue, 08/10/2010 - 12:44am

Is there a way to cross-tab this analysis with utilization patterns for running backs? The way the big bruising guys are all clustered is suggestive that there's more story to dig into here.

38
by sswoods (not verified) :: Tue, 08/10/2010 - 1:21am

Forgive me if I came across as arguing that the Lions had decent O-Line play. I certainly don't believe that, but I do believe that the line play wasn't as putrid as commonly argued. Compared to Dallas of the early to mid 90's, yeah. Compared to, say, league average, then no. Regardless, I agree that Barry made his line look better than it really was.
I'm also not suggesting that every play was either a 10 yard burst or a 10 yard loss. I am saying that a 2 yard loss on 1-10 hurts more than a no gainer. It changes strategy, play-calling, etc. 2nd and 9 can still be a running down - 2nd and 13 is most likely a passing down. As you've pointed out, inconsistent QB play (to put it kindly) makes that scenario even less appealing. And those losses came in bunches, as do the big plays. It isn't every game that he had a 30 yard run, and it wasn't every game he had a couple 6 yard losses. One would expect regression (progression?) to the mean over time. Often times Barry's style would result in big plays, but a failed gamble kills a drive. There are typically only 11-12 of those per game, and touchdown drives are relatively rare (2 per game or less for most of his career, league average - and most of the time the Lions weren't average).

I don't know anything about the Packers game you refer to - I'll look it up - but I'm not in the camp that believes the league was so watered down compared to today's game. Different argument for a different time. I stand by my original opinion, Brown/Dickerson/Payton in the first tier, Sanders/Smith/Simpson (and a few others) in the next.

39
by sn0mm1s :: Tue, 08/10/2010 - 1:36am

The game was called the Chicago Charities All Star game. It was held annually starting sometime in the 1930s and ending in the 1970s. The 1962 Packers (beat in 1963 by the all-star squad) was the last NFL Champion to lose to the college players. The game was discontinued when it became obvious that the college players weren't going to win any longer and NFL players were making enough money that they didn't want to risk injury. IIRC college stars last won about 1/3 of the games from inception to their last victory. I know that while Brown was playing the Lions also got beat by the all stars and I believe the Browns got beat by the all stars prior to Jim arriving.

For me the order is 1) Sanders 2) Payton 3) Brown.

53
by Mr Shush :: Tue, 08/10/2010 - 9:09am

All depends what sort of top three you're after, doesn't it?

If I had a time machine and were picking an all-time team to play a match or series of matches under today's rules and conditions (and with time for old-time players to do modern weight training, learn modern systems, etc.) and had to have one every down first string RB, I think it would probably be Marshall Faulk, because passing these days is just so damn important. If the game were to be played under 70s conditions, or those of some other massively run-dominated era, I would lean towards Brown or Simpson. If my Spacetimetrotters might have to play all sorts of games under all sorts of rules and conditions against all sorts of opponents but I was picking a three back rotation to give me an appropriate player for every down, distance, opponent and conditions-set it would be Faulk, Sanders and one of Brown and Simpson. But if I was allowed only one back for the same range of situations it might be Payton. The thing is, changing the parameters can produce a fair number of different answers, but I'm not sure it's ever going to lead to the answer being Smith.

60
by sn0mm1s :: Tue, 08/10/2010 - 10:52am

Faulk was only great when he had 4 other potential HOFers on offense with him. On the Colts he was Reggie Bush with more carries and less receptions. I believe that Faulk's rushing average was 3.7 or so with the Colts (and it wasn't like this was a small sample size). Other than his receiving stats his last year with the Colts (catching dumpoff passes from Manning) you could easily say Faulk was a huge disappointment. By any standard metric or DVOA/DYAR Faulk just wasn't good if he was the only option.

100
by Mr Shush :: Tue, 08/10/2010 - 8:27pm

Faulk averaged 10y/r or better over a season three times with the Colts, including in the year you describe as him catching dump-offs from Manning. If he was turning dumpoffs into that sort of yardage, his play in the passing game was clearly pretty incredible. He was selected for three pro bowls as a Colt, and led the league in yards from scrimmage in 1998. No, his yards per carry in Indy weren't impressive. Then again, those Colts teams were pretty crappy: 32-48 over his time there. Offensive linemen who started 8 or more games in at least one season for the Colts in that time include such luminaries as Joe Staysniak, Tony Mandarich and Eric Mahlum. He got a couple of years at the end of Will Wolford's career (while he was still playing at a pretty high level, and a couple at the very start of Tarik Glenn's. That appears to be about it, for anything resembling quality blocking. In any case, my point is that Faulk was a far better receiver than anyone who was as good or better a runner, and a significantly better runner than anyone who was remotely comparable as a receiver. He was also, probably, the greatest pass-catching back of all time. My contention is that in the modern game, a very good but not elite runner who is a coverage miss-match for most starting corners, never mind nickel backs, safeties or (God forbid) linebackers is probably more valuable than a truly elite runner who's only a pretty good receiver (Brown, for example).

107
by sn0mm1s :: Tue, 08/10/2010 - 9:18pm

I know his stats and his situation. However, Faulk didn't vault into elite back status until teams couldn't focus on him. That was all I am saying. He wasn't a good runner in Indy no matter how you slice the stats. He was an above average receiving back but it was only his last year there that he was really good.

117
by Whatev (not verified) :: Wed, 08/11/2010 - 12:49am

Well, he did preface his post with "If I had a time machine and were picking an all-time team..." so presumably he has as many Hall of Famers as he needs around him.

140
by BigCheese :: Fri, 08/13/2010 - 12:57am

I couldn't disagree with you more. If you want a RB for the modern game the answer is clearly Water Payotn over Marshall Faulk.

I think we can both agree that Payton is a far better runner than Faulk, which is still pretty important in today's game. He's also the better blocker and, if it comes up, passer by an huge margin.

So, you're basically basing your argument on Faulk being so superior as a reciever over Payton that it makes up for al the other areas Payton is better at. However, Payton was one hell of a reciever as well. How good of one? Well, he's still the all-time Bears leader in receptions and third in yards. If you exclude players which played part of their carrer as WR, TE or FL, the list of backs with most recieving yards goes:

Marshall Faulk
Larry Centers
Marcus Allen
Tiki Barber
Roger Craig
Herschel Walker
John Williams
Earnest Byner
Walter Payton

Oh, and Payton has a higher Yards per Catch than Faulk (9.2 to 9.0).

And that's before taking into account that Faulk played with Warner, Bruce, Holt and the Martz system.

Payton's teams averaged 40.63% pass to total plays, going over .500 only in the 1987 strike season (and over 44% only in the 82 strike season).

Faulk's teams averaged 56.77% pass to total plays, going under .500 only in 94 and 95 (and 94 was the only season that doesn't top every one of Payton's non-strike seasons), and as high as an absurd 64.93% in 02.

So I'm going to go ahead and say that wile Faul might have been a little better than Payton catching the ball, the difference is nowhere near enough to make up for all the other facets of the game where Sweetness was superior. So, if I was buliding an all-time team to play under modern rules / trends, there are only three no-brainer, not even open to debate players: Walter Payton at RB, and Jerry Rice at WR1 amd LT at OLB.

Faulk can probably be the Sproles to Payton's Tomlinson.
- Alvaro

43
by tuluse :: Tue, 08/10/2010 - 2:34am

2nd and 13 might not be a rushing down for your average team, but it most certainly was for the Sander's Lions.

41
by socctty :: Tue, 08/10/2010 - 2:12am

Without reading through all the rest of the comments on here (through 25 or so as of now), I'd note this: being an above-average RB for a really, really long time is pretty damn good!

As a Texans fan that fulfills my allowance of nice things I can say about Cowboys.

42
by socctty :: Tue, 08/10/2010 - 2:18am

Another fault in this sort of approach is that you should probably do this on an individual year basis. That would make things mighty more complex and have a lot more noise, but you probably already came across the required data doing this article.

44
by DeltaWhiskey :: Tue, 08/10/2010 - 2:41am

One of the most enjoyable discussions I've read here in a long time.

In my opinion, I'd rather face Barry Sanders than Emmitt Smith. With defensive discipline and a little luck breaking your way, he could be contained.

45
by Mike Elseroad (not verified) :: Tue, 08/10/2010 - 3:35am

I agree 100% with Delta.

For instance, Emmitt gained substantial yardage in playoff games against, for the most part, high caliber opposition.

It seems, to me anyway, that when Barry faced a tough run defense, he could be more easily contained than Emmitt. For example, the 1994 Green Bay Packers defense led by Reggie White and LeRoy Butler with Fritz Schumer as D coordinator held Sanders to less than 10 yards rushing in the Wild Card round. Meanwhile, Emmitt tuned up the very same Packers defense for 100+ plus rushing yards in three straight playoff games from 1993 through 1995.

Sanders was comparable in many ways to Michael Vick. In the 2004 NFC Divisional Playoff game versus a pretty awful Rams defense, Vick as well as Warrick Dunn ran wild. But in the NFC Championship game the following week versus the Eagles, Eagles D coordinator Jim Johnson swapped defensive ends Jevon Kearse and Darren Howard (I think) so as to put Kearse on Vick's blind side (i.e. his right). Johnson also instructed his ends to corall Vick insead of rushing towards him, preventing Vick from breaking containment and tearing up the field as he did the week before versus the Rams. Johnson's plan worked perfectly as Vick never broke containment and the Eagles won easily.

I might be over simplifying things, but for the most part, Barry did a lot of his best work against really weak defenses while Emmitt's more methodical style better enabled him to perform more consistently against his foes, strong and week alike.

52
by Hurt Bones :: Tue, 08/10/2010 - 9:06am

I was at Barry's last game in Baltimore. He rushed 19 times for 41 yards. One run was for 31 yards, the remaining 18 runs netted 10 yards. Ouch.

55
by flypay (not verified) :: Tue, 08/10/2010 - 10:28am

I was at that game too! One of very few Lions fans (or any fans for that matter- was not even sold out as the Ravens were still pretty bad). Had no idea it would be the last time he would be seen in uniform...

77
by Hurt Bones :: Tue, 08/10/2010 - 1:45pm

The game was a sellout. That not everyone chose to attend is a different story.

56
by Tim Wilson :: Tue, 08/10/2010 - 10:33am

Career numbers back up your theory. Emmitt was at his best in the playoffs, while Barry came up short in the playoffs versus his regular season performances:

http://sports.espn.go.com/nfl/columns/story?columnist=paolantonio_sal&id...

As has been mentioned, Barry also lost more yardage than any other RB in NFL history. His running style was sensational, but if I was going to build an offense around a running back, I'd rather have Emmitt. He was more likely to gain yards on every carry, and he was an EXCELLENT blocker and receiver, which Barry was not.

Sure, Barry was a phenomenal pure runner, maybe the best in NFL history. But for me, Emmitt is the better all-around back and the better player to build a team around.

61
by sn0mm1s :: Tue, 08/10/2010 - 11:01am

Sure, but you can only run those sorts of D's if you don't respect the other players on the team. Also, Sanders ran for 160 yards the previous year against Reggie White and Leroy Butler in a playoff game and still lost. Emmitt was a concern - but so was Aikman and Irvin.

66
by Temo :: Tue, 08/10/2010 - 11:23am

Emmitt was a concern - but so was Aikman and Irvin.

This is false, at least to imply that they were equal threats. Irvin was awesome, no doubt, but no one geared their defense to stop Aikman. Aikman was either throwing to Irvin, or he was tossing the ball underneath to a RB/TE (this is especially true after they lost Alvin Harper).

Those Cowboys teams had a very simple concept of winning football: get the lead, rely on your defense, and grind it out on the ground. If you couldn't stop the running game from racking up first downs, you weren't going to stop them from winning.

Check out Troy Aikman's FO stats from '93 onwards: always near the top of the league in DVOA, but never in DYAR/Effective Yards. The running game had both, however.

70
by sn0mm1s :: Tue, 08/10/2010 - 11:34am

No, what that means is that when it was 3rd and long Aikman and Irvin regularly connected to move the chains. A RB has never just carried a team regular season to championship. Teams couldn't ignore Aikman and Irvin in the same was you could ignore the QBs and WRs for the Lions (other than 1995). The Cowboys were a run first team yet their passing numbers were often better than the Lions' who were running a Run n Shoot offense. Despite what everyone seems to think Smith got stuffed as well. It wasn't like he churned out 4.2 yard carries every single carry. My guess that over the course of their careers Sanders probably averaged getting stuffed (as defined by FO) only 1 more time per game than Smith.

71
by Temo :: Tue, 08/10/2010 - 11:39am

You're exaggerating my claim. I said it's false to imply that they were equal threats, because they weren't. The running game always came first. Aikman and Irvin (well, more Irvin than Aikman, but whatever) were still really awesome.

I also do not ever compare those Cowboys teams to the Lions. The Cowboys teams were some of the best teams to ever play the game. The Lions were just good some years and average in others (though, no, they were not horrible as some claim).

78
by Tim Wilson :: Tue, 08/10/2010 - 2:32pm

Read the link. The idea that the Lions did not have any other competent players on offense for defenses to focus on is a myth. Not saying they had a HoF WR, but they had other weapons.

The point, though, is not that defenses were able to focus on Barry, because theoretically they were able to do that in the regular season as well. The point is that Barry had a much tougher time in the postseason, against higher quality opponents, while Emmitt excelled in the postseason.

To me, this is a reflection of their running styles, and it's why I'd pick Emmitt to build a team around, no question.

81
by sn0mm1s :: Tue, 08/10/2010 - 3:12pm

I have read the link and Sal's article is one of the worst ever written. It is like most Barry articles (both pro and con) - not based on fact but based on hearsay. How many drives could Barry really be stopping if he averaged 20 carries a game and 17 of them were for 6.3 YPC? He averaged 3 negative runs per game for -2.5 YPC. How many championships do you really think he could win with inconsistent QBs who could rarely win starting jobs outside of Detroit and a D that was generally bad. A RB *cannot* carry a team on his own if his QBs are poor (one good QB season doesn't equal good QBs) unless he has a spectacular D.

86
by Temo :: Tue, 08/10/2010 - 4:40pm

How many drives could Barry really be stopping if he averaged 20 carries a game and 17 of them were for 6.3 YPC?

Here's Emmitt's DVOA rank vs. Barry's DVOA rank:

1993: 2/20
1994: 1/2
1995: 2/11
1996: 22/2
1997: 19/3
1998: 5/28
1999: Emmitt was 9th, Barry retired

Sanders had 2 seasons out of the 6 in the DVOA era where he had the better DVOA. I'd say those negative runs really add up.

90
by sn0mm1s :: Tue, 08/10/2010 - 5:30pm

And I have repeatedly stated in this thread that DVOA and DYAR devalue big plays. Two 10 yard runs are worth more in FO's system than an 80 yard run.

1st and 10 on your 20 what would you rather have?

A 79 yard run which puts you on your opponent's 1?

Or 2 consecutive 10 yard runs?

According to FO, the two consecutive 10 yards run is better in terms of success points - which makes no sense. It leads to issues like this past year where FO says a replacement level RB would rush for 1660 yards if he took the same carries Chris Johnson took. Do you believe that? Chris Johnson would be a better player under FO's system if he just kneeled down every 10 yards to try to pick up another 10 yards.

FO isn't even internally consistent. They state that long runs are almost exclusively based on the talent of the RB and that negative rushes are more the fault of the Oline. However, in determining their success points they penalize long runs with diminishing returns but don't lessen the impact of negative runs. Terrell Davis led the league in DYAR but carried for less than his ALY. The stats don't even pass an eyeball test. The ALY indicates that Davis was underperforming yet he was a monster DYAR compiler.

92
by Eddo :: Tue, 08/10/2010 - 5:52pm

"Two 10 yard runs are worth more in FO's system than an 80 yard run."

Did FO ever actually say this? The relevant point I've ever heard them make is that an 80-yard run is not eight times more valuable than a 10-yard run.

------

There are also a ton of regular commenters who don't trust ALY very much at all, myself included. Though, to be fair, it's not meant to correlate with DYAR, necessarily.

By the way, as FO writers often point out, "Terrell Davis's DYAR" is really "The DYAR of Terrell Davis, running behind the Broncos' offensive line and in the context of their offensive system".

95
by sn0mm1s :: Tue, 08/10/2010 - 6:20pm

From their DVOA description:

We then expand upon that basic idea with a more complicated system of "success points." A successful play is worth one point, an unsuccessful play zero points. Extra points are awarded for big plays, gradually increasing to three points for 10 yards, four points for 20 yards, and five points for 40 yards or more. There are fractional points in between. (For example, eight yards on third-and-10 is worth 0.63 "success points.") Losing four yards is -1 point, while losing 12 yards is -1.8 points. Interceptions average -6 points, with an adjustment for the length of the pass and the location of the interception (since an interception tipped at the line is more likely to produce a long return than an interception on a 40-yard pass). A fumble is worth anywhere from -1.70 to -3.98 points depending on how often a fumble in that situation is lost to the defense -- no matter who actually recovers the fumble. Red zone plays are worth 25 percent more for teams (and 10 percent more for players), and there is a bonus given for a touchdown.

So two 10 yard runs from your own 20 are worth 6 pts while a 79 yard run is worth 5 pts. The underlying formula is undoubtedly more complex but the point still remains.

101
by Jeff Fogle :: Tue, 08/10/2010 - 8:30pm

So, applying this to a previous discussion...it's possible for a team (like the San Diego Chargers) to create the illusion of "completely outplaying" someone (like the New England Patriots) in a playoff game if they start off five different drives in their own territory with double digit runs by LT against a soft defense that didn't clamp down until the Chargers reached midfield.

If you focus on how the drives ended, the game seems pretty close. If you reward "big play bonus points" to first down plays against a soft defense that eventually clamps down later in the drive...you might help create a DVOA blowout that may not be an accurate reflection of the game as a whole. Is that a fair assessment?

109
by sn0mm1s :: Tue, 08/10/2010 - 9:35pm

Sort of. The sweet spot in FO's scoring is 10 yard plays resulting in 1st downs. So, lets say LT runs for 4 first downs of exactly 10 yards (netting 12 points). Rivers throws 3 incomplete passes. The Chargers punt and NE downs the ball in the endzone to start at their own 20. Corey Dillon then busts an 80 yard run for a TD (this would be a maximum of 10 points since TDs are worth less than 6 pts in the FO world). LT would be the more successful back. If this happened 4 times in a game LT would have 16 carries for 160 yards and 48 success points just off yardage. Corey Dillon would have 4 carries for 320 yards and 4 TDs and have a maximum of 40 pts but only 20 points off of yardage. To make this even more silly let's say Corey Dillon ran for only 79 yards and Tom Brady punched it in from the 1 with a QB sneak
Tomlinson still has 48 pts.
Dillon now has 4 carries for 316 yards and 20 pts.
Tom Brady more than likely has more success points than Dillon because he gets a 10% bonus in the red zone + points for the TD + 1 pt for a successful run.

Team DVOA is a little different and Rivers' 3 incomplete passes may bring it down but you can see my point.

112
by tuluse :: Tue, 08/10/2010 - 9:57pm

I'm almost positive that's not how it works.

For one, I know the final yard that results in a TD is worth 3 yards at midfield, so there is no way half of Dillon's success points could be from the TDs.

114
by sn0mm1s :: Tue, 08/10/2010 - 10:13pm

That is why I said "maximum". FO defined a TD as less points than an actual TD so, assuming it isn't a decimal, that places the highest amount for a TD at 5 pts. Of course, if a TD isn't worth 5pts it makes that scenario even more lopsided towards Tomlinson.

116
by Temo :: Tue, 08/10/2010 - 11:38pm

So two 10 yard runs from your own 20 are worth 6 pts while a 79 yard run is worth 5 pts. The underlying formula is undoubtedly more complex but the point still remains.

I think you have a fundamental misunderstanding of DVOA. All it does is rate a play, and compare how the league rates on that play, and spit out a DVOA rating.

So no, 2 10 yard runs aren't worth more than a 79 yard run. They average 3 points a play, which is then compared to league average. The 79 yard run will always be worth more than 2 10 yard runs per DVOA.

Simply put, a player with a 79 yard run WILL ALWAYS rate higher than a player with two 10 yard runs.

119
by sn0mm1s :: Wed, 08/11/2010 - 3:31am

I was thinking DYAR not DVOA - my mistake. However, the flaw exists with DVOA just as with DYAR I just have to change the scenario. Two different drives Player A rushes for 10 yards from the 20 yard line each drive. Player B rushes for 79 yards on the first drive and -1 yards on the next drive.

Stat line of 2 carries for 20 yards player A
Stat line of 2 carries for 78 yards player B

I am pretty sure in this case that player A has the higher DVOA.

120
by Jerry :: Wed, 08/11/2010 - 5:12am

Let's extend it a bit. Player A has 8 10-yard carries. Player B has an 80-yard TD carry and 7 rushes for no gain. (We'll assume that field position, down, and distance are the same for both players.) Both players now have the same stat line, but A will (deservedly) have the better DVOA.

You're probably right about how some single-play DVOA results look odd. It's not designed to be a single-play tool, though, and as it's aggregated over a game, or a season, it looks reasonable.

124
by Temo :: Wed, 08/11/2010 - 12:13pm

What he said; DVOA is always going to look silly in tiny batches of plays like that. But I was using season-long (300+ carries) sample sizes.

The reason DVOA has the these kinds of scoring rules in place is to make it a better indicator of football success. If an 80 yard run was made to be the same as 8 10 yard runs, then it makes less of an advanced statistic and really just more of the same.

So it's strange to suggest that DVOA undervalues big runs, and that somehow YPC does it "right".

Besides, even for Barry Sanders, how many long runs of that kind of yardage did he really have? Maybe like 25 runs of over 50 yards over his career? I don't know where to check that (though PFR has him with 16 TD runs of over 50 yards). A more typical line for Sanders was like: 20 yards, -1, 0, 3, 40, 0, 2, -3, 10, 0, 0 and his YPC would be 6.5.

125
by Travis :: Wed, 08/11/2010 - 12:34pm

According to NFL.com, Barry Sanders had 42 40+ yard runs from 1991-98 (long runs for 1989 and 1990 aren't on the site, but his long carry in 1989 was only 34 yards).

Emmitt Smith had 17 40+ yard runs from 1991-2004 (1990 not included).

129
by Temo :: Wed, 08/11/2010 - 7:04pm

A bit more than I thought (I said 25 50+ yard runs, would have said 30-35 40+ yards). But still, relatively small in relevance to the number of carries a RB usually has. And yet has a disproportionate effect on YPC.

126
by Aaron Schatz :: Wed, 08/11/2010 - 12:41pm

Not quite. First of all, runs are not capped at 5 points. They simply go up slower after that. A 79-yard run is worth 5.7 points. Of course, most 79-yard runs are going to score touchdowns. Frank Gore ran for a 79-yard touchdown against Seattle in Week 2 last year. That was worth 7.73 "success points."

By comparison, a 10-yard run on first-and-10 is worth 2.15 "success points." So the 79-yard touchdown is worth a lot more than the two 10-yarders. In fact, even if one of those 10-yarders is ALSO a touchdown, the 79-yard run is worth more, as a 10-yard touchdown run on first-and-goal from the 10 is worth 4.15 "success points."

132
by DeltaWhiskey :: Thu, 08/12/2010 - 8:57am

(a) 79 yd TD = 7.73

(b) (7 consecutive 10 yd runs @ 2.15 points) + (~10yd TD run)
(7*2.15) + 4.15
15.05 + 4.15 = 19.2

With 8 min to go in the fourth quarter and a one score lead, (b) seems about right to me as far as value, that is 2.5 times greater than (a).

135
by sn0mm1s :: Thu, 08/12/2010 - 8:37pm

Really? What about in the first quarter? What about in the 4th quarter when you are down by a score? Putting the runs in an ideal game context means absolutely nothing. If you are up 1 score in the 4th quarter would you rather have the 79 yard TD run? Or 4 10 yard runs that result in a punt? The 4 10 yard runs that result in no points is worth more "success points" than a 79 yard run resulting in a TD that puts you up by two scores. 4 10 yard runs that result in you not scoring and losing the game are worth more than a 79 yard TD run that puts you ahead.

136
by tuluse :: Thu, 08/12/2010 - 8:44pm

Well this comes down to the fact that DVOA is meant to predictive. 7 successful predicts that you'll have more successful runs in the future. 1 successful run, doesn't really tell you anything, no matter how successful it was.

The whole point of DVOA is to try to predict which teams will win more games in the future.

138
by sn0mm1s :: Thu, 08/12/2010 - 9:04pm

DYAR isn't predictive. As far as DVOA, can you give a source of it being a predictive stat and not a descriptive stat? It seems to me DYAR = descriptive sum stat, DVOA = descriptive rate stat.

139
by tuluse :: Thu, 08/12/2010 - 11:43pm

They're both predictive, as they're both based on the same system. DYAR just has a lower baseline and is cumulative.

As for a source, you can look up any article where they compare year+1 performance, or any article where they make a change to DVOA, and use the fact that it correlates better to future performance as a reason for why it should or should not be included.

141
by sn0mm1s :: Fri, 08/13/2010 - 2:47am

That doesn't mean they are predictive stats - especially when dealing with individual players.

142
by DeltaWhiskey :: Fri, 08/13/2010 - 3:46am

Taking my arbitrary time constraint and various scenarios, this is what it looks like using Brian Burke’s Win Probability (WP) Calculator.

Assuming plays take on average 0:45 each, beginning with the ball on Team A’s own 20

Base Scenario
Team A up by one score (+4) with 8:00 to go in 4th , ball on the 20 to start the drive.
Team A WP: 0.76

Scenario A (Boom Barry Sanders)
Team A rips off 80 yd TD run to go up 11, consumes 0:45 of clock plus 0:15 on ensuing kickoff, which goes into endzone for touch back, Team B assumes possession with 7:00 to go on their own 20.
Team A WP: 0.94

Scenario B (Grind it out Emmit Smith)
Team A grinds out four consecutive 10 yard runs moving the ball to their opponent’s 40yd chewing up 3:00 minutes, and an additional 2:15 on 0yd runs on 1st, 2nd, and 3rd down, before consuming another 0:45 on the punt play. Team B takes over with 2:00 to go on their own 20.
Team A WP: 0.78

Scenario C (Bust Barry Sanders)
Team A grinds out three consecutive 0 yard runs consuming 2:15 on 1st, 2nd, and 3rd down, before consuming another 0:45 on the punt play, which moves the ball to their opponent’s 40. Team B takes over with 5:00 to go.
Team A WP: 0.69

However, the other part of the equation has to do with the Emmit v. Barry discussion.

Let’s assume that Scenario B has a high probability (~100%) due to Emmit’s rep (obviously the 4 consecutive 10 yd runs is optimistic, but certainly is a simplified representation of grinding it out and is consistent with his late game rep as described elsewhere in this discussion). But what about Barry, given the scenario, what is the likelihood that you’ll get Scenario A out of Barry versus Scenario C? If Barry is only likely to generate Scenario A 1 in 4 times and the other three you get Scenario C (arbitrary pick) then Team A’s WP with Barry is WP = [(3*0.69)+.94]/4 = 0.75. Even if Barry is a 50/50 proposition to boom/bust the WP = 0.81.

Based on Aaron’s above description, Barry’s Boom play is worth ~7.73 success points and presumably Barry’s bust would be worth 0. So, success points given above 1 in 4 probability of Boom = [(3*7.73)+.0]/4 =5.79. These are pretty close, and based on the above analysis, I stand by my original assertion.

143
by sn0mm1s :: Fri, 08/13/2010 - 4:14pm

We are discussing 2 different things here. FO states those 4 consecutive 10 yard runs is more successful than an 80 yard TD run. Obviously, you believe that Win Share tool is a pretty valid measurement and it shows that the 80 yard TD run is much more successful than 4 10 yard first down runs in terms of winning - even if you allow for 3 more stuffs to bleed the clock.

You are arguing that 16 10 yard runs when time is an issue *and* your team winning is more valuable than 1 80 yard TD run. I agree.

145
by DeltaWhiskey :: Sat, 08/14/2010 - 5:15am

No, you suggested the 4 runs and a punt scenario, I initially compared the ~80 yd TD run to 7 ten yd runs w/ a TD at the end and felt the value looked right to me. Next, I looked at your scenario using a different metric, hypothesized about the likeklihood of your scenarios coming true (~100% for 4 x 10 yds and punt v. ~25% for an 80 yd TD run) and came to the conclusion that they appear to be equally valuable. I don't know where the notion of sixteen 10 yd runs comes from, but I suspect it is the same place the 4 runs and a punt came from.

51
by JimZipCode :: Tue, 08/10/2010 - 9:04am

Everyone should definitely read the PFR articles. It's shocking to see the list of RBs who benefited the most from their O-line(s). Changed my opinion entirely: before I read that stuff, I was always one of those guys who said "I could run for x yards behind that O-line."

The old Bill James comment about records is apropos for Emmitt. James wrote that records are set when a number of factors come together to make an environment conducive to setting the record. For ex, to get a career home run leader, you need a great home run hitter (Hank Aaron) playing his home games in little bandbox parks, where home runs come fairly easily. Put a different great home run hitter (Mark McGwire) in a pitchers park (Oakland Coliseum), and you don't come close to threatening the home run record. (Not trying to say that McGwire was as great as Aaron.)

To get the NFL career rushing record, you need a great RB, with outstanding durability, running behind a great O-line, playing for an offensive coordinator (Norv) who really believes in pounding the rock, with a great team that often generates leads to protect. Emmitt's greatness is the first ingredient. But as with all records, not the only ingredient.

Did Emmitt have a higher career success rate than Barry? Sanders was dropped for a loss pretty often; it seemed to me that Smith almost always had positive yardage, even though he wasn't the home-run hitter Sanders was.

@Dean in #6 -- Funny, my impression of Smith is very different. In my memory Smith was always chugging forward for an extra yard in the 4th quarter, with about half the opposing defense hanging off of him. Some of Smith's most impressive runs were those clock-killing first downs with the whole D playing the runs and the Cowboys closing out a win.

(I am most emphatically *NOT* a Cowboys fan. But it's difficult not to respect Emmitt.)

59
by Capt. Anonymous (not verified) :: Tue, 08/10/2010 - 10:45am

only a small list of guys have managed to average 100yds/game in 3 different seasons(An even smaller amount has managed to lead the league in rushing 3 times and averaged 100yds/gm while doing so). Other players have led the league in yds/game but never reached 100 yds/gm 3 times(joe perry,steve van buren,gale sayers,earl campbell).

Adrian Peterson has already led the league 2 times. We are probably watching a legend. Though, he has only topped 100 yds per game once.

jim brown 7(led league 8 times)
oj simpson 3(3)
walter payton 3(1)
eric dickerson 5(5)
barry sanders 4(4)
tomlinson 3(1)
emmitt smith 3(3)

As Kornheiser would say, "that's it! that's the list!"

63
by ChasBreit (not verified) :: Tue, 08/10/2010 - 11:19am

Also a discussion that could involved Favre. Sort of like Smith in that longevity begets numbers and sort of like Sanders in that he was boom-or-bust, often in the same game.

67
by Temo :: Tue, 08/10/2010 - 11:27am

I anticipate a great discussion on Favre whenever he decides to hang 'em up for good. Not on whether he belongs in the Hall, but whether he belongs on the short list of "greatest ever".

69
by Temo :: Tue, 08/10/2010 - 11:34am

In the period from 1991-1995, Emmitt Smith carried the ball, in order, 365,373,283,368, and 377 times. The one year with 283 he only started 13 games due to a contract hold out.

That is probably the greatest workload over a 5 year period in the history of the game. And yet he survived, thrived, and helped carry the offense for teams that won 3 super bowls during that period.

121
by Skafko (not verified) :: Wed, 08/11/2010 - 7:03am

This carries are only for regular season, you forgot carries in post season

122
by DeltaWhiskey :: Wed, 08/11/2010 - 10:05am

'...unless he is named Eric Dickerson."

Or apparently Emmitt Smith.

127
by chemical burn :: Wed, 08/11/2010 - 2:05pm

Uh, after his 373 year his missed several games due to injury and had his worst production. Just sayin'...

130
by DeltaWhiskey :: Thu, 08/12/2010 - 2:10am

Thought Temo said he started 283 due to contract holdout, guess somebody, me, you or Temo needs to fact check.

131
by Andrew Potter :: Thu, 08/12/2010 - 7:33am

The holdout has it:

Smith's first contract expired in 1993, and as the new season neared he clashed with the Cowboys' owner about how much he was worth. Once again he missed training camp, and still he did not have a satisfactory contract. The season began without him. Dallas played its first two games--and lost them both-- while Smith sat at home in Pensacola watching them on television.

Not sure where the injury bit came from, though he was injured toward the end of the season:

In the first week of 1994 the Cowboys met the Giants in a divisional showdown to determine NFC East first place (carrying an automatic home field advantage throughout the playoffs). This game, perhaps more than any other, highlighted the singular talents that have made Smith famous. Despite a serious separation of the shoulder sustained early in the second half, he was eager to play.

( http://www.answers.com/topic/emmitt-smith )

His production during 1992 (the 373 carry season) was just over 24 carries per game, averaging 107 yards per game (4.6 yards per carry).

His production during 1993 (the 283 carry season) was just under 22 carries per game, averaging 114 yards per game (5.25 yards per carry).

( Stats for the above from www.pro-football-reference.com/players/S/SmitEm00.htm )

133
by James (not verified) :: Thu, 08/12/2010 - 9:59am

Smith had only one carry in one game due to injury that season. I think it was the Atlanta game. I'd like to point out the Cowboys lost all three games Smith didn't play in that year, and would have gone undefeated (including playoffs) with him if Leon Lett wasn't an idiot.

134
by Dean :: Thu, 08/12/2010 - 11:12am

All of the metrics in the world can prove one thing - that Emmitt was a world-class back. The thing is, though, that none of his detractors - myself included - dispute this.

While statistics can tell you how good he was or wasn't, they can't measure how overrated he was or wasn't.

Statistics don't measure the concept that for all his ability, he still never measured up to the enormous hype which surrounded him, and never measured up to the image which the punditry presented.

That doesn't mean he wasn't a hell of a back, just not as good as advertized.

It also doesn't change the fact that we're dealing with opinions and another person every bit as knowledgable could look at the exact same set of facts and draw a different conclusion. It makes for a fun discussion, but much like Manning/Brady from a few years ago, it's not a discussion which can ever reach a definitive conclusion.

144
by Tim Wilson :: Fri, 08/13/2010 - 8:26pm

What "enormous hype"? I think the point of the initial FO post here was that whenever Smith's name comes up, there are often an abundant number of detractors who call him overrated or an average back, who bring up Barry Sanders as an infinitely superior player, who call Smith lucky due to his OL, etc.

Even on the weekend of his HoF induction, when hyperbole is usually at an all-time high for a player (Jerry Rice was dubbed the greatest football player ever by the NFL Network crew, for instance), I never once heard Emmitt called one of the greatest backs of all time-- just the NFL's all time rushing leader. I don't think that the media or fans in general fall over themselves to praise Emmitt. Certainly he gets some praise, which is to be expected for an all-time great player, but to me he is underrated if anything, due to the ready criticisms which seem so plentiful whenever he is mentioned.

74
by Mansteel (not verified) :: Tue, 08/10/2010 - 1:21pm

Two words explain Tiki Barber's ranking 2nd on this metric: "Ron" and "Dayne".

75
by Dan :: Tue, 08/10/2010 - 1:33pm

Using ypc makes boom-bust backs look good, and Emmitt's last few years bring down his average, so career ypc (vs. teammates) isn't a good way to compare Barry and Emmitt.

Look at the 6 years of the DVOA era when they both played (1993-1998): Barry had a much higher ypc (5.12 vs. 4.32), but Emmitt actually had a slightly higher DVOA (8.8% vs. 7.1%), thanks a much better success rate (52% vs. 45%) as well as more TDs (84-47) and short-yardage carries. YPC strongly favors the boom-bust back, even though the actual value of their carries was very close.

How many times did Emmitt's teammates have a higher ypc than he did? I just glanced at the numbers rather than doing exact calculations, but it looks like it happened 6 times: once in 1996 and then in five straight years at the end of his career, 2000-2004. The first ten years of Emmitt's career (1990-1999) are more comparable to Barry's 10-year career in showing what he could do as an elite back; the fact that he held on for a few more years in his 30s and became more of a plodder shouldn't count against him.

76
by sn0mm1s :: Tue, 08/10/2010 - 1:40pm

His years in Arizona don't move his average much though. He was 4.2 when he went to Arizona and 4.2 when he retired.

80
by Bobman :: Tue, 08/10/2010 - 3:05pm

Am I missing something? Shouldn't Edgerrin James be on thst list somewhere? Is he not officially retired yet (like Marvin Harrison)? Oy.

Anyway, thinking of Edge and Emmitt together bring us to usage and backup RBs. When you have a guy who is a complete workhorse, there may well be a tendency to undervalue the importance of #2 (See Manning, Peyton, Indianapolis Colts). Until Edge's injury in Game 6 of his third year, he basically led the NFL in rushing for 2.4 seasons and carried 95% of his teams runs, maybe more. Who the hell needs a backup? My impression of Smith's Cal Ripkin-like career (I thought that was an apt analogy above--both great players, both insanely durable and consistent, neither the greatest ever) might lead to a situation where the fall-off to #2 was hardly an issue, or might have been statistically irrelevant for half his career if he had 350 carries and his backup had 20.... (or something like that)

Anyway, none of this takes into account his contribution to the English langiuage: "debacled." That alone is worthy of the gold jacket.

146
by pbrane (not verified) :: Wed, 09/15/2010 - 7:02pm

Just saw this while perusing historical stats... why are people wasting time trying to use FO rushing DVOA/DYAR to argue points about Sanders and Smith? Those stats don't separate out the RB from the team and context at all... they simply measure how effective the Dallas and Detroit rushing attempts were when Smith and Sanders were on the field.

And I appreciate the point made by Mr Shush about the parameters of era affecting player value, a great point that's often overlooked. There are many contextual factors that make any attempt at statistical comparison of individual players next to useless. Don't get me wrong, I'm a stat-head who loves sabrmetrics in baseball, but in football I simply don't find much of the quantitative analysis particularly enlightening *when it comes to player valuation.* I don't want to sound like I don't like anything that football stat analysis has to offer.

At the end of the day, much as I would prefer to have a valid and reliable statistical assessment of individual player value, this is one area where I think qualitative assessment and opinion are about as far as you'll get....

So, that said, for Sanders v. Smith... I'd choose Sanders in a heartbeat over Smith.