Writers of Pro Football Prospectus 2008

19 Sep 2007

Willie Parker and the Curse of 370

Suddenly, everybody seems concerned about running back workloads -- except for Mike Tomlin, who promises (threatens?) to "ride Willie until the wheels come off". The "Curse of 370" (and RB overuse in general) seems to be the first FO theory to approach the ranks of conventional wisdom. It's early to be speculating about this season's potential victims, but that didn't stop the Pittsburgh Post-Gazette's Ed Bouchette from asking Mike Tomlin about Willie Parker's workload. After two weeks, three running backs have the 47 or more carries that would put them on pace for 370 over the full season -- Travis Henry (49 carries) plus Edgerrin James and Parker (50 carries each).

Posted by: Stuart Fraser on 19 Sep 2007

84 comments, Last at 20 Sep 2007, 8:51pm by Alex

Comments

1
by Lou (not verified) :: Wed, 09/19/2007 - 9:54am

3 running backs, 3

2
by Stuart Fraser :: Wed, 09/19/2007 - 10:09am

1 - D'oh. I think at that point I was trying to write "two running backs beside Parker" or something. Fixed.

3
by Yaxley (not verified) :: Wed, 09/19/2007 - 10:23am

As a Steelers fan, I'm very worried about Parker getting the Larry Johnson treatment. But I'm also excited that the Curse of 370 is being discussed in the Post-Gazette. That means there's at least a chance that the message will get through to Tomlin at some point, right?

4
by TGT (not verified) :: Wed, 09/19/2007 - 10:43am

Good coaches ignore writers, right?

5
by Fnor (not verified) :: Wed, 09/19/2007 - 10:46am

Especially the Post-Gazette.

6
by JCRODRIGUEZ (not verified) :: Wed, 09/19/2007 - 10:59am

Is it just my perception, or must of the Fast Willie's carries have been on the second half of two loopsided games?...why not give Najeh some more load?...is it a getting-Willie-on-rithm kind a thing?...

7
by Ben (not verified) :: Wed, 09/19/2007 - 11:01am

I don't think it's reasonable to extrapolate his current # of carries over a season considering that the Steelers had a huge lead in both games. the second half playcallig will be very different in games that they are behind or when they have a small lead. He should have been pulled from the Browns game sooner than he was, but I think it's a little early to get worried about it.

8
by Starshatterer (not verified) :: Wed, 09/19/2007 - 11:08am

I don’t think it’s reasonable to extrapolate his current # of carries over a season considering that the Steelers had a huge lead in both games.
Except, as "JCRODRIGUEZ" pointed out, you would expect the backup running back (Davenport) to soak up a bunch of those late-game-with-a-lead carries.

It may be a little early to worry about it, but Tomlin's “ride Willie until the wheels come off� seems a little cavalier. Maybe he was joking?

9
by Theo, Holland (not verified) :: Wed, 09/19/2007 - 11:09am

[FO and Steelers fan mode]

Whoo hooo! The Curse of 370 goes mainsteam!!
Oh damnit, it's Parker.

[/FO and Steelers fan mode]

10
by Ian (not verified) :: Wed, 09/19/2007 - 11:11am

I agree with 7. As much as I am a Steeler homer, I doubt the team will have such huge leads in every game. Remember, although there is a new coach, the Steelers have been quick to drop the running game at times. As well, I don't think the discussion around no-huddle, spread sets was just a smokescreen.
As for the Post-Gazette, they've always been big in narrative with very little analysis. A great sports town deserves better.

11
by The Other Vlad (not verified) :: Wed, 09/19/2007 - 11:14am

Good baseball coverage at the P-G since Dejan took over, though.

12
by Mr Shush (not verified) :: Wed, 09/19/2007 - 11:16am

My question is whether running Parker into the ground is necessarily a big problem, from the team's point of view. Obviously burning out an elite runner like LJ or Alexander seriously hurts your team, but Parker isn't one. He's a below average back who puts up big conventional numbers because he gets the ball a lot and breaks some big gains. His production would not be hard to replace, and his contract doesn't contain enough guaranteed money to make cutting or trading him after this season prohibitive. He will count around $4m against next year's cap if he is on the team, or a little under $2m if he is cut or traded. If he racks up 1500-plus yards this year, as seems possible, he might well bring a third round pick or better in trade. I'm not saying this is what Tomlin plans to do, but if he did it, I'd call it a pretty smart move.

13
by fish shure (not verified) :: Wed, 09/19/2007 - 11:19am

It's great that FO is getting this publicity. What's interesting is that I think the "idea" of running back over-use has penetrated even farther than the specifics. Even though you might not see mention to the "Curse of 370" everywhere, I feel like commentators and sports writers are much more aware of running back overuse.

It helps that the public has become conditioned to know that going way over 100 pitches in baseball is a bad idea; this is sort of a translation of that into football.

14
by DrObviousSo (not verified) :: Wed, 09/19/2007 - 11:19am

#9: Exactly. Exactly.

15
by beedubyuh (not verified) :: Wed, 09/19/2007 - 11:21am

For Steeler fans concerned with Parker's usage, I'd pay more attention to what Tomlin does than what he says. Last week, didn't Davenport get all of his yardage during the final drive, when PIT was just running out the clock? That speaks volumes about Tomlin's intent to rest Parker when feasible. He just probably won't admit it, for a variety of good reasons.

16
by CA (not verified) :: Wed, 09/19/2007 - 11:31am

There is documented evidence that if a runner carries 370 times a season, including the postseason, nothing but bad things happen to him.

Actually, it's 370 times in the regular season, 390 including the postseason.

I have to say that Tomlin, whom I generally strongly respect as a coach, comes off as pretty dumb in that article. Not quite "They want it to be a thinking game instead of a football game" dumb, but dumb nonetheless. Dumb enough, in fact, that I suspect he may be bluffing.

17
by Scott de B. (not verified) :: Wed, 09/19/2007 - 11:33am

So the main hope Steelers fans have that Parker won't be over-used is that they will suck in a bunch of games this year? Talk about a Catch-22.

18
by Mike F. (not verified) :: Wed, 09/19/2007 - 11:40am

Using Curtis Martin as an example doesn't make much sense, does it?

368 carries in 1995; 369 in 1998; 367 in 1999. Sure, he was a few carries away from the "magical 370," but if anything, Curtis Martin proves this theory wrong, not right.

19
by Mikey (not verified) :: Wed, 09/19/2007 - 11:51am

How does college use factor into the 370 rule? Does it factor at all?

Parker had 285 career carries at North Carolina. To take two other examples, Larry Johnson had 460 carries at Penn St at Shaun Alexander had 727 carries at Alabama.

In seven years of college and pro football prior to this season, Parker had a total of 909 carries, or 130 per year.

Does this make him less susceptible to effects of overuse?

Hopefully it's a moot arguement. Najeh has been running hard in spot duty and I would be fine with seeing him used more often as a change-of-pace back.

20
by CA (not verified) :: Wed, 09/19/2007 - 11:55am

Re: 18 Using Curtis Martin as an example doesn’t make much sense, does it?

You are conveniently ignoring 2004, in which Martin had 371/408 carries and a league-leading DPAR of 54.9 among running backs. In 2005, he had a DPAR of -2.8, making him a classic victim of the curse. That said, the fact that Martin had 369/418 carries in 1998 and managed to improve his yards per carry and rushing total the following season is a noteworthy exception to the rule.

21
by mawbrew (not verified) :: Wed, 09/19/2007 - 12:01pm

I would guess tha Parker will get hurt at some point this year and miss at least a few games, thus leaving him short of the magic number

22
by hooper (not verified) :: Wed, 09/19/2007 - 12:03pm

Credit where credit is due - the discovery of 370 as the breakpoint of running back abuse should certainly be well-received and well-attributed. However, I don't remember how much people actually thought of overuse of running backs before FO put a number to it. In other words, how much attention was paid to running backs with large carry numbers before FO pronounced "370". Was the innovation the concept of abuse, or was it simply the number?

On another front, given that we're only 2 weeks in, I don't think there's a lot that can be read into this. In Pittsburgh, I wouldn't be surprised if Tomlin is playing safe by having Parker do the work until he gets a full handle on the rest of the team. I'm sure he's been a little cautious about Ben, given last season. Maybe he's not quite to the point of giving Davenport all the cleanup duty. (Just thoughts, no real analysis there.) I'll give about 3 more game weeks before I worry about Parker.

For another point of oddities in 2 weeks, note that Travis Henry is another back on pace for the curse. While Denver would be a team that doesn't seem to worry about running back shelf life, it could still seem odd that one back would get so many carries when they have Young, Bell, and Sapp, and Denver could probably get my grandmother to run for 800 yards.

The Pittsburgh/Denver contrast is especially interesting because Pittsburgh has been in blowouts, while Denver has been in squeakers. Parker's garbage time carries may explain his numbers, but that couldn't apply to Henry. Denver has been averaging around 70 plays a game, but that couldn't apply to a clock-killing Pittsburgh. It seems that the reasons for their large loads are opposing and mutually exclusive, yet here we are.

I'd give it a couple weeks.

23
by CA (not verified) :: Wed, 09/19/2007 - 12:11pm

Re: 19 How does college use factor into the 370 rule? Does it factor at all?

Mikey, this doesn't exactly answer your question, but, last year, Bill wrote a BSMW article last year that discussed workloads among Big Ten running backs and how that translates to NFL success. If I am remembering correctly, the magic number is 550. That is, historically, Big Ten running backs who had more than 550 career carries in college tend not to perform well in the NFL. The link to the FO Extra Point on that piece is in my name, but, unfortunately, the link within that to the actual article doesn't work. You may be able to cull some of the info from the comments. Bill, if you're reading this, do you have a working link to the article? By the way, I'm aware that Parker did not play in the Big Ten.

24
by rk (not verified) :: Wed, 09/19/2007 - 12:12pm

In 2005, Parker had 47 carries after 2 games (both blowouts like this year). He finished the year with 255. It's way too early to guess how many carries he'll get this year regardless of what Tomlin says.

25
by Stephen (not verified) :: Wed, 09/19/2007 - 12:25pm

#18: Exactly. I'm always skeptical of putting a number on something like pitches or catches. 370 for parker might be 400 for LDT or 290 for Bush. Teams should be able to evaluate their own personnel and decide if they are tiring or not. 370 seems like just a number to me, because is there really that much of a difference between 370 and 360? I think a spectrum of carries would be a better tool to measure the dangers of overuse.

26
by Stephen (not verified) :: Wed, 09/19/2007 - 12:26pm

sorry, i meant carries instead of catches

27
by Mike F. (not verified) :: Wed, 09/19/2007 - 12:40pm

You are conveniently ignoring 2004

I wasn't ignoring anything: yes, he "broke down" after that season. But he didn't after 1995, 1998, and 1999, did he? He continued to put up solid numbers. So the "curse" was right 1 out of 4 seasons.

And here's another one: Terrell Davis: 1997, 369 carries. The year after that, he rushed for 2000 yards. (Year after that, he was done). But if the "curse" is defined as such:

A running back with 370 or more carries during the regular season will usually suffer either a major injury or loss of effectiveness the following year, unless he is named Eric Dickerson

Then Curtis Martin and Terrell Davis really shouldn't be cited as examples.

28
by Independent George (not verified) :: Wed, 09/19/2007 - 1:00pm

#27 - that's why it's the curse of 370, not the curse of 369.

Snark aside, you're absolutely right - an ACL Gremlin doesn't suddenly materialize on carry no. 370, and there's no real difference between 369 carries and 370. That's not the point, though; despite the name, it's really more of an easy-to-remember convention than a hard and fast rule. The idea is that RB workloads matter, and 370 marks a place where you should start being concerned (and 416 is completely off the charts).

It's like pitch counts in baseball; there's nothing mystical about 100 pitches that causes promising young starters to blow out their elbows, but it's nevertheless something coaches should be concerned with.

29
by Mike F. (not verified) :: Wed, 09/19/2007 - 1:21pm

re: 28 -- I wasn't trying to play semantics with the actual number. It's not that I think Martin and TD and Campbell shouldn't be involved in this whole "curse" discussion because they fell a few carries short of 370 . . . it's because their performance did not significantly decline the year following their heavy workload.

In fact, TD and Campbell's production actually went up. (The year after that, they went down, yes).

30
by Phil (not verified) :: Wed, 09/19/2007 - 1:26pm

"It’s like pitch counts in baseball; there’s nothing mystical about 100 pitches..."

Unless your name is Pedro...

31
by Stuart Fraser :: Wed, 09/19/2007 - 1:59pm

From the Football Outsiders glossary:

Curse of 370: The theory that a running back with 370 or more carries during the regular season will usually suffer either a major injury or loss of effectiveness the following year, unless he is named Eric Dickerson. Occasionally, the loss of effectiveness or injury takes place two years later.

It isn't *always* the next year. Sometimes it does take more than one year - LaDanian Tomlinson (2002-2004) is another example.

32
by Alex (not verified) :: Wed, 09/19/2007 - 2:05pm

Teams should be able to evaluate their own personnel and decide if they are tiring or not.

Trouble is, teams are frequently too optimistic in their evaluations of how many carries RBs can take, and this results in RBs having ruined seasons/careers. And teams haven't learned from this, despite seeing numerous examples. It happens almost every year, yet they keep running people into the ground.

If there were evidence that NFL teams were able to accurately predict which RBs could handle 370+ carries, then maybe I'd agree. But there isn't. Herm Edwards alone has been wrong twice now. A coach looks at his RB, thinks that since he's young, strong, physical, and feel's fine, that he'll be able to handle the workload. But feeling fine isn't the issue. The effects don't always show up right away, so by the time the RB starts feeling worn out, the damage has mostly already been done.

33
by Brian (not verified) :: Wed, 09/19/2007 - 2:11pm

Has anyone ever taken a scientific look at the 370 carry rule?

For example, the "breakdown rate" of a RB with 370+ carries may be high. But the breakdown rate of all starting RBs may be nearly as high.

Another possibility is that most of the observed effect of high numbers of carries is simply due to regression to the mean. If a RB has an exceptionally good year in terms of yds/carry, he'll be fed the ball very frequently. But it's rare that any player at any position has two elite/exceptional years in a row. As the RB's performance comes back down to earth (regresses to his mean or below), it will appear as if the previous year's workload caused the dropoff.

Just a thought. I'd like to hear what other people think.

34
by Bjorn (not verified) :: Wed, 09/19/2007 - 2:21pm

Travis Henry is just about 30 years old anyways. Might as well rush him as much as you can.

35
by coldbikemessenger (not verified) :: Wed, 09/19/2007 - 2:22pm

Landry pretty much refused to give Dorsett more than 20 touches a game.
He said it year after year after year.
And Dorsett whined about it year after year...

To be fair
Dorsett was something like 185 pounds
Even 25 years ago nobody thought of him as a bruiser
But he was pretty awesome between the tackles

36
by Costa (not verified) :: Wed, 09/19/2007 - 2:34pm

33: I'm sure there's a fair bit of truth to that in terms of just explaining drop in production part of the equation, but then there's still the frequency that serious injuries occur that remains.

I'm sure with time that more exceptions will appear, but there's still enough evidence to show that there's something very real there. Not necessarily with the specific number "370" per se, but with general overuse of running backs in general. Average RB career length remains the shortest of all positions, and it seems as though a large part of that is due to the fact that coaches need to learn that the position is different from others in that they need more rest, similarly to how the MLB needed to learn that the usage pattern of pitchers needed to change from the old days when they used to pitch 400+ innings a season and no one counted their pitches.

37
by Stephen (not verified) :: Wed, 09/19/2007 - 2:50pm

#32: I agree, but teams "should" also know not to hire Herm Edwards as their coach, in my opinion.

38
by Jerry (not verified) :: Wed, 09/19/2007 - 2:51pm

I heard the question and answer at Tomlin's press conference yesterday (with no mention of the number 370) and thought of FO. I was happy to see that today's article referenced FO so much.

It's odd that in Cowher's first year, he ran Barry Foster into the ground. Maybe the rule of 370 is something that a coach has to learn about for himself, unless he is named Herm Edwards. As a Steeler fan who doesn't think Parker is "a below average back", I hope Tomlin is made aware of what he's risking; this article will probably help.

39
by Bobman (not verified) :: Wed, 09/19/2007 - 3:20pm

Does running style play into this at all? (Sorry to add more noise to the calcs.) I get the impression that Willie is an around-the-end speedster and not a between the tackles bruiser who is likely to get a LOT more dinged up. (of course who pulls a hammy or torques a knee more often? Probably the guy sprinting outside.)

Nevertheless, how they finish runs might also factor in--Edge (and Walter Payton if memory serves) almost never took a direct hit. He'd angle off a bit at the end so the impact was more glancing, or he'd be bounced out of bounds. Edge had a ton of stretch run plays that gave him the sideline to work rather than go head to head with Ray Lewis up the gut every Sunday. My impression is that that helps on the old wear and tear (probably also reduces career fumble rate as well).

This is coming from a guy counting Joseph Addai's carries nervously already.

40
by David (not verified) :: Wed, 09/19/2007 - 3:20pm

To be fair, Parker's not below average. He's a league-average rush muncher.

Last year, his DPAR put him 14th among league running backs and he's 16th in PAR right now. He was 24th in DVOA last year and is 21st now.

Over 350 rushes, a league-average rush muncher can have value, if I recall my FO books correctly.

41
by brian (not verified) :: Wed, 09/19/2007 - 4:10pm

The more I think about this theory, the more I think this theory simply doesn't hold water. I am not saying that excessive carries are a good thing, or that they don’t hurt a back in the long term, but clearly not in the way that the theory claims. There are simply too many exceptions. It almost reminds me of the Madden Curse, where once you accept the curse exists, it is easy to manipulate the data to support the curse. But, an objective review shows it to be nothing more then just a manipulation of the data and ignoring any data that doesn’t support it.

Some one else has already mentioned Curtis Martin; we are supposed to ignore the first three times he hit 370 (or about) because they don’t support the “curse�, and focus only on the last time because he broke down after that. What about Emmitt Smith? He did it four times, with no discernable performance decrease the first three times, and only a small decrease the fourth time. What about Walter Peyton? He did it twice, and actually improved the following year both times. More recent examples include Tomlinson, who did just fine after 370+ carries. Jamal Lewis, whose DVOA went up the following year, etc.

This brings us to the second point. Why do post season carries not count? I can see why you don’t include catches, but what makes a post season carry any different then a regular season one? How is a back who carries four yards up the middle before getting gang tackled in week 17 affected any more then a back who carries four yards up the middle and gets gang tackled the following week in the Wild card round of the playoffs? The only possible reason I can see for not including post season carries is because once you do include them, the number of “exceptions� to the “curse� increases far too much for the curse to even be considered anymore.

The third thing the “curse� fails to consider is that for many backs, their 370+ carries year is a career year. The “drop-off� the next year isn’t so much a drop-off as a regression to the mean. I am actually surprised that this site of all places would ignore this effect, as it is considered in so many other things (high or low fumble recovery rate, high or low third down conversion rate, etc.). As an example of this, lets look at Ricky Williams. In his first year with the Dolphins, he had a career year and carried for 370+ times. DVOA of 6.5. The following year his DVOA was down to -12.1. Proof of the theory right? Not really, it is simple regression to the mean. His two seasons prior to carrying 370+ he had a DVOA of -11.6 and -14.1.

42
by brian (not verified) :: Wed, 09/19/2007 - 4:13pm

Re 31: If that is the case though, then the theory is pretty useless isn't it? I think we can all agree that a back who carries 370+ times will eventually either suffer a serious injury or loss of effectiveness at some future point in their career. Of course, ANY starting back, regardless of number of carries will eventually suffer either a serious injury or loss of effectiveness at some future point in their career.

43
by senser81 (not verified) :: Wed, 09/19/2007 - 4:13pm

re: #22

I thought at the end of his career Earl Campbell was used as an example of RB abuse, and before him Larry Brown was used as an example of RB abuse. I don't think RB abuse/overuse was something invented by FO. Of course, I don't believe Al Gore invented the internet, either....

44
by CA (not verified) :: Wed, 09/19/2007 - 5:03pm

It looks like it may be time to re-post the "370 Carries Revisited" article. The critics of the concept of the "curse," particularly brian, should read the PFP 2006 piece linked in my name, as well as the original "Ricky Williams is Pretty Much Screwed" piece. They address most of your points.

45
by Tom (not verified) :: Wed, 09/19/2007 - 5:24pm

Re 39:

Except Payton almost never went out of bounds, and sought out contact. Although he did know how to turn hits into glancing blows.

46
by joe football (not verified) :: Wed, 09/19/2007 - 5:25pm

I agree with 12. Parker is a former undrafted FA who has a modest contract and, according to FO, isn't that good anyhow. Not to sound cold, but I'm willing to break him if it helps us this year

47
by Sundown (not verified) :: Wed, 09/19/2007 - 5:50pm

When it comes to Travis Henry, does the "Curse of 370" refer to carries or number of illegitimate children?

48
by brian (not verified) :: Wed, 09/19/2007 - 5:53pm

Re 44: I have read the original post, several times in fact. However, contrary to what you claim, it does not address the points I raised.

My first point, it ignores all the backs that had great (or at least good) years after carrying the ball for 370 times. There is no explanation in the article as to how these exceptions could exist if the theory is correct.
My second point, regarding post season carries. It tells us that 390 carries (including post season) is equal to 370 carries (regular season). But it doesn't tell us WHY post season carries count for less. There is no obvious reason why they should count for less. Now of course it is possible there is a non-obvious reason why they should count for less, and if the sample size was in the thousands I might agree, but it isn't. The other possible explanation is that there is no curse, and that the number of carries chosen is arbitrary and just happens to coincide with a small data set.
And then the third point it doesn't address either. The article points out that after carrying for 370+ times, Williams' DVOA dropped to double digit negatives. This is cited as proof of the curse. But what it ignores is that he had double digit negative DVOA in every year PRIOR to carrying 370+ times too! It wasn't the 370+ carries that did it, he was always a negative DVOA rusher who happened to have one great year.

This is exactly why I used the analogy of the Madden Curse. While many players DO in fact regress after appearing on the cover of Madden, it isn't BECAUSE they appeared on the cover.

49
by Brian (not verified) :: Wed, 09/19/2007 - 5:56pm

#44: Thanks, I was looking for that.

Also, I don't think we should discount "lower-case brian's" point responding to #31. If we expand the curse window to 2 years, you're bound to see frequent breakdowns.

It's not enough to say that high-carry RBs 'breakdown' at x% rate the following year. You have to compare that with the general 'breakdown' rate of all starting RBs at y%. if x minus y is statistically significant than there may be weight to the theory.

That said, I'll read the article now, which I trust does just that!

50
by JeffW (not verified) :: Wed, 09/19/2007 - 6:12pm

#20: C. Martin also had seasons of 368 and 367 carries to go along with the 369 carry season.

The 370 carry break point is misrepresented as a hard break point IMO. The sample size is small enough that it may just be a coincidence that 370 is the break point--this blind spot comes up a bit too often in FO research. I'd rather see a model of RB breakdown as a function of carries.

51
by Waverly (not verified) :: Wed, 09/19/2007 - 6:19pm

I wonder if the general public is attracted to this theory in part because it's called a "curse", involving a "magic" number.

Not everyone will understand the subtleties, either, as is typical for statistical observations. As it gets better known it will be simplified. Some people will apply it in absolute terms.

Finally it will gradually become obsolete as a better theory comes along, even as it achieves greater popularity amongst general football fans.

52
by Alex (not verified) :: Wed, 09/19/2007 - 6:26pm

For example, the “breakdown rate� of a RB with 370+ carries may be high. But the breakdown rate of all starting RBs may be nearly as high.

It isn't. From FO Basics:

"Plenty of running backs get injured without hitting 370 carries in a season, but there is a clear difference. On average, running backs with 300 to 369 carries and no postseason appearance will see their total rushing yardage decline by 15 percent the following year and their yards per carry decline by two percent. The average running back with 370 or more regular-season carries, or 390 including the postseason, will see their rushing yardage decline by 35 percent, and their yards per carry decline by eight percent."

Another possibility is that most of the observed effect of high numbers of carries is simply due to regression to the mean.

Except they don't regress to the mean, they regress to below the mean in most cases. If the extra carries weren't causing the regression, then they would only regress to their average performance level the next year. But on average, they regress below their normal level of performance.

53
by zlionsfan (not verified) :: Wed, 09/19/2007 - 6:36pm

I think the entire concept is being misrepresented by some people.

To me, the articles say that one reason RBs are likely to decline after a heavy workload is because of the workload itself. They don't say that it will happen, or that it only happens at 370 and above, or that there are no other possible reasons for this. They just say that there is a negative correlation in the group between carries in season X and yards in season X+1.

Aaron actually says this in the foreword to the more recent article:

"Just so people understand, there’s nothing magical about carry number 370 that makes a running back blow out his ACL, any more than there is something special about pitch 100 that makes a pitcher’s arm fall off. It’s simply a useful shorthand to represent the fact that overworking your running back with too many carries is a bad thing."

There is no "curse", they don't present "proof", it isn't a hard-and-fast rule. It's simply a danger zone. If you want to make a bunch of noise about how RBs with a heavy workload aren't guaranteed to decline the following season, well, I'm sure FO would agree with you. If you're arguing that you "proved" them "wrong", well, whatever.

54
by bradluen (not verified) :: Wed, 09/19/2007 - 6:48pm

370 vs 390 inc. postseason seems consistent to me, for the same reason that (numbers plucked from thin air) 30 carries in one week might be equally as wearing as 40 carries in 2 weeks. The equal-wear relationship between weeks and carries should be neither flat nor proportional.

55
by Brian (not verified) :: Wed, 09/19/2007 - 7:43pm

re 52:
I think your first point is a good one. But the difference between a 2% decline and an 8% decline is 0.24 yds/rush for a starting NFL RB. Assuming that very-high-carry seasons are often peak or "career" years for RBs, a 0.24 yd/rush decline seems very mild. That's about 80 yds spread out over a 16-game season, or about 5 yds/game.

On your second point, I think there are far too few years in even the longest career to really know a RB's "true" mean performance. I'm not sure how one would know if he regressed below his mean without knowing the mean. We can only know he declined in the direction of his mean, and we could never expect an RB to regress to exactly his mean.

Also keep in mind the tendency is always towards decline as players age.

Despite my skepticism, the workload theory seems plausible intuitively.

56
by tunesmith (not verified) :: Wed, 09/19/2007 - 8:10pm

Re: 43 - neither does al gore. :)

I wouldn't worry about Travis Henry hitting 370, he'll get injured long before that number is reached. He's already been gimpy twice this season.

57
by Dunbar (not verified) :: Wed, 09/19/2007 - 11:53pm

I don't understand why the 370 rule should apply two seasons down the line. You'd think that would be enough time for the RB to get over the effects of the 370-carry season.

I'm not saying that the 370 rule isn't a nice guideline for RB overuse. But I'm thinking that the reason for the correlation isn't so much the cumulative effects of having that many carries over the course of the season (although obviously there must be some effect, maybe a fairly significant one); I think the reason you see running backs break down a year or two after the 370-carry season is just probability--if there's, say, a 1 in 500 chance that a back will sustain a major injury on any given carry (just pulling that number out of my ass there), you're more likely to see it happen if you actually run the guy that many times over ANY period of time, as opposed to running him like 200 times over the same time. Add whatever cumulative effects there are and yeah, sure, you're going to see some injuries.

I hope that doesn't seem condescending--it's not like I think I'm the first person to think of that--but it seems like a lot of people do use the 370-carry mark as some sort of magical milestone. I think that if you run a guy 25 times a game, he's probably not that much more likely to get hurt in ANY GIVEN GAME a year or two after a 370-carry season than he is during Week 8 of a potential 370-carry season. So there are probably a fair number of seasons where a RB was on pace to get 370 carries or thereabouts, and got injured before he could ever reach that point.

Anyway, that's probably a totally pointless post, but that's kind of been bothering me for a while.

58
by Alex (not verified) :: Wed, 09/19/2007 - 11:58pm

On your second point, I think there are far too few years in even the longest career to really know a RB’s “true� mean performance. I’m not sure how one would know if he regressed below his mean without knowing the mean.

But we're not just looking at one RB's career. We're looking at dozens of RBs. If you take the average "non-Cursed" year of all those RBs, and then compare that to the years directly following the 370+ carry season, and find that the average "Cursed" year is significantly worse than the average "non-Cursed" year, then you have good evidence that it's not just regression to the mean, but regression below the mean. They can't all be overachieving for the majority of their careers, and then falling back down to their "true" levels after 370+ carry seasons.

Also keep in mind the tendency is always towards decline as players age.

Not true. Early in their careers, RBs tend to improve with age. Eventually, age-related decline sets in.
But the Curse of 370 has struck several RBs when they're still quite young, well before they would be expected to start declining due to age. It hit Edge when he was 23, Jamal Lewis when he was 25, and Terrell Davis and Jamal Anderson when they were 27.

Edge went on to have a very good career after recovering from his "Cursed" year. If the decline were due to age, he would've kept declining, or at least stayed down, instead of bouncing right back and remaining productive for several years.

And you could even make the case that Curtis Martin was hit by the "Curse of Almost 370" when he was 23, after having 368 carries the year before. After all, there's nothing magical about 370. 368 is close enough. Martin rebounded after that, then was again hit by the "Curse of Almost 370" when he was 27, after having 369 and 367 carries in the previous two seasons. He then rebounded from it again, playing well for a few years until he finally met the true "Curse of 370" at the age of 32, and had by far the worst season of his career. He missed as many games in that season as he had in the previous 10 years, and only gained 735 yards, averaging 3.3 YPC, both career lows.

Yes, some of that last one was age. But keep in mind that just one year prior, at the ripe old age of 31, he had gained 1697 yards, averaging 4.6 YPC, both career highs. And if this were all decline due to age, how did Martin recover, not once, but twice, and have successful seasons later on?

Also, Curtis Martin only had 4 seasons with a YPC below 4 in his 11 year career. 3 of those years came immediately after seasons when he had between 367-371 carries. Doesn't sound like regression to the mean to me.

59
by oldnumberseven (not verified) :: Thu, 09/20/2007 - 12:19am

@#

Good baseball coverage at the P-G since Dejan took over, though.

:: The Other Vlad — 9/19/2007 @ 10:14 am

He is really missed in the Penguins coverage.

60
by JKL (not verified) :: Thu, 09/20/2007 - 12:25am

I wrote a two-part post on Running back overuse and injuries on the pro football reference blog back in July. Part one is linked in my name.

There is nothing magical about 370. However, high workload games do increase the risk of injury. Certainly, guys who get to above 370 carries in a 16 game season will for the most part have a high workload at the end of that season, many will be playing in the playoffs (correlation between rushes and wins and all), and will see the high use extended when the championship is on the line.

#12 raises a good point, if it were the end of the season, and a team were fighting for a superbowl. If the player is not a substantial cap risk if he is injured the next year, the risk may be worth it. I am not sure tacking carries on an already assured victory game is worth the increase risk early in the year, because he is still likely better than others on the roster. In my article, I use the word overuse rather than something like "abuse". Everyone is an adult here, and the running backs want as many carries as possible (more money, more fame, etc). Still, I think teams overuse running backs, often unneccessarily, without weighing the risk/reward. Some situations may justify the risk, most do not. I think the Pats have the right idea when they are giving Sammy Morris 10+ carries in these first two games in which they are up by double digits in the second half.

61
by johnt (not verified) :: Thu, 09/20/2007 - 12:56am

FO's statistics (as they will admit) are not especially good at separating running back performance from offensive line performance. There are attempts, of course, but nothing particularly definitive and certainly not rushing DVOA. Watching games is still the best way to figure out which is which.

And anybody with eyes who has watched the Steelers in the past 2 years will probably recognize the offensive line has had substantial problems opening holes for Parker. His cutback skills leave something to be desired, but when he gets a good hole he can regularly burst for a good 7-8 yard gain. One need only watch Verron Haynes and Najeh Davenports attempts at rushing with the same line to realize losing Parker would be a lot worse than losing an average RB.

62
by Brian (not verified) :: Thu, 09/20/2007 - 1:00am

I think I've pretty much made up my mind on the "curse."

5 yards per game drop off? After a career year? Come on.

If LT had a 0.24 YPC drop off from last year, he'd be thrilled. No one would call him cursed or abused.

Cherry-picked anecdotes don't convince anyone. Well, maybe some people.

63
by Alex (not verified) :: Thu, 09/20/2007 - 3:09am

I think I’ve pretty much made up my mind on the “curse.�

You seem to make up your mind on a lot of things pretty quickly. Maybe a little too quickly.

5 yards per game drop off? After a career year? Come on.

Where did you get that number? It's not 5 yards per game. RBs that have 370+ carries average 1732 yards. Since they decline by 35% on average, that's a decline of more than 600 yards, or more than 35 yards per game. That's huge.

You can't just say, "Yeah, but if they had the same number of carries the next year, with the 8% decline in YPC they had, they only would've declined by 5 yards per game." One of the main reasons they don't have as many carries the next year is that they get injured or worn down, and therefore can't carry the ball as often. It doesn't help you if your YPC only goes down a little bit, but you miss 4 games due to injury. Health is a skill, one that RBs generally lose when they go over 370 carries.

The 8% decline in YPC as opposed to the normal 2% decline for other RBs just shows that it's not only increased injury risk that causes the decline. They also have decreased effectiveness when they can play.

If LT had a 0.24 YPC drop off from last year, he’d be thrilled. No one would call him cursed or abused.

Again, it's not just about the YPC a RB gets, it's about the number of carries that a RB can sustain that YPC for. If you took off 35% of Tomlinson's rushing yards (and 8% of his YPC) from last year, you'd get about 1180 yards on 247 carries. And since he usually gets around 20 carries a game (his lightest full season load has been over 19.5 carries per game) that implies that he'd probably miss 3 or 4 games to injury if he were to suffer such a decline. That'd be the worst year of his career, by far.

So, gaining fewer yards than he had in any previous season, while losing 3 or 4 games to injury (after missing only 1 game in the previous 6 years), right out of the blue, after a career year? Yeah, that would qualify as cursed or abused to me.

64
by Alex (not verified) :: Thu, 09/20/2007 - 3:15am

an ACL Gremlin doesn’t suddenly materialize on carry no. 370

No, ACL Gremlins only materialize when Eagles fans start to really believe, deep down, that we might win a Super Bowl.

65
by langsty (not verified) :: Thu, 09/20/2007 - 3:57am

"I agree with 12. Parker is a former undrafted FA who has a modest contract and, according to FO, isn’t that good anyhow. Not to sound cold, but I’m willing to break him if it helps us this year

:: joe football — 9/19/2007 @ 4:25 pm"

I'm bringing this up next time you give me shit about the everett thing.

66
by brian (not verified) :: Thu, 09/20/2007 - 4:32am

Re 51: I don't think people are attracted to the theory because it involves a "curse" or a "magic number" but rather because it is brought up on this site fairly frequently and always stated as a fact, without there being solid evidence that it actually is a fact. At least, not in the statistical manner in which it is presented.

Re 54: This is correct. One particular week might be harder on a particular back then another week, even if he has less carries that week. However, there is no reason why every back (or most backs) would suddenly experience less wear carrying the ball in the post season. This is perhaps my biggest problem with the theory. Let’s say you flip a coin three times and all three times it comes up heads. One possible conclusion you might come to is that the coin is rigged and will always land on heads. Another possible conclusion you might come to is that it was simply coincidence that you got three straight heads. How can you know for sure? Simple, increase the sample size. If you flip it 1000 more times and they all come up heads, you probably have a rigged coin. If it comes up tails about half the time, then it was just coincidence the first three flips were heads. Now what does this have to do with football? Well, certainly quite a number of backs who ran 370+ times in the regular season “broke down� the following year. But you’re dealing with a small sample size because not many backs carry the ball 370+ times in the regular season. How do we know it isn’t just coincidence? Well, we increase the sample size. How do we do that? Include the post season. Once you include the post season, you have a much larger sample size of 370+ carry seasons. The problem for the theory though of course, is that many of those backs did not “break down� the year after. Now, to make the theory “work�, we’re told that some how, some way, post season carries bring less wear on the back. But this is circular logic. You have to start with the assumption that the theory is correct in order to accept discounting those post season carries. Unless there is a reason (other then just stating it to make the theory work) why post season carries shouldn’t count, a more objective analysis would be that the theory of 370+ carries is significantly flawed in some way at best, flat-out wrong at worst.

Re 63: There is no need to speculate about what might happen to Tomlinson as he has already carried 370+ times in one season (in the regular season). And, the following year he improved in every measurable way, both in conventional stats and DVOA.

67
by Alex (not verified) :: Thu, 09/20/2007 - 9:33am

But you’re dealing with a small sample size because not many backs carry the ball 370+ times in the regular season.

A sample size of 27 may not be great, but it's enough to establish a rough estimate of the "danger zone" for RB carries. It's no coincidence.

Now, to make the theory “work�, we’re told that some how, some way, post season carries bring less wear on the back.

Look, the whole point of the 370 carry rule is that an NFL RB's body can't usually take 370 carries in 17 weeks without sustaining significant damage. If you spread the carries out, giving the RB's body more time to heal in between carries, it would make perfect sense that they would be able to take a few more in a slightly longer time frame. They get a few more weeks for their bodies to heal when the carries come in the playoffs.

Remember, the theory isn't saying that a RB shouldn't get more than 370 carries total, for their entire career. The point is that a lot of carries in a short period of time is more damaging than the same number of carries spread out over a longer time frame. And it turns out that RBs are usually able to handle roughly 20 more carries in about a 20 week time period than they can in a 17 week time period. That's really only one more game's worth of carries in an additional 3 or 4 weeks. Not much.

Re 63: There is no need to speculate about what might happen to Tomlinson as he has already carried 370+ times in one season (in the regular season). And, the following year he improved in every measurable way, both in conventional stats and DVOA.

I wasn't speculating about what would happen to Tomlinson if he were to have 370+ carries. I was responding to the comment that suggested that, were he to suffer the average decline of a RB after a 370+ carry season, that he would still have a good year this year.

To address your point, yes, Tomlinson did go 2 carries over the recommended limit, and did fine the next year. But, despite being a durable RB, not getting carries in the preseason, and only barely going over the 370 carry mark, he did suffer a modest decline two years later, when his YPC dropped to 3.9, and he had the only missed game of his 6 year career.

68
by mactbone (not verified) :: Thu, 09/20/2007 - 9:42am

Re 66:
You do know that they play more games in the postseason and the carries therefore are spread out over more games... right?

69
by Smeghead (not verified) :: Thu, 09/20/2007 - 9:44am

Man, #12 and #46 make me think this might also be the first FO theory to be used in a lawsuit.

It's got to be only a matter of time before there's a test case, either in the courts or the locker room, right? Or a back/agent conscious of this issue who foul up a coach's game plan by refusing to take a certain workload or go onto the field after a certain point? If you thought having an extra three or four carries a game in December was going to be the difference in millions on your contract -- either because of breakdown yourself or the perception of overuse among GMs -- would you go run through a wall for your equally self-interested coach?

I'm getting a partial image here. I think it says "v. Herm Edwards" ...

70
by Rich Conley (not verified) :: Thu, 09/20/2007 - 11:25am

"Also, Curtis Martin only had 4 seasons with a YPC below 4 in his 11 year career. 3 of those years came immediately after seasons when he had between 367-371 carries. Doesn’t sound like regression to the mean to me."

no, but it also doesn't prove that running back carries mean anything.

Do we have any data on how offences that rush an extremely large amount of times do the next year? Could this just be "overly heavy rushing offenses tend to decline?"

71
by SGT Ben (not verified) :: Thu, 09/20/2007 - 11:25am

Brian seems to be way caught up in these magical numbers. He seems stuck completely on the 390/370 fix without considering that they backs have to take not only more actualy carries...but extra PRACTICE reps as well. We can't forget that each week is probably closer to 100+ more plays that the back is caught up in due to all the practice reps for preparing for the next season.

Also, while it may not seem like much, every week that they are practicing in the offseason (or playing) is another week that they are not resting their bodies. The more wear and tear you do to your body, the longer it will take to rest up (particularly when you factor in the effects of age.)

Also, it's not just a RB breakdown that causes the regression to the mean. Every additional carry is an additional piece of information for defensive coordinators to study. Also, if a team is running their back 370+ times, they are (more than likely) enjoying a successful season in the NFL.

Regression to the mean affects every other aspect of the teams performance. If they're having a stellar year (say, a 14-2 season before losing to the Patriots) it's only logical to assume that the team (as a whole) will not perform as well. When the team (as a whole) does not perform as well, the running game will not be utilized as often. LT is more likely to break the big run with more carries per game...but when he's not seeing the ball as much, his big runs (that boost his YPC) aren't going to be there...reducing his YPC significantly the next season.

Everything about the Chargers team last year would indicate that this season they will look much more pedestrian(LT with 370+ carries or not)

There is validity to the 370 carry theory...but you have to be able to look past JUST the running backs performance and factor in the rest of the teams performance as well.

72
by Rich Conley (not verified) :: Thu, 09/20/2007 - 11:27am

"One of the main reasons they don’t have as many carries the next year is that they get injured or worn down, and therefore can’t carry the ball as often."

ORRR, their team isnt as good. Remember that correlation between winning and carries?

73
by SGT Ben (not verified) :: Thu, 09/20/2007 - 11:29am

Sorry for the spelling errors in my post earlier. Fatfingered on my laptop!

74
by Rich Conley (not verified) :: Thu, 09/20/2007 - 11:30am

Look, theres absolutely no evidence that the 370 carries is bad for the running back. There IS evidence that it is a predictor for decline in the running OFFENSE.

Larry Johnson is going to have a bad year this year, but its not because Larry Johnson is breaking down. Its because the chiefs were decent last year, and suck this year.

75
by SGT Ben (not verified) :: Thu, 09/20/2007 - 11:51am

If 370 carries for the Running Back is bad for the Running Offense the next year...wouldn't that pretty much automatically imply that it's bad for the running back?

76
by Rich Conley (not verified) :: Thu, 09/20/2007 - 12:23pm

76. no. It implies that with overuse, some part of the running game is breaking down. Is it the RB? Maybe? Blocking? Maybe? Is it the fact that teams that run a ton tend to be good teams, and good teams regress to the mean? Maybe.

77
by Jericho (not verified) :: Thu, 09/20/2007 - 12:29pm

Well, when I think logically about pitch counts in baseball, I think of two things. One, pitching is itself an unnatural act. Repeating that act will put stress on the arm. Two, pitching is very mechanical. Keeping good mechanics reduces the stress on the arm, but when fatigued, mechanics break down. More pitches = more stress = bad mechanics. All of which leads to injuries.

And let's not forget the most impotant element of pitch counts in baseball. Age. The idea is to limit pitch counts on young pitchers, as they're still developing physically and can't handle as much stress.

When I look to football, I see running as a natural act, not an unnatural act. The unnatural part is getting pounded by 300 lb. lineman. And logically I can say that over time, more pouding will lead to injuries and reduced effectiveness.

What I have a problem saying is putting a threshold on the exact number. What is okay at 350 that becomes dangerous at 370? That's less than 2 carries a game! What's the big problem resulting from that, particularly given an offseason of rest?

I see guys like Eric Dickerson or Emmitt Smith (who violated the rule in 1991, 1992 - 444 carries!, 1994, and 1995), who seemed to do okay following multiple warning seasons. I see a guy like Curtis Martin who was effectively at the threshold three times prior and did okay.

So the problem seems to be where is the threshold? And the answers seems to be unknown. It will vary from player to player. And you won't know it until you break it. So it's best to exercise caution, but I also think one never really knows what will happen with these guys.

78
by Alex (not verified) :: Thu, 09/20/2007 - 12:53pm

It implies that with overuse, some part of the running game is breaking down. Is it the RB? Maybe? Blocking? Maybe? Is it the fact that teams that run a ton tend to be good teams, and good teams regress to the mean? Maybe.

Well, to figure out whether it's the RB or the rest of the team that suffers from overuse, you could just compare teams that had lots of carries from a RB by committee, where none of their RBs passed 370, to teams that have a similar number of carries, and similar effectiveness, but have a feature back go over 370 carries on his own. If the two groups of teams see a similar regression the next year, then it's probably just the team. But, if the teams that use a RBBC approach are better able to sustain good rushing performance the following year, then it's more likely that the RB is the one declining.

Also, you could just look at the teams that have a RB go over 370, and look at the rushing performance of other RBs on those teams, in both the 370 year, and the following year. If the other RBs also decline along with the feature back, then it's probably the team. But if they stay just as effective as they were, it's probably RB overuse.

79
by SGT Ben (not verified) :: Thu, 09/20/2007 - 1:40pm

Emmit Smith had one of the best offensive lines in front of him, a defense that played exceedingly well and a few other playmakers on offense to help keep teams from completely keying in on him. Dallas was looking to run out the clock as they were often winning games in Emmits heyday.

For Curtis Martin...there are exceptions to every rule...and the original talk concerning this isn't that 370 is a hard thresh-hold but a rough estimate (or a generalization) that backs will generally decline within two years after hitting that. Yes, Curtis approached the threshold several times, no one is arguin that. You could round his carries up to 370 and ask why...but for some reason, he continue to play well. Since I don't purchase game film...nor do I record games to re-watch, I can only guess that he didn't take a lot of hard shots. He pulled a "Marvin Harrison" where he would go down as he was being hit...to absorb the blow. There's a reason why "generally" big backs don't rush for 10K plus yards...because they take a bigger pounding than their smaller/fleet of foot contemporaries. Are there exceptions to this, yes!

As for the RBBC approach vs the Sole Possessor of 370+, the findings probably wouldn't be "concrete" in favor of a RBBC having less dropoff than the Single Back...but I'm guessing that the single back will (the majority of the time) have a more significant dropoff.

I'm pretty sure that there haven't been a lot of SB winners who have gone with the RBBC approach. Yes, the Colts did it...and other teams have manged it...but it's generally the 1 guy who pulls a team through.

Jericho is right in post #77, there is no known "hard number" that will tell us when something happens. Humans bring too many variables to the table to give that. But, 370 seems to be the "lowest number" that gives us these kinds of results. There are exceptions and will continue to be exceptions...but there are going to be far more Jamaal Andersons (Falcons) than Curtis Martin/Emmit Smiths around the league. Guys who are good...then have one "pounding" year...and never look the same again. The older they are when they pound the league...the more likely they are to fail to perform to a standard that is comporable to their Mean up to that point.

80
by Sid (not verified) :: Thu, 09/20/2007 - 4:00pm

problem is they've been winning easily. I don't have a problem with how they've used Najeh. He's come in late in games and saved Parker a few carries.

this week against San Fran could be a rout. Sit Parker in the 2nd half.

81
by Brian (not verified) :: Thu, 09/20/2007 - 6:58pm

If regression to the mean is sufficient to explain the observation, why are we straining so hard to add an additional explanation? It's simply unnecessary and very unscientific.

If the wear-and-tear factor is real, wouldn't we see a steady decine from week-to-week during the "overuse" season? Why would the overuse effect sneak up on someone months and months after the overuse? (Or, as some people suggest, two years later!)

The answer is because it probably doesn't. I think it might be time for a basic statistics and logic refresher.

Everyone here understands the correlation doesn't equal causation rule. One variable could be causing a second, or the second could be causing the first. That's true, but there are other possibilities too.

If two variables are correlated, there could be no relationship at all--just luck. Most people probably understand that too.

But I get the impression most people here overlook the "lurking" variable that causes the correlation in both variables. In other words, the two correlating variables do not affect each other at all, but are both affected by a third factor.

I think the "270 curse" correlation may be a good example of this.

The high number of carries and the following year's drop off in performance can both be explained by exceptional performance.

A RB having a career year (varying highly above his career mean in YPC) would be given a high number of carries. His team, benefitting from his exceptional year would also likely be ahead in more games than average, which also leads to even more carries.

The following year, it is very unlikely a RB will have back-to-back career-type seasons, so his YPC performance would be expected to decline towards his career mean. Accordingly, he would not see as many carries in the following year. (Keep in mind this is a general tendency, and exceptions do happen.)

The FO theory about line continuity suffers from the same analytical error. O-line health leads to both better performance and to continuity. Continuity doesn't necessarily cause the performance.

There is one additional logical problem with the overuse theory regarding injuries. High-carry seasons can only happen in injury-free years. So there is a very strong selection bias in any correlation with injuries for years following high-carry years. Of course non-high carry years will appear to have more frequent injuries.

We, including the FO writers, should be far more careful about jumping to conclusions when we see a couple correlations.

Even strict academic studies that pass formal peer review are later found to have problems and mistakes that even reverse the findings. FO articles and theories are always clever and interesting, but not always true. Even some of Bill James's baseball studies have been discredited over the years.

I get the feeling some guys have an emotional bias towards FO dogma. That's ok, really. But please allow for those of us who like to apply some critical thought to these topics. I hope agreement is not a prerequisite for discussion.

82
by Xao (not verified) :: Thu, 09/20/2007 - 7:44pm

If regression to the mean is sufficient to explain the observation, why are we straining so hard to add an additional explanation? It’s simply unnecessary and very unscientific.

I'm assuming that this is a rhetorical question, but there's been no evidence that regression to the mean is sufficient to explain the observation.

If the wear-and-tear factor is real, wouldn’t we see a steady decine from week-to-week during the “overuse� season? Why would the overuse effect sneak up on someone months and months after the overuse? (Or, as some people suggest, two years later!)

Physical stress has cumulative as well as incremental effects. Ask anyone who's spent a few years in a light infantry unit. You can also frequently mask immediate physiological responses to such stress with, ah, medical supplements. Suppressing the side effects doesn't eliminate the damage, though it does allow you to function temporarily, and you will pay the body's bill later on.

A RB having a career year (varying highly above his career mean in YPC) would be given a high number of carries. His team, benefitting from his exceptional year would also likely be ahead in more games than average, which also leads to even more carries.

This is another pair of unsupported assumptions.

The following year, it is very unlikely a RB will have back-to-back career-type seasons, so his YPC performance would be expected to decline towards his career mean. Accordingly, he would not see as many carries in the following year.

Yet another bit of speculation with no evidence behind it. About that logic refresher...

83
by SGT Ben (not verified) :: Thu, 09/20/2007 - 8:22pm

As indicated by XAO...long term effects rarely show up during "the war". Having spent the last 10 years in the military and deployed three times...I've been around a lot of different people and types of people.

A lot of our "Air Assualt" guys do a great job for years...but it's the continuance of doing it that enables them to do so. The instant they get away from their jobs, they find that there are far more problems associated with what they've done...and have a harder time going back to it. Same for my infrantry brothers.

The same rules apply to the running backs of the NFL. While they are going through the rigors of the season, the damage isn't nearly so apparent and easily dismissed psychologically. However, once the offseason begins, the damages come to light. This is why the hypothetical 370 is such a big deal. It's harder to come back from the pounding of 370 carries (not even including the catches out of the backfield or the 1000s of practice reps) than it is when you go RBBC with 200 carries (and less reps in practice, because you share the load there too.)

While the "370" isn't a rule (per se), I'm going to call it that for simplicity's sake.

Backs that get that 370 are (generally) having a good season...they're healthy and running well. That statement can be agreed in a generality...because what coach is going to feed a back 370 times if he's averaging less than 3.5 YPC?

Because they're having a good season...they're probably going to see some regression the next season. Whether it's age, health, coaching changes (San Diego comes to mind) or whatever...there will be some regression.

The greater your success, the more likely your regression to the mean. No one seems to be arguing this point.

Coupling it together, the 370 rule is what it is...not because of just the back...but because of ALL the factors that play into a running backs success. It's nearly impossible for a back to reach and maintain that thresh-hold season after season. It's even more difficult in the Free Agency era (for you Dickerson pundits.)

So, while there have been backs who've enjoyed multiple success of the 370 thresh-hold rule...they are few and far between...and Hall-of-Fame caliber type runningbacks. We all know that very few teams have those...

That's why it seems as if people are discounting Curtis Martin or Emmit Smith...The "Average or Mean" runningback or season will be exceeded by few people. It will happen, but that's why we can go with a generalized 370 rule. There are always exceptions to rules...always will be. Look at NO last year. Everything about them screamed a team that wasn't as good as their record/stats indicated. Look at their First/Second down DVOA. This year, their 3rd down rate has reduced towards the mean (and passed it by on it's way down) and the are back to looking like the 'Aints.

DVOA, DAVE, VOA is not infalliable. Neither is the 370 rule. But, it's fitting in most scenarios that you'll run into in your NFL-watching career.

84
by Alex (not verified) :: Thu, 09/20/2007 - 8:51pm

I see a guy like Curtis Martin who was effectively at the threshold three times prior and did okay.

He did okay, but wasn't nearly as good in the years right after approaching the threshold. He suffered a significant decline within two years of each of the three seasons when he came close to 370. He only had 4 seasons in his 11 year career where he averaged below 4 YPC. Three of them came right after years with 367+ carries.

Curtis Martin isn't an exception to the rule at all. If anything, he's proof that RB overuse can be a problem even before you reach 370 carries.

So the problem seems to be where is the threshold? And the answers seems to be unknown. It will vary from player to player.

True. That's why, for instance, if you have a durable elite RB who doesn't get any carries in the preseason, and you're chasing a playoff bid, it's not unreasonable to push him a few carries over 370 while you're desperately trying to win 2 close games at the end of the season. The risk might very well be worth the reward in that case, which is why giving Tomlinson 372 carries in 2002 wasn't a terrible idea. It didn't work, unfortunately, but at least he didn't suffer a decline in performance until two years later, and even then it wasn't all that bad.

On the other hand, pushing your RB almost 40 carries over the 370 mark, like Herm Edwards did to LJ in 2006, is absurd. When you give your RB 416 carries, you haven't just passed the 370 carry barrier, you've gone to plaid!

I’m pretty sure that there haven’t been a lot of SB winners who have gone with the RBBC approach. Yes, the Colts did it…and other teams have manged it…but it’s generally the 1 guy who pulls a team through.

I guess I should've been a bit more clear. I didn't necessarily mean that they have to split the carries evenly, or anything like that. They can still have one RB as the clear #1. What I meant was that teams that run a lot, but spread enough of the carries out to their backup RBs to keep the starter out of the danger zone, are more likely to sustain their rushing success the next year.

If you look at the last 20 years, only three SB winners have had a RB with 370+ carries in the regular season. Only 7 of the last 20 SB winners, and only 1 of the last 5, have even had a RB with 300+ carries in the regular season! And these weren't teams that avoided the run, either. 14 of the last 20 SB winners have been in the top 10 in total carries.