Writers of Pro Football Prospectus 2008

Most Recent FO Features


» Factors: Andre Johnson

One of the NFL's best receivers notched a -2.3% DVOA last year. Does a target-by-target breakdown show he was better than that?

17 Jul 2009

Do Running Backs See a "Verducci Effect"?

by Aaron Schatz

You may have noticed Will Carroll mention in his column today that, after considering the likely rise in carries that Maurice Jones-Drew will get this season, he asked me to research if there is some sort of "Verducci Effect" in football. For those who don't know, the "Verducci Effect" comes from Sports Illustrated's Tom Verducci, who noticed that pitchers who go far past their previous career high in innings pitched tend to have injury or decline issues the following season. So the question here is whether the same thing happens with part-time running backs who become high-workload full-time running backs. Is there an effect from increased workload, and is it anything like the effect of ultra-high workload that we call "The Curse of 370."

(As an aside, I know that some people have questioned the Curse of 370 recently. It was always meant to be a rough guideline, not a hard and fast rule -- it isn't as if 369 carries is "safe" and 372 carries automatically means "screwed" -- and I will add that I'm totally open to considering criticism as to why this theory may be flawed. However, the best way to make that criticism is not to leave discussion thread responses but rather to e-mail well-explained, constructive comments to info-at-footballoutsiders.com -- and the key word here is "constructive" as opposed to "indignant." Digression over.)

Don't consider this fully-vetted, 100-percent scientifically accurate research, but I did a quick-and-dirty search to look at backs with a major workload increase since 1978. First, we need a control group: Every running back since 1979 with at least 250 carries in a season. These running backs on average dropped the following year by 325 yards and .25 yards per carry. (The yardage average includes guys who retired or did not play the following year due to injury; the yards per carry average does not.)

Next, I looked at "workload increase" backs. This group included those who:

1) Carried the ball at least 250 times
2) Previously had a career high below 150 carries
3) Increased their workload by 200 or more carries compared to the previous season

This gives us 19 running backs, including Michael Turner. The other 18 running backs in the sample dropped the following year by an average of 230 yards and .26 yards per carry. The problem is that not all these backs are guys you would think of as "part-timer with sudden workload increase." Seven of them were in their second year -- those aren't part-timers, those are guys who weren't given a heavy workload as rookies, such as LenDale White and Shaun Alexander.

If we take out the second-year players, there does seem to be a small "increased workload effect." The remaining 11 players dropped the following year by an average of 368 yards and .38 yards per carry. Of course, you can't really go by such a small sample size, and part of the reason for the drop in yards per carry is that some of these guys were absurdly good in their "increased workload" year. After Larry Johnson averaged 5.2 yards per carry in 2005, would you blame "increased workload" for his drop to 4.3 yards per carry the next season? Or would you blame plain old regression to the mean, not to mention the deteriorating Kansas City line of that time?

I tried increasing the sample a little bit by looking at backs who:

1) Were in their third year or older
2) Carried the ball at least 225 times
3) Previously had a career high below 150 carries
4) Increased their workload by 150 or more carries compared to the previous season

Now our sample is back to 19 backs, including Michael Turner. The 18 other backs here dropped the following year by an average of 355 yards and .32 yards per carry. I put those backs in a table below for those curious.

Finally, let's compare these "increased workload" groups to the good ol' Curse of 370. Twenty-eight backs have gone over 370 carries during the regular season, including "Curse-Buster" Eric Dickerson four times. Those backs dropped in the following year by an average of 714 yards and .55 yards per carry -- a much larger drop.

Overall, I would say the difference is too thin and the sample size is too small to say that running backs with a massively increased workload are in particular danger -- unless, of course, that workload went up into the high 300s.

Increased Workload Backs Since 1979
Name Years Team Prev Career
High Carries
Increased Workload Season Following Season
Runs Yds TD Yd/Car Runs Yds TD Yd/Car
James Allen 2000-01 CHI 58 290 1120 2 3.86 135 469 1 3.47
Jamal Anderson 1996-97 ATL 39 232 1055 5 4.55 290 1002 7 3.46
Ladell Betts 2006-07 WAS 90 245 1154 4 4.71 93 335 1 3.60
Stephen Davis 1999-00 WAS 141 290 1405 17 4.84 332 1318 11 3.97
Reuben Droughns 2004-05 DEN 30 275 1240 6 4.51 309 1232 2 3.99
Barry Foster 1992-93 PIT 96 390 1690 11 4.33 177 711 8 4.02
Ahman Green 2000-01 GB 35 263 1175 10 4.47 304 1387 9 4.56
Gaston Green 1991-92 DEN 68 261 1037 4 3.97 161 648 2 4.02
Troy Hambrick 2003-04 DAL 113 275 972 5 3.53 63 283 1 4.49
Mark Higgs 1991-92 MIA 49 231 905 4 3.92 256 915 7 3.57
Larry Johnson 2005-06 KC 120 336 1750 20 5.21 416 1789 17 4.30
Lamont Jordan 2005-06 OAK 93 272 1025 9 3.77 114 434 2 3.81
Dorsey Levens 1997-98 GB 121 329 1435 7 4.36 115 378 1 3.29
Joe Morris 1985-86 NYG 133 294 1336 21 4.54 341 1516 14 4.45
Erric Pegram 1993-94 ATL 101 292 1185 3 4.06 103 358 1 3.48
Gerald Riggs 1984-85 ATL 100 353 1486 13 4.21 397 1719 10 4.33
Lewis Tillman 1994-95 CHI 121 275 899 7 3.27 29 78 0 2.69
Michael Turner 2008-09 ATL 80 376 1699 17 4.52 -- -- -- --
Charles White 1987-88 LARM 97 324 1374 11 4.24 88 323 0 3.67
Harvey Williams 1994-95 LARD 97 282 983 4 3.49 255 1114 9 4.37

Posted by: Aaron Schatz on 17 Jul 2009

10 comments, Last at 29 Jul 2009, 11:39am by dbostedo


by Bartles N. James (not verified) :: Fri, 07/17/2009 - 9:54pm

this study seems inconclusive, as the numbers don't account for circumstances.

for example, Betts' increase was due to coming off the bench to start for an injured Portis, his regression the next year was due to his returning to a back-up role.

Hambrick took over as a starter for the departed Smith in 03, and wasn't even on the same team the year following his increased workload season, having followed Smith to AZ and returning to a back-up role, also.

by Anonymous* (not verified) :: Sat, 07/18/2009 - 1:04pm

Overall, I would say the difference is too thin and the sample size is too small to say that running backs with a massively increased workload are in particular danger -- unless, of course, that workload went up into the high 300s.

Drink wine coolers less, read more.

by 21 Cards (not verified) :: Sun, 07/19/2009 - 9:11am

Agree with Bartles. If you had a filter for 'starter level production' at 4 ypc minimum the first year, we actually see that there is absolutely no dropoff in production the second. The only case of a significant drop with that filter was with Charles White whose career was already gasping for breath. His '87 season came as a grizzled veteran who already had plenty of miles on his body.

by S :: Mon, 07/20/2009 - 11:17am

So I guess the answer to Will Carroll's question would be "who knows?" This gorup is an odd mix of the types of players described in comment #1, sprinkled with a few guys experiencing breakout seasons (usually regressing to the mean the next year). The only two on this list who I remeber having a "curse of 370" type experience the next year were Dorsey Levens and Charles White.

Would there be any way to modify the baseline i.e. adjust it according to league average number of carries per year? That may generate a larger sample that would include more years of data (though you would have to deal with "between-era" effects).

by bengt (not verified) :: Wed, 07/22/2009 - 9:57am

I'm unhappy with including to-be-retiring players in the yardage total of the control group. That doesn't seem right. For instance, it probably includes Ricky Williams earlier this millennium.

by anonymiss (not verified) :: Thu, 07/23/2009 - 3:33pm

So all constructive criticism of this site should be done in private and not in public? That's silly. If someone has a good and respectful deconstruction, there's absolutely no reason why it shouldn't be offered - and respectfully debated - on these pages.

by Bowl Game Anomaly :: Thu, 07/23/2009 - 8:19pm

Yes there is- because the writers don't necessarily read the comments, but they do read emails.

If you want to have a debate with other readers, the comments are fine. But if you want the writers to take notice, the comments are just not an effective forum for that.

by anonymiss (not verified) :: Fri, 07/24/2009 - 1:24am

Your reply makes me think that a reasoned counter should be sent to both places, here and there. But I think it's pretty obvious the outsiders don't like being questioned in public, no matter how reasonable and rational the critique may be. And that's a shame - the truth should come before ego.

by Eddo :: Fri, 07/24/2009 - 12:05pm

The writers do read the comments here, much of the time (think of how often you see the green comment boxes). And I have seen them reply to constructive criticism in comment form as well.

I think Aaron's point was that if you want your question or critique addressed, emailing someone is a more reliable method, as often the writers are too busy writing to check the comment threads regularly.

by dbostedo :: Wed, 07/29/2009 - 11:39am

"But I think it's pretty obvious the outsiders don't like being questioned in public, no matter how reasonable and rational the critique may be."

Care to explain in what way that is obvious? I'd argue exactly the opposite, with the existence of these comment threads being a prime example.