The Curious Case of 350+ Carry Backs... (1 Viewer)

[CalBear]

Those that claim that Football Outsiders does not understand statistics does not understand what he is saying. There's a point where you can disagree with someone or a hypothesis and have a healthy debate around the premise. Perhaps, poke some fun. But, I don't see anything here that suggests the theory is totally debunked, nor do I do I see anything here to suggest that the FOs don't understand statistics. You're on my side of this debate, but I'm standing at a distance from the other inferences you draw here.

One of these two things is true:* FO doesn't understand statistics

* FO is being disingenuous

Because the way they're abusing statistics is totally obvious to anyone who knows statistics.

I have a rule that's a little like the 370-carry rule. I call it the 369-carry rule. It says that guys who get exactly 368 or 369 carries in a year are destined to have one of their best career seasons in year N+1. Here's the run-down:

Walter Payton 1979 (followed up with 1460 rushing yards)

Earl Campbell 1979 (followed up with 1934 rushing yards, best of his career)

Emmitt Smith 1994 (followed up with 1773 rushing yards and 25 TD, both career bests)

Curtis Martin 1995 (followed up with a 17-TD season, best of his career)

Curtis Martin 1998 (followed up with 1464 rushing yards, third-best of his career)

Terrell Davis 1999 (followed up with 2003 rushing yards, best of his career)

Edgerrin James 1999 (followed up with 1709 rushing yards and 18 TD, both career bests)

That is the complete list of 368-369 carry backs. Other than Payton, who had a decent but not outstanding season, every one of these backs had an exceptional year N+1 after getting 368 or 369 carries. Football Outsiders has to have known this if they were doing statistical analysis--it's the reason they chose 370 carries. So, again, one of three things is true:

* FO believes that backs who get 369 carries are destined to have great seasons in year N+1

* FO believes that something magical happens to backs when they get one more carry after getting 369 on the year

* FO is disingenuously curve-fitting to fit their hypothesis

I'm betting on #3.

The point is, you can do this with any stat; just keep going higher until the data points that contradict your hypothesis disappear. If any of these seasons were included in the tiny sample sizes, they would have thrown off the results that FO wanted, so they moved the yardsticks to exclude these seasons. That's why their study is meaningless; if Earl Campbell got one more carry in 1979, their extremely weak correlations would have disappeared. (But then they would have moved the yardsticks to 375 carries instead of 370).

 

[Jason Wood]

Those that claim that Football Outsiders does not understand statistics does not understand what he is saying. There's a point where you can disagree with someone or a hypothesis and have a healthy debate around the premise. Perhaps, poke some fun. But, I don't see anything here that suggests the theory is totally debunked, nor do I do I see anything here to suggest that the FOs don't understand statistics. You're on my side of this debate, but I'm standing at a distance from the other inferences you draw here.

One of these two things is true:* FO doesn't understand statistics

* FO is being disingenuous

Because the way they're abusing statistics is totally obvious to anyone who knows statistics.

I have a rule that's a little like the 370-carry rule. I call it the 369-carry rule. It says that guys who get exactly 368 or 369 carries in a year are destined to have one of their best career seasons in year N+1. Here's the run-down:

Walter Payton 1979 (followed up with 1460 rushing yards)

Earl Campbell 1979 (followed up with 1934 rushing yards, best of his career)

Emmitt Smith 1994 (followed up with 1773 rushing yards and 25 TD, both career bests)

Curtis Martin 1995 (followed up with a 17-TD season, best of his career)

Curtis Martin 1998 (followed up with 1464 rushing yards, third-best of his career)

Terrell Davis 1999 (followed up with 2003 rushing yards, best of his career)

Edgerrin James 1999 (followed up with 1709 rushing yards and 18 TD, both career bests)

That is the complete list of 368-369 carry backs. Other than Payton, who had a decent but not outstanding season, every one of these backs had an exceptional year N+1 after getting 368 or 369 carries. Football Outsiders has to have known this if they were doing statistical analysis--it's the reason they chose 370 carries. So, again, one of three things is true:

* FO believes that backs who get 369 carries are destined to have great seasons in year N+1

* FO believes that something magical happens to backs when they get one more carry after getting 369 on the year

* FO is disingenuously curve-fitting to fit their hypothesis

I'm betting on #3.

The point is, you can do this with any stat; just keep going higher until the data points that contradict your hypothesis disappear. If any of these seasons were included in the tiny sample sizes, they would have thrown off the results that FO wanted, so they moved the yardsticks to exclude these seasons. That's why their study is meaningless; if Earl Campbell got one more carry in 1979, their extremely weak correlations would have disappeared. (But then they would have moved the yardsticks to 375 yards instead of 370).

Interesting find...there certainly is some logic to your contention given the seemingly arbitrary nature of 370 carries. 320 you could dismiss as innocent thinking they just looked at "20-carry per game" backs, but 370? Odd.

 

[ConstruxBoy]

Hasn't FO admitted they curve fit the data? I thought they had.

 

[Kool-Aid Larry]

Myth of 370

It's been thoroughly debunked by those that understand statistics. Those that say it has not been debunked do not understand statistics.

from the comments:

I should say that I think the published article on the Curse was complete rubbish and was really shoddy with the statistics to the point where it was a violent assault upon reason

lolzI think we can call this myth BUSTED.

 

[Kool-Aid Larry]

Those that claim that Football Outsiders does not understand statistics does not understand what he is saying. There's a point where you can disagree with someone or a hypothesis and have a healthy debate around the premise. Perhaps, poke some fun. But, I don't see anything here that suggests the theory is totally debunked, nor do I do I see anything here to suggest that the FOs don't understand statistics. You're on my side of this debate, but I'm standing at a distance from the other inferences you draw here.

One of these two things is true:* FO doesn't understand statistics

* FO is being disingenuous

Because the way they're abusing statistics is totally obvious to anyone who knows statistics.

I have a rule that's a little like the 370-carry rule. I call it the 369-carry rule. It says that guys who get exactly 368 or 369 carries in a year are destined to have one of their best career seasons in year N+1. Here's the run-down:

Walter Payton 1979 (followed up with 1460 rushing yards)

Earl Campbell 1979 (followed up with 1934 rushing yards, best of his career)

Emmitt Smith 1994 (followed up with 1773 rushing yards and 25 TD, both career bests)

Curtis Martin 1995 (followed up with a 17-TD season, best of his career)

Curtis Martin 1998 (followed up with 1464 rushing yards, third-best of his career)

Terrell Davis 1999 (followed up with 2003 rushing yards, best of his career)

Edgerrin James 1999 (followed up with 1709 rushing yards and 18 TD, both career bests)

That is the complete list of 368-369 carry backs. Other than Payton, who had a decent but not outstanding season, every one of these backs had an exceptional year N+1 after getting 368 or 369 carries. Football Outsiders has to have known this if they were doing statistical analysis--it's the reason they chose 370 carries. So, again, one of three things is true:

* FO believes that backs who get 369 carries are destined to have great seasons in year N+1

* FO believes that something magical happens to backs when they get one more carry after getting 369 on the year

* FO is disingenuously curve-fitting to fit their hypothesis

I'm betting on #3.

The point is, you can do this with any stat; just keep going higher until the data points that contradict your hypothesis disappear. If any of these seasons were included in the tiny sample sizes, they would have thrown off the results that FO wanted, so they moved the yardsticks to exclude these seasons. That's why their study is meaningless; if Earl Campbell got one more carry in 1979, their extremely weak correlations would have disappeared. (But then they would have moved the yardsticks to 375 yards instead of 370).

Interesting find...there certainly is some logic to your contention given the seemingly arbitrary nature of 370 carries. 320 you could dismiss as innocent thinking they just looked at "20-carry per game" backs, but 370? Odd.

d00d, check the link.ps

SSHHHH!!! Stop giving away the good angles, CalBear.

 

[Kool-Aid Larry]

I thought I would throw a list in so we can see who we're talkin' about:

2008:

Turner - 377 carries (+6 catches) -- he only had 6 catches??

Peterson 364(+21)

Portis 342(+28) --- hmmm...some tough luck on that one....

Forte 315(+64)

2007:

Portis 325(+47)

Tomlinson 315(+60)

Westbrook 278(+90) --- 2008 should've been a career year

2006:

L Johnson 416(+41)

Tomlinson 349(+56)

Jackson 346(+90)

T Barber 327(+58) - knocked into retirement

Gore 313(+61)

R Johnson 341(+23) close....

 

[switz]

Westbrook 278(+90) --- 2008 should've been a career year

Actually, following an equation I've developed, in 2006 all signs pointed to '07 being Westy's career year, and 2008 being a big drop off.The formula worked, and has worked on other players - it's one of the reasons I'm pretty high on Turner and Peterson having HUGE years in '09.

 

[Kool-Aid Larry]

who will have huge drop offs?

 

[Maurile Tremblay]

Hasn't FO admitted they curve fit the data? I thought they had.

I think so. They also had a tongue-in-cheek Eric Dickerson exception.Which is to say, in FO's defense, that I don't think they ever presented the Curse as being statistically sound. Just something that was kind of fun, and may (or may not) eventually prove to have merit. (Same with their method of identifying which college QBs would go on to be successful pros.)

 

[JKL]

The concern I have with anecdotal stories about players who had big workloads and the...*gasp*...broke down is that, without any sort of analysis to compare the groups, you're just sort of stuck doing an eyeball test without fixating on anything to compare your test to. You've given a neat and detailed account of guys with large workloads who fell off the cliff a year later (or, in some cases, on the injury sheet a week later). But, there is no accounting for the regularity with which guys break down, regardless of workload.. . . I think what you need is a comparison group and to run some analyses comparing these guys. I would predict that there is no correlation between workload and injury (or performance).

I'm sorry, but I do not think it is a fair criticism to say that I told anecdotal stories, did not account for the regularity with which guys break down regardless of workload, or did not have a comparison group. In post one, I listed all backs who played in each of the first 6 games and had at least 90 carries in those six games, from 1995-2006. I broke the backs down into subgroups. Of the backs who had between 90 and 137 carries through 6 games (averaged 15.0 to 22.9), 10 out of 179 (5.6%) missed at least half of the games the rest of the year. I would count that as a comparison group of "other starting running backs". The number was 5.6%. The backs with 138-149 carries, 3 of 24 missed half the remaining games. The backs with 150+ carries, 2 out of 9 missed at least half the remainder.Now we can talk about small sample sizes at the performance extremes, what p-value is appropriate (does a real world NFL team really want to require a p-value of .05 before they consider that excessive work rate may lead to injuries? what is the appropriate confidence level in this case?), and such. But to say I didn't compare or provide any basis to assess whether the injury rates were higher is wrong.Let's move to the end of season post. Similar methodology. There, I looked at season ending injuries within the first 6 games of the next season, based on workload. The similar "control" group, those starting backs who finished with between 15.0 and 22.9 carries over the final 6 weeks, suffered a season ending injury 5.2% of the time (8 of 153). 8.9% of backs who averaged 23.0 to 24.9 had season ending injuries (2 of 23) and 28.6% of backs who averaged 25.0 or more had season ending injuries (4 of 14). Again, small sample sizes at the extreme, and injury rates within those small sample sizes in excess of the comparison group of backs with between 15.0 and 22.9 carries per game.Now, did I also do specific research into the nature of each injury because the sample sizes were small. Yes. But that doesn't mean I told anecdotal stories or didn't use specific methodology established before I looked at the data.

 

[Kool-Aid Larry]

A couple things I would note on that would be to hammer on the small sample size issue, which you're obviously already aware of, but also on injury rates per carry.

Each touch a back gets is an added chance for injury, so I think it's the natural state of things for a 25 tpg back to outproduce the 15 tpg back in injuries, as well as yards, etc.

If you really want to avoid injuries, find the guy who doesn't play at all.

 

[cubbie5150]

A couple things I would note on that would be to hammer on the small sample size issue, which you're obviously already aware of, but also on injury rates per carry.Each touch a back gets is an added chance for injury, so I think it's the natural state of things for a 25 tpg back to outproduce the 15 tpg back in injuries, as well as yards, etc.If you really want to avoid injuries, find the guy who doesn't play at all.

...or practice, or get out of bed in the morning.

 

[Fodasme69]

who will have huge drop offs?

 

[AB in DC]

One of these two things is true:* FO doesn't understand statistics* FO is being disingenuous

You missed the third possibility:*Calbear et al don't realize that FO was only being half-serious and completely missed the joke.

 

[cobalt_27]

The concern I have with anecdotal stories about players who had big workloads and the...*gasp*...broke down is that, without any sort of analysis to compare the groups, you're just sort of stuck doing an eyeball test without fixating on anything to compare your test to. You've given a neat and detailed account of guys with large workloads who fell off the cliff a year later (or, in some cases, on the injury sheet a week later). But, there is no accounting for the regularity with which guys break down, regardless of workload.. . . I think what you need is a comparison group and to run some analyses comparing these guys. I would predict that there is no correlation between workload and injury (or performance).

I'm sorry, but I do not think it is a fair criticism to say that I told anecdotal stories, did not account for the regularity with which guys break down regardless of workload, or did not have a comparison group. In post one, I listed all backs who played in each of the first 6 games and had at least 90 carries in those six games, from 1995-2006. I broke the backs down into subgroups. Of the backs who had between 90 and 137 carries through 6 games (averaged 15.0 to 22.9), 10 out of 179 (5.6%) missed at least half of the games the rest of the year. I would count that as a comparison group of "other starting running backs". The number was 5.6%. The backs with 138-149 carries, 3 of 24 missed half the remaining games. The backs with 150+ carries, 2 out of 9 missed at least half the remainder.Now we can talk about small sample sizes at the performance extremes, what p-value is appropriate (does a real world NFL team really want to require a p-value of .05 before they consider that excessive work rate may lead to injuries? what is the appropriate confidence level in this case?), and such. But to say I didn't compare or provide any basis to assess whether the injury rates were higher is wrong.Let's move to the end of season post. Similar methodology. There, I looked at season ending injuries within the first 6 games of the next season, based on workload. The similar "control" group, those starting backs who finished with between 15.0 and 22.9 carries over the final 6 weeks, suffered a season ending injury 5.2% of the time (8 of 153). 8.9% of backs who averaged 23.0 to 24.9 had season ending injuries (2 of 23) and 28.6% of backs who averaged 25.0 or more had season ending injuries (4 of 14). Again, small sample sizes at the extreme, and injury rates within those small sample sizes in excess of the comparison group of backs with between 15.0 and 22.9 carries per game.Now, did I also do specific research into the nature of each injury because the sample sizes were small. Yes. But that doesn't mean I told anecdotal stories or didn't use specific methodology established before I looked at the data.

Can you answer the question whether the difference in these groups is reliable? Is it predictive of anything? Does it mean that players who work hard get injured more, or does it mean that players who have greater opportunities to get injured get injured more? You're implying causation without any effort given to establishing a proper correlation, let alone the nature of the relationship (if it even exists, in the first place).

 

[cobalt_27]

Westbrook 278(+90) --- 2008 should've been a career year

Actually, following an equation I've developed, in 2006 all signs pointed to '07 being Westy's career year, and 2008 being a big drop off.The formula worked, and has worked on other players - it's one of the reasons I'm pretty high on Turner and Peterson having HUGE years in '09.

Did your formula fail to predict anyone accurately?

 

[Kool-Aid Larry]

I'm not sure if you need a formula for Peterson, but I think I might actually be a little down on Turner.

It's possible I'm just biased because I was a skeptic last year, but I think he did a lot of damage against soft targets last year, and the schedule might not be so favorable in '09.

 

[cobalt_27]

I'm not sure if you need a formula for Peterson, but I think I might actually be a little down on Turner.It's possible I'm just biased because I was a skeptic last year, but I think he did a lot of damage against soft targets last year, and the schedule might not be so favorable in '09.

As long as it's not because he had a high number of carries last year. That myth didn't hold up well in this thread.

 

[switz]

Westbrook 278(+90) --- 2008 should've been a career year

Actually, following an equation I've developed, in 2006 all signs pointed to '07 being Westy's career year, and 2008 being a big drop off.The formula worked, and has worked on other players - it's one of the reasons I'm pretty high on Turner and Peterson having HUGE years in '09.

Did your formula fail to predict anyone accurately?

I didn't apply it to everyone, only players that matched entry criteria. In those cases the only player it failed on - and note miserably - was Joseph Addai. And the big difference between him and the other players included was that he split time his rookie season (but so did Peterson, so I'm not sure whether that's a valid criteria or not)

 

[cobalt_27]

Westbrook 278(+90) --- 2008 should've been a career year

Actually, following an equation I've developed, in 2006 all signs pointed to '07 being Westy's career year, and 2008 being a big drop off.The formula worked, and has worked on other players - it's one of the reasons I'm pretty high on Turner and Peterson having HUGE years in '09.

Did your formula fail to predict anyone accurately?

I didn't apply it to everyone, only players that matched entry criteria. In those cases the only player it failed on - and note miserably - was Joseph Addai. And the big difference between him and the other players included was that he split time his rookie season (but so did Peterson, so I'm not sure whether that's a valid criteria or not)

Well, I think your formula rocks!

 

[Hipple Long Ware & Peete]

can we get a recap for those involved. reading all the N +1 's hurts my brain.

 

[Kool-Aid Larry]

Curse of the 370 is a crock, but some people have differing opinions.

edit: Also, Switz has a secret formula and he wants us to know it.

 

[cobalt_27]

can we get a recap for those involved. reading all the N +1 's hurts my brain.

Exactly what Kool-Aid said. Despite much buzzing here and there that suggests players wear down after significant use, there are no data to suggest that RBs underperform the year after heavy use when compared to any other group of RBs who had less use. And, Switz wants you to know that he's got a secret formula.

 

[IHEARTFF]

Man threads like this are why I love this site. I am also thrilled that I am reading all of this content in May for my 1st ever dynasty league. I love how much you guys are already talking football, it is great.

 

[cobalt_27]

Gratuitous bump, since Matthew Berry Tristan Cockroft and others seem to still believe in the "Curse of 370" nonsense.

For those new to this thread, the fun starts on post 16 and goes from there.

Moral of the story, you can expect Michael Turner to regress to the mean and produce less than last season, but...not based on any so-called curse.

Go.

 

[Steed]

Gratuitous bump, since Matthew Berry and others seem to still believe in the "Curse of 370" nonsense.

For those new to this thread, the fun starts on post 16 and goes from there.

Moral of the story, you can expect Michael Turner to regress to the mean and produce less than last season, but...not based on any so-called curse.

Go.

Matthew Berry has Michael Turner #1http://sports.espn.go.com/fantasy/football...fldk2k9lovehate

 

[cobalt_27]

Gratuitous bump, since Matthew Berry and others seem to still believe in the "Curse of 370" nonsense.

For those new to this thread, the fun starts on post 16 and goes from there.

Moral of the story, you can expect Michael Turner to regress to the mean and produce less than last season, but...not based on any so-called curse.

Go.

Matthew Berry has Michael Turner #1http://sports.espn.go.com/fantasy/football...fldk2k9lovehate

Yeah, my bad. ESPNMag has a one-page thing starting with Berry, and it looked like the remaining pages were all written by him. But, it's an assortment of other writers. Tristan Cockroft is the one responsible for communicating this myth.