21 Oct 2008
The subject of this week's Smarter Stats column for the Washington Post won't come as news to faithful FO readers, but with Clinton Portis on pace to enter Full Curse Territory and former victim Shaun Alexander now on the Redskins' roster, it's time to bring it up again. Will they never learn?
8 comments, Last at 30 Oct 2008, 7:36pm by bravehoptoad
The Eagles and Redskins could use some fixes in the secondary, while the Cowboys are in need of some interior offensive line help. And who is the eventual successor to Osi Umenyiora in New York?
Comments
Re: Smarter Stats: The Curse of 370
I wish you had emphasized more the main point that using a running back too much will usually result in decreased performance and not "the curse of 370".
The whole "don't go over 370" thing suffers from selection bias, as explained here: http://www.advancednflstats.com/2008/07/drunkards-light-posts-and-myth-o...
And the thing is, it's been mentioned here as well (I think by Aaron): there is nothing special about 370 carries. A RB will be approximately the same amount of trouble if he rushes for 365,366,367 carries etc.
Re: Smarter Stats: The Curse of 370
Actually, the post you site suggests that 370 is special only because of multiple endpoints, and that choosing 369 or 371 as the cutoff doesn't suggest that regression is expected.
I think FO's point is valid; if a RB puts up big numbers due to overuse, he'll likely take a major step back the following season. However, putting a hard and fast value to the overuse cutoff is not appropriate.
Re: Smarter Stats: The Curse of 370
I'm sure it's been covered, but I'm too lazy to dig it up. If the NFL adds two more games to the regular season, will the number then be 416? Would spreading the (roughly) 370 carries over two more games reduce its impact? Would the number have been 324 in a 14 game season?
Re: Smarter Stats: The Curse of 370
Pro Football Reference has some posts on running back overuse that suggest it's not the season total that puts RBs at risk, but the frequency of high-carry games in a given period.
http://www.pro-football-reference.com/blog/?p=328
http://www.pro-football-reference.com/blog/?p=330
http://www.pro-football-reference.com/blog/?p=483
The PFR theory makes more sense to me. I'm interested to hear the FO writers' thoughts on this since it contradicts the curse of 370. I wonder if this is a correlation vs. causation issue. In other words, is it simply that RBs who go for 370 carries in one year are extremely likely to be overused the next year, and therefore, 370 carries has a high correlation with being injured the following year?
Re: Smarter Stats: The Curse of 370
Sure, but the "Curse of Frequent High-Carry Games in a Given Period" doesn't have the same ring to it.
Re: Smarter Stats: The Curse of 370
I've always winced at FO's hubris on this issue. A few of the posts above mine are excellent reasons to question the foregone conclusion of 'the curse of 370.' It is the besetting sin of statisticians the world over to confuse correlation with causation.
Re: Smarter Stats: The Curse of 370
Staubach12, most of the backs who get 370+ carries don't get overused the next year; they get hurt.
High-carry games doesn't make more sense to me. Of course, what would make sense is knowing how many times and how hard a back was hit, and I don't know of any way to measure that.
There is something that FO (or really, someone with the Internet and some time on their hands) could probably tell us:
How many games do backs with 370+ carries in a season play in over the next two years?
How many games do backs with 100+ carries in a season play in over the next two years?
How many games do backs with 1000+ rushing yards in a season play in over the next two years?
How many games do backs with 600+ rushing yards in a season play in over the next two years?
Running backs are fungible precisely because they burn out quickly. It's not clear to me that the guys with 370+ carries burn out faster, or at the same rate.
And if they do burn out at the same rate or slower, then the coaches should be giving their best backs *more* carries, not less.
Re: Smarter Stats: The Curse of 370
Subrata Sircar, you're missing my point. I'm saying that overuse presents an injury risk over a shorter period of time than a season. Getting 75 carries in three games is still overuse. Backs who got 370 carries in one year dodged a bullet by not getting injured. I'm saying that the odds are they won't dodge the bullet again the following year and are extremely likely to get hurt because they will continue to have a high number of carries per game the following season (which will probably lead to an early to mid season injury).
Post new comment