Fantasy Football - Footballguys Forums
This is a sample guest message. Register a free account today to become a member! Once signed in, you'll be able to participate on this site by adding your own topics and posts, as well as connect with other members through your own private inbox!
RB Production After Seasons W/400+ Touches (1 Viewer)
- Thread starter David Yudkin
- Start date
That's some great work, David.
The only RB I'd dispute those conclusions with this year would be Tiki, since the Giants O seems to have morphed into a very high octane version, which might allow him to take less punishment as running lanes are bigger due to the emphasis on the pass & Manning progressing up the learning curve. His age might negate that advantage, though.
The only RB I'd dispute those conclusions with this year would be Tiki, since the Giants O seems to have morphed into a very high octane version, which might allow him to take less punishment as running lanes are bigger due to the emphasis on the pass & Manning progressing up the learning curve. His age might negate that advantage, though.
Last edited by a moderator:
Jason Wood
Zoo York
Really nice work David 
Winning IS Everything
Footballguy
Very good work David! I always figured that this was the case after a 400 touch season, but after seeing all those stats it is quite a common dropoff the following year.
GregR_2
Footballguy
The comment about regression to the mean has me off and doing some thinking. Want to make some comments on it, but think I need to do some more homework first. Can someone remind me, was there a FBG article about it, and if so could someone point it to me as I haven't found it after looking in the archives, googling, and doing forum searches?
SSOG
Moderator
Just to add some numbers, I have a spreadsheet with every RB to get 350+ regular season carries and their average numbers the next season. 47 RBs qualified. Here were the results.
Average in year N:
15.8 games, 374.1 carries, 1622.6 yards, 4.33 ypc, 12.8 TDs, 238.9 fantasy points.
Average in year N+1:
13.1 games, 275.5 carries, 1156.5 yards, 4.11 ypc, 8.6 TDs, 167.5 fantasy points.
Change:
-2.7 games, -98.6 carries, -29.7% yards, -.23 ypc, -29.9% points.
4 RBs scored 10+% more points in year N+1 than in year N.
10 RBs scored the same number of points (+/- 10%) in Year N+1 as in year N.
32 RBs scored 10% or more FEWER points in year N+1 than in year N.
The 37th RB was Ricky Williams, who retired (citing overuse as one of the primary reasons).
Also worth noting: 10 RBs failed to even score 50% as many fantasy points in year N+1 as in year N.
Final thought- If you want to use this data BACKWARDS (i.e. not to predict how RBs with 350+ carries are going to do in year N+1), then here are some numbers for you to look at:
Average age of RB receiving 350+ Carries was 25.5. Average years of experience was 4.3. Only four of the 47 RBs to receive 350+ carries were over the age of 27.
Four RBs received 350+ carries last season. They were Shaun Alexander (370), Edgerrin James (360), Tiki Barber (358), and Clinton Portis (351).
Average in year N:
15.8 games, 374.1 carries, 1622.6 yards, 4.33 ypc, 12.8 TDs, 238.9 fantasy points.
Average in year N+1:
13.1 games, 275.5 carries, 1156.5 yards, 4.11 ypc, 8.6 TDs, 167.5 fantasy points.
Change:
-2.7 games, -98.6 carries, -29.7% yards, -.23 ypc, -29.9% points.
4 RBs scored 10+% more points in year N+1 than in year N.
10 RBs scored the same number of points (+/- 10%) in Year N+1 as in year N.
32 RBs scored 10% or more FEWER points in year N+1 than in year N.
The 37th RB was Ricky Williams, who retired (citing overuse as one of the primary reasons).
Also worth noting: 10 RBs failed to even score 50% as many fantasy points in year N+1 as in year N.
Final thought- If you want to use this data BACKWARDS (i.e. not to predict how RBs with 350+ carries are going to do in year N+1), then here are some numbers for you to look at:
Average age of RB receiving 350+ Carries was 25.5. Average years of experience was 4.3. Only four of the 47 RBs to receive 350+ carries were over the age of 27.
Four RBs received 350+ carries last season. They were Shaun Alexander (370), Edgerrin James (360), Tiki Barber (358), and Clinton Portis (351).
SSOG
Moderator
I'm also planning on doing some more analysis later this offseason looking at numbers from year N-1 to see if there are any sort of trends to help identify the next 350+ carry RBs.
Also, one last thing I forgot to mention...
If you sort out the RBs who had 350+ carries *AND* 5.0+ ypc, the results are catastrophic, with none of them managing to score more than 71% as many points the next season, which doesn't bode well for Alexander and Barber next season. Of course, don't freak out, since the sample size is so small (5 data points) as to be pretty much useless, and since Terrell Davis, Eric Dickerson, and Shaun Alexander all had career years (2000 yards, 2000 yards, 27 TDs) coming off of a season with 350 carries and ypc of 4.64+.
Also, one last thing I forgot to mention...
If you sort out the RBs who had 350+ carries *AND* 5.0+ ypc, the results are catastrophic, with none of them managing to score more than 71% as many points the next season, which doesn't bode well for Alexander and Barber next season. Of course, don't freak out, since the sample size is so small (5 data points) as to be pretty much useless, and since Terrell Davis, Eric Dickerson, and Shaun Alexander all had career years (2000 yards, 2000 yards, 27 TDs) coming off of a season with 350 carries and ypc of 4.64+.
David Yudkin
Footballguy
I believe the article you are looking for was posted on profootballreference, but I don't think the articles are still up anymore. You might want to check with Doug.
Jason Wood
Zoo York
SSOG...I know I've looked at those very same trends for varying amounts of carries per season (looking at the resultant impact in Year N+1). Regression to the mean plays a huge role in that type of analysis, particularly the 5+ yards per rush component.One quirky component of this kind of analysis however is that RBs who go over 2,000 yards from scrimmage are virtual locks to be fantasy studs again in Year N+1 barring injury.
EBF
Footballguy
This is something that I've been looking at for a few years. Last year I told people that Curtis Martin and Corey Dillon were major injury risks based on age and the previous year's stats. This year, I think Tiki Barber, Shaun Alexander, and Edgerrin James are the most likely to disappoint. Tomlinson is also a good candidate, but I think he has above average durability.
Couch Potato
Footballguy
GregR_2
Footballguy
Definite on the article as a whole. One bit of feedback though, I think you could make the article even more useful if you provided the missing context in which to frame your results.
We do need to know what happened to players at that level, which you've provided. From what you did we can see how many of them were then worthy of a top five pick, a top 10 pick, a first round pick, got X carries, suffered a major injury, or otherwise insert your favorite criteria.
But the thing that is largely missing is the context in which to put that information. Are those results good or bad? Without knowing how others fared, we can't really say for certain.
For instance, 6 of the 20 RBs (30%) who had that many carries ended up as top 10 RBs the next year. That's part of the puzzle, but without knowing whether that is good or bad compared to other RBs, one can't put it in its proper perspective. If RBs with between 200 and 300 carries reach the top ten 40% of the time, it would imply to me that 400+ carry seasons might indeed be a detriment vs a lighter workload, and all things equal you might want to go with the guy who had less work the year before, but otherwise you have tiered the same. But if those others only reach the top ten 20% of the time, then it would indicate that 400+ carries the previous year is a good thing, not a bad thing, for picking a top ten back.
It isn't missing everywhere. You gave it in the case of average number of games missed the following season, for example. And I think that extra bit of context makes the conclusions drawn from it much more useful and reliable.
Definite on the article as a whole though. 
on the article as a whole. One bit of feedback though, I think you could make the article even more useful if you provided the missing context in which to frame your results.We do need to know what happened to players at that level, which you've provided. From what you did we can see how many of them were then worthy of a top five pick, a top 10 pick, a first round pick, got X carries, suffered a major injury, or otherwise insert your favorite criteria.
But the thing that is largely missing is the context in which to put that information. Are those results good or bad? Without knowing how others fared, we can't really say for certain.
For instance, 6 of the 20 RBs (30%) who had that many carries ended up as top 10 RBs the next year. That's part of the puzzle, but without knowing whether that is good or bad compared to other RBs, one can't put it in its proper perspective. If RBs with between 200 and 300 carries reach the top ten 40% of the time, it would imply to me that 400+ carry seasons might indeed be a detriment vs a lighter workload, and all things equal you might want to go with the guy who had less work the year before, but otherwise you have tiered the same. But if those others only reach the top ten 20% of the time, then it would indicate that 400+ carries the previous year is a good thing, not a bad thing, for picking a top ten back.
It isn't missing everywhere. You gave it in the case of average number of games missed the following season, for example. And I think that extra bit of context makes the conclusions drawn from it much more useful and reliable.
Definite
on the article as a whole though.
Last edited by a moderator:
GregR_2
Footballguy
Hrm, didn't find it, nor in searching the blog.Using google's cashe I did find a 2004 pre-season link list that had a thread on it, but I guess we lost the thread during the board cleanup. Can a staffer answer is it gone for good or would it be possible to retrieve that still? The link was: http://forums.footballguys.com/forum/index...howtopic=153089
SSOG
Moderator
350+ carry RBs are almost always top-10 picks in year N+1... because they got 350 carries the year before. That's a lot of production.For what it's worth, my analysis last year convinced me to stay far away from Curtis Martin and Corey Dillon and to be very wary about Shaun Alexander and Rudi Johnson. As you can see, the results were hit or miss. I avoided the Curtis Martin trainwreck, but Dillon managed to hold some fantasy value thanks to all the short TD plunges. I also wound up trading away Alexander based on the results, although I wasn't too beat up over it since I got Holmes/Johnson in return.
David Yudkin
Footballguy
Should anyone be interested, here are all the players and where they ranked from year to year . . .
Emmitt Smith 1994, 1, 1
Emmitt Smith 1993, 1, 1
Emmitt Smith 1992, 1, 1
Edgerrin James 1999, 1, 2
Earl Campbell 1979, 1, 2
Eric Dickerson 1986, 1, 3
Emmitt Smith 1995, 1, 6
Barry Sanders 1991, 1, 7
Eric Dickerson 1988, 1, 9
Eric Dickerson 1984, 1, 10
Priest Holmes 2003, 1, 12
Marshall Faulk 2001, 1, 14
Marcus Allen 1985, 1, 17
Terrell Davis 1998, 1, 77
Shaun Alexander 2005, 1, ?
Terrell Davis 1997, 2, 1
Thurman Thomas 1991, 2, 2
Terrell Davis 1996, 2, 2
Thurman Thomas 1992, 2, 3
Walter Payton 1979, 2, 4
James Wilder 1984, 2, 7
Ricky Williams 2002, 2, 9
Roger Craig 1988, 2, 10
Earl Campbell 1980, 2, 12
Ahman Green 2003, 2, 13
Edgerrin James 2000, 2, 32
Jamal Anderson 1998, 2, 119
Emmitt Smith 1991, 3, 1
Marshall Faulk 1998, 3, 2
LaDainian Tomlinson 2004, 3, 3
LaDainian Tomlinson 2003, 3, 3
LaDainian Tomlinson 2002, 3, 3
Eddie George 1999, 3, 3
Thurman Thomas 1993, 3, 7
John Riggins 1983, 3, 7
Rickey Watters 1996, 3, 9
Wilbert Montgomery 1979, 3, 11
Herschel Walker 1988, 3, 11
Ahman Green 2001, 3, 13
Barry Foster 1992, 3, 17
Eddie George 2000, 3, 19
Dorsey Levens 1997, 3, 51
Walter Payton 1984, 4, 4
Walter Payton 1985, 4, 5
Curtis Martin 1996, 4, 14
Jerome Bettis 1997, 4, 20
Jamal Lewis 2003, 4, 25
Curtis Martin 2004, 4, 29
Curt Warner 1983, 4, 131
Tiki Barber 2005, 4, ?
Rickey Watters 1995, 5, 3
Gerald Riggs 1985, 5, 8
Curtis Martin 2001, 5, 18
Natrone Means 1994, 5, 31
Edgerrin James 2005, 5, ?
Edgerrin James 2004, 6, 5
Emmitt Smith 1996, 6, 18
Clinton Portis 2005, 6 , ?
Curtis Martin 1998, 7, 8
Tiki Barber 2002, 7, 15
Corey Dillon 2004, 7, 16
Deuce McAllister 2003, 7, 17
James Wilder 1985, 7, 28
Tony Dorsett 1981, 8, 9
Curtis Martin 1999, 8, 10
Edgar Bennett 1995, 9, 24
Ricky Williams 2003, 9, DNP
Edgerrin James 2003, 10, 6
Jamal Lewis 2000, 16,147
Emmitt Smith 1994, 1, 1
Emmitt Smith 1993, 1, 1
Emmitt Smith 1992, 1, 1
Edgerrin James 1999, 1, 2
Earl Campbell 1979, 1, 2
Eric Dickerson 1986, 1, 3
Emmitt Smith 1995, 1, 6
Barry Sanders 1991, 1, 7
Eric Dickerson 1988, 1, 9
Eric Dickerson 1984, 1, 10
Priest Holmes 2003, 1, 12
Marshall Faulk 2001, 1, 14
Marcus Allen 1985, 1, 17
Terrell Davis 1998, 1, 77
Shaun Alexander 2005, 1, ?
Terrell Davis 1997, 2, 1
Thurman Thomas 1991, 2, 2
Terrell Davis 1996, 2, 2
Thurman Thomas 1992, 2, 3
Walter Payton 1979, 2, 4
James Wilder 1984, 2, 7
Ricky Williams 2002, 2, 9
Roger Craig 1988, 2, 10
Earl Campbell 1980, 2, 12
Ahman Green 2003, 2, 13
Edgerrin James 2000, 2, 32
Jamal Anderson 1998, 2, 119
Emmitt Smith 1991, 3, 1
Marshall Faulk 1998, 3, 2
LaDainian Tomlinson 2004, 3, 3
LaDainian Tomlinson 2003, 3, 3
LaDainian Tomlinson 2002, 3, 3
Eddie George 1999, 3, 3
Thurman Thomas 1993, 3, 7
John Riggins 1983, 3, 7
Rickey Watters 1996, 3, 9
Wilbert Montgomery 1979, 3, 11
Herschel Walker 1988, 3, 11
Ahman Green 2001, 3, 13
Barry Foster 1992, 3, 17
Eddie George 2000, 3, 19
Dorsey Levens 1997, 3, 51
Walter Payton 1984, 4, 4
Walter Payton 1985, 4, 5
Curtis Martin 1996, 4, 14
Jerome Bettis 1997, 4, 20
Jamal Lewis 2003, 4, 25
Curtis Martin 2004, 4, 29
Curt Warner 1983, 4, 131
Tiki Barber 2005, 4, ?
Rickey Watters 1995, 5, 3
Gerald Riggs 1985, 5, 8
Curtis Martin 2001, 5, 18
Natrone Means 1994, 5, 31
Edgerrin James 2005, 5, ?
Edgerrin James 2004, 6, 5
Emmitt Smith 1996, 6, 18
Clinton Portis 2005, 6 , ?
Curtis Martin 1998, 7, 8
Tiki Barber 2002, 7, 15
Corey Dillon 2004, 7, 16
Deuce McAllister 2003, 7, 17
James Wilder 1985, 7, 28
Tony Dorsett 1981, 8, 9
Curtis Martin 1999, 8, 10
Edgar Bennett 1995, 9, 24
Ricky Williams 2003, 9, DNP
Edgerrin James 2003, 10, 6
Jamal Lewis 2000, 16,147
David Yudkin
Footballguy
I guess this one is debatable. Only 25% matched or exceeded their ranking in the following year, and if we set the baseline for where they got drafted in Year X + 1 as where they ranked in Year X, then that's discouraging. Three quarters of the time, the RB in question did not earn back what they cost. To your point earlier, I am not sure how this compares to other early first round picks. If RB WITHOUT 400 carries have a 90% "not equal or greater ROI" rate, then this could be a positive. I don't know right now, nor do I have the luixury to explore it right now (although I have started looking into these things in general).IMO, anytime you get Top 10 production out of a first round RB you really shouldn't complain, and that happened 56% of the time (albeit on several occasions at a decent production dropoff from one year to the next).It's that other 44% that is troubling, as there is a collection of big injuries, nagging injuries, and reduced production in there that is where the risk stems from.
Team Legacy
Footballguy
This is why he's a staff member. Excellent stuff!
Jason Wood
Zoo York
Actually he's a staff member because he gets us all discounts at Best Buy; but the fact he's a good writer was an added bonus.
CA_7
Footballguy
Have started to hear more about Brandon Jacobs getting more carries this year.Pat Kirwan just talked about it on Sirius and found this:
http://www.sportingnews.com/yourturn/viewtopic.php?t=96095
GregR_2
Footballguy
I don't think that's a realistic way to look at things. "Outperforming his previous year" shouldn't be anyone's measure of success (nor is it yours as you mention later on). It's easy to get into a "I have the 3rd pick, so if I don't end up with at least the 3rd most valuable player, the pick was wasted" mentality. And I think that's a mistake, because it isn't a realistic view of what the situation really is. There is way too much uncertainty in this hobby for any given pick to be viewed that way as you're slotting every single player.
Realistically the measure of success of a pick was did the player contribute at a level, relative to the round taken, that will allow our overall team to succeed. And as you say below, as long as you get top 10 production out of a first round RB, you really shouldn't have anything to complain about.
The injuries don't bother me at all, because that's expected at anywhere along the positional curve. The reduced production, and why it happened, is more of what gets at me.
Correct. The third pick, on average, probably finishes as something like the 9th best fantasy player overall. So if you draft some guy at #3 and he finishes #6, you done good. The player exceeded rational expectations.This is related to the whole AVT debate.
Jason Wood
Zoo York
I recall looking at RB carries in a similar light a few years ago and thanks to trustee Google caching, found this post from a few years back...
http://www.draftsheets.com/cgi-bin/nph-pro.../stategist9.cfm
http://www.draftsheets.com/cgi-bin/nph-pro.../stategist9.cfm
Jeff Pasquino
Footballguy
Actually he's a staff member because he gets us all discounts at Best Buy; but the fact he's a good writer was an added bonus.
DirecTV for everyone!Turbo Punch
Footballguy
David,
First off - very good article.
I was wondering what affect age in year N+1 had, and how much these numbers were affected by serious injury in year N+1. Using your sample of 19 seasons (excluding Williams 2003) since 2000 I found that only 2 of these 19 RB had injuries that cost them at least half the following season – Holmes 2003 and Edge 2000.
Here is a table in descending order of age in year N+1. Generally it looks like the younger RBs were able to roughly maintain their PPG even after a 400 touch season but the older RBs had trouble maintaining their PPG. Except for Holmes in 2003 the PPG takes a serious hit for RBs over 27 or 28 in year N+1. Also Holmes did miss half the season the following year. This season Tiki will be 31, Alexander 29, Edge 28 and Portis 25.
Games Games PPG PPG % Age
Year N Year N+1 yr N yr N +1 change N+1
Martin 04 16 12 17.4 9.6 -45% 32
Dillon 04 15 12 16.8 14.1 -16% 31
Priest 03 16 8 23.3 24.8 6% 31
Martin 01 16 16 15.5 11.8 -24% 29
Faulk 01 14 14 24.4 14.9 -39% 29
Tiki 02 16 16 16.5 11.6 -30% 28
George 00 16 16 18.3 9.5 -48% 28
Edge 04 16 15 16.1 17.9 11% 27
A Green 03 16 15 21.6 13.1 -39% 27
LT 04 15 16 19.2 19.9 4% 26
Deuce 03 16 14 16.5 13.1 -21% 26
Edge 03 13 16 17 16.1 -5% 26
R Williams 02 16 16 20.3 14.5 -29% 26
LT 03 16 16 21.5 18 -16% 25
Lewis 03 16 12 19.4 12.8 -34% 25
A Green 01 16 14 16.5 15.5 -6% 25
LT 02 16 16 19.2 21.5 12% 24
Edge 00 16 6 21.1 17.3 -18% 23
Lewis 00 16 16 12.6 13.7 9% 22
First off - very good article.
I was wondering what affect age in year N+1 had, and how much these numbers were affected by serious injury in year N+1. Using your sample of 19 seasons (excluding Williams 2003) since 2000 I found that only 2 of these 19 RB had injuries that cost them at least half the following season – Holmes 2003 and Edge 2000.
Here is a table in descending order of age in year N+1. Generally it looks like the younger RBs were able to roughly maintain their PPG even after a 400 touch season but the older RBs had trouble maintaining their PPG. Except for Holmes in 2003 the PPG takes a serious hit for RBs over 27 or 28 in year N+1. Also Holmes did miss half the season the following year. This season Tiki will be 31, Alexander 29, Edge 28 and Portis 25.
Games Games PPG PPG % Age
Year N Year N+1 yr N yr N +1 change N+1
Martin 04 16 12 17.4 9.6 -45% 32
Dillon 04 15 12 16.8 14.1 -16% 31
Priest 03 16 8 23.3 24.8 6% 31
Martin 01 16 16 15.5 11.8 -24% 29
Faulk 01 14 14 24.4 14.9 -39% 29
Tiki 02 16 16 16.5 11.6 -30% 28
George 00 16 16 18.3 9.5 -48% 28
Edge 04 16 15 16.1 17.9 11% 27
A Green 03 16 15 21.6 13.1 -39% 27
LT 04 15 16 19.2 19.9 4% 26
Deuce 03 16 14 16.5 13.1 -21% 26
Edge 03 13 16 17 16.1 -5% 26
R Williams 02 16 16 20.3 14.5 -29% 26
LT 03 16 16 21.5 18 -16% 25
Lewis 03 16 12 19.4 12.8 -34% 25
A Green 01 16 14 16.5 15.5 -6% 25
LT 02 16 16 19.2 21.5 12% 24
Edge 00 16 6 21.1 17.3 -18% 23
Lewis 00 16 16 12.6 13.7 9% 22
GregR_2
Footballguy
That's probably what I'd do. Use ADP data to chose the player at each spot each year, and then see how they end up in the scoring system the ADP is from.You could do with the results from your individual leagues too. It wouldn't have the possible fringe picks removed by doing that like an ADP result would, but it would tell you how much the third pick had actually been worth in your particular league.
Yes. I've done it before (using about four years' worth of data several years ago), but I'm not sure I'd be able to find it if I looked. Maybe I could.
Actually, I did it about a month ago as well, although for a different reason (so I didn't think of it right away), using the last three years of data from the Great White league in No Mercy. It's not in a form right now that I can get the answer from right away, but let me play around with it for a few minutes tomorrow (or maybe Wednesday).
BTW, when I use just a few years' worth of data, the results are obviously kind of all over the place. So after averaging the years together, I look for the best-fitting function that describes the results, so things are smoothed out a bit. I think that's the best way to do it (other than just using more years' worth of data, which isn't always practical).
Last edited by a moderator:
KnowledgeReignsSupreme
Footballguy
Year N:23.6 carries per game.
Year N+1:
21 carries per game.
Looks like a normal RB injury rate (missing 2.7 games in year N+1) with a regress to mean in workload.
IOW, don't expect everything to go right (no injuries, tops in the league workload) two years in a row.
Footballhead
WWPRD?
The problem with regression to the mean is that it tells you everything that goes up must come down.....but what it doesn't tell you is who will stay up for a bit longer.
Cajun Cannon
Footballguy
Has anyone analyzed Year N-1?
At the end of the article, it seems that DY offers a set up for a regression analysis to weed out correlating factors and try to find a solid causation for predicting N+1 performance.
However, has anyone tried any sort of analysis in year N-1 to "predict" a forthcoming 400+ touch season?
IMHO, I'd rather pick the upcoming 400 touch RB than avoid the drop off to 300 touch RB.
At the end of the article, it seems that DY offers a set up for a regression analysis to weed out correlating factors and try to find a solid causation for predicting N+1 performance.
However, has anyone tried any sort of analysis in year N-1 to "predict" a forthcoming 400+ touch season?
IMHO, I'd rather pick the upcoming 400 touch RB than avoid the drop off to 300 touch RB.
