Projecting First Downs for RBs (1 Viewer)

[burd]

im in a league that awards First Down points. Ive been using Doug's formula for projecting WR first down points and it's worked well. I was wondering if a formula for projecting RB first down points exists. Ive been using the WR formula to project RB first down points but i was wondering if there was a formula for RBs. thanks,

 

[dgreen]

What's the WR formula?

 

[greggorymac]

im in a league that awards First Down points. Ive been using Doug's formula for projecting WR first down points and it's worked well. I was wondering if a formula for projecting RB first down points exists. Ive been using the WR formula to project RB first down points but i was wondering if there was a formula for RBs. thanks,

Interesting... This is the first I've ever heard of this type of scoring structure. Is the purpose to raise the value of RBs that don't stay in on 3rd downs (ala Jamal Lewis, LJ, etc.)?

 

[burd]

im in a league that awards First Down points. Ive been using Doug's formula for projecting WR first down points and it's worked well. I was wondering if a formula for projecting RB first down points exists. Ive been using the WR formula to project RB first down points but i was wondering if there was a formula for RBs. thanks,

Interesting... This is the first I've ever heard of this type of scoring structure. Is the purpose to raise the value of RBs that don't stay in on 3rd downs (ala Jamal Lewis, LJ, etc.)?

yeah, everyone in the league seems to like 1 point for each Rush or Rec First Down (atleast they like it better than 1 point per reception). Im not sure it raises the value of 3rd down RBs enough to warrant keeping them on your team ... Kevin Faulk had a good amount of first downs but his lack of yardage and TDs didn't make him attractive enough to keep for very long. I remember guys like Faulk etc moving from team to team as other owners needed to cover Byes. But after playing in a league that scores first downs, i can say that I definitely like it over the traditional type league. One big difference is that every 3rd down play is a potential point play because the team is obviously shooting for a first down. I think theres a lot more suspense and action because the scores are constantly changing. It also improves the value of "possession receivers" as well as 3rd down backs

 

[burd]

What's the WR formula?

i dont want to give out member info but it's Doug's formula ...

 

[cobalt_27]

If you have a prediction as to the number of yards each RB will rush for, you could plug in this formula and get a pretty reliable estimate of 1st Downs:



1st Downs = {RuYds*.043) + 5.2

Example:

Adrian Peterson 2008: {1760 RuYds * .053} + 5.2 = 80.9 (Actual was 81)

Steve Slaton 2008: {1282 * .043} + 5.2 = 60.3 (Actual was 60)

You should be able to predict about 2/3rds of your RBs within 6 (+/-) using this formula (well, assuming you could accurately predict their yardage).

 

[greggorymac]

If you have a prediction as to the number of yards each RB will rush for, you could plug in this formula and get a pretty reliable estimate of 1st Downs:



1st Downs = {RuYds*.043) + 5.2

Example:

Adrian Peterson 2008: {1760 RuYds * .053} + 5.2 = 80.9 (Actual was 81)

Steve Slaton 2008: {1282 * .043} + 5.2 = 60.3 (Actual was 60)

You should be able to predict about 2/3rds of your RBs within 6 (+/-) using this formula (well, assuming you could accurately predict their yardage).

Couple questions?a) Did you just 2008 data for your linear curve fit?

b) How big was your data set (were all rushers involved or selected RBs)?

 

[Chase Stuart]

If you have a prediction as to the number of yards each RB will rush for, you could plug in this formula and get a pretty reliable estimate of 1st Downs:



1st Downs = {RuYds*.043) + 5.2

Example:

Adrian Peterson 2008: {1760 RuYds * .053} + 5.2 = 80.9 (Actual was 81)

Steve Slaton 2008: {1282 * .043} + 5.2 = 60.3 (Actual was 60)

You should be able to predict about 2/3rds of your RBs within 6 (+/-) using this formula (well, assuming you could accurately predict their yardage).

You could improve on this by using carries, too.I.e., 1200 yards on 300 carries will produce a few more first downs than 1200 yards on 250 carries.

 

[burd]

If you have a prediction as to the number of yards each RB will rush for, you could plug in this formula and get a pretty reliable estimate of 1st Downs:



1st Downs = {RuYds*.043) + 5.2

Example:

Adrian Peterson 2008: {1760 RuYds * .053} + 5.2 = 80.9 (Actual was 81)

Steve Slaton 2008: {1282 * .043} + 5.2 = 60.3 (Actual was 60)

You should be able to predict about 2/3rds of your RBs within 6 (+/-) using this formula (well, assuming you could accurately predict their yardage).

You could improve on this by using carries, too.I.e., 1200 yards on 300 carries will produce a few more first downs than 1200 yards on 250 carries.

its even MORE complicated when you factor in 1st down receptions by RBs. For RBs, roughly 20% of their rushes are for 1st downs & roughly 35% of their receptions are for 1st downs.

so any good RB formula would have to account for both rushes and receptions. Are there any math wizzes out there that can come up w/ something?

 

[Jeff Pasquino]

If you have a prediction as to the number of yards each RB will rush for, you could plug in this formula and get a pretty reliable estimate of 1st Downs:



1st Downs = {RuYds*.043) + 5.2

Example:

Adrian Peterson 2008: {1760 RuYds * .053} + 5.2 = 80.9 (Actual was 81)

Steve Slaton 2008: {1282 * .043} + 5.2 = 60.3 (Actual was 60)

You should be able to predict about 2/3rds of your RBs within 6 (+/-) using this formula (well, assuming you could accurately predict their yardage).

This might work but would completely ignore the short yardage / goal line type RB.This formula seems to want to make a linear fit to the top RBs that produced first downs.

I think the purpose of the OP's question is to how to find RBs that will produce 1sts, and probably at a higher rate than the norm. If I'm in a league like that, that's what I would want to know - who gives me above-average production?

Based on rushing alone, I took that formula and looked at the Top 50 players for rushing first downs in 2008. (Minimum was 24).

Then I flagged ones that were either 10% above or 15% below that "best fit" line.

Player Team Pos Efficiency

T.J. Duckett SEA RB 206%

David Garrard JAC QB 163%

Tim Hightower ARI RB 152%

Pierre Thomas NO RB 134%

Michael Pittman DEN RB 132%

Marion Barber DAL RB 132%

Ryan Fitzpatrick CIN QB 131%

Sammy Morris NE RB 123%

Peyton Hillis DEN RB 120%

Joseph Addai IND RB 119%

Mewelde Moore PIT RB 118%

Steven Jackson STL RB 116%

Le'Ron McClain BAL FB 116%

Maurice Jones-Drew JAC RB 116%

Fred Jackson BUF RB 114%

LenDale White TEN RB 112%

Brandon Jacobs NYG RB 110%

Clinton Portis WAS RB 108%

Brian Westbrook PHI RB 108%

Matt Forte CHI RB 108%

Ronnie Brown MIA RB 108%

Derrick Ward NYG RB 108%

Michael Turner ATL RB 107%

Marshawn Lynch BUF RB 107%

Ricky Williams MIA RB 104%

Kevin Smith DET RB 104%

Larry Johnson KC RB 103%

Thomas Jones NYJ RB 102%

Jonathan Stewart CAR RB 102%

Adrian Peterson MIN RB 100%

Willis McGahee BAL RB 100%

Steve Slaton HOU RB 99%

LaDainian Tomlinson SD RB 98%

Tashard Choice DAL RB 98%

Cedric Benson CIN RB 96%

Dominic Rhodes IND RB 95%

DeAngelo Williams CAR RB 94%

Ryan Grant GB RB 93%

Chris Johnson TEN RB 93%

Maurice Morris SEA RB 90%

Darren McFadden OAK RB 90%

Kevin Faulk NE RB 89%

Frank Gore SF RB 88%

Earnest Graham TB RB 88%

Edgerrin James ARI RB 88%

Jamal Lewis CLE RB 87%

Julius Jones SEA RB 82%

Warrick Dunn TB RB 79%

Justin Fargas OAK RB 76%

Willie Parker PIT RB 74%

As you can see, the Top 17 were over 110%, or 10% better than that formula. Many of those RBs would be considered 3rd down / short yardage guys.

Contrastingly, the players at the bottom often got yanked in those short yardage situations or on third downs on a regular basis.

 

[dgreen]

I think the purpose of the OP's question is to how to find RBs that will produce 1sts, and probably at a higher rate than the norm.

Right. Why would anyone want one formula to predict 1st downs for all RBs? We don't predict yards or TDs that way, so why 1st downs?

 

[Jeff Pasquino]

I think the purpose of the OP's question is to how to find RBs that will produce 1sts, and probably at a higher rate than the norm.

Right. Why would anyone want one formula to predict 1st downs for all RBs? We don't predict yards or TDs that way, so why 1st downs?

The formula is useful as a BASELINE to determine what you should be looking for in a player providing an above-average level of performance.Knowing that the NFL averages 4.0 yards per carry is only useful if you are looking for guys that beat it.

 

[cobalt_27]

If you have a prediction as to the number of yards each RB will rush for, you could plug in this formula and get a pretty reliable estimate of 1st Downs:



1st Downs = {RuYds*.043) + 5.2

Example:

Adrian Peterson 2008: {1760 RuYds * .053} + 5.2 = 80.9 (Actual was 81)

Steve Slaton 2008: {1282 * .043} + 5.2 = 60.3 (Actual was 60)

You should be able to predict about 2/3rds of your RBs within 6 (+/-) using this formula (well, assuming you could accurately predict their yardage).

You could improve on this by using carries, too.I.e., 1200 yards on 300 carries will produce a few more first downs than 1200 yards on 250 carries.

Actually, statistically, carries adds nothing to the equation. Once you factor in yards, the effect of # rushes is wiped out (on regression).

 

[cobalt_27]

If you have a prediction as to the number of yards each RB will rush for, you could plug in this formula and get a pretty reliable estimate of 1st Downs:



1st Downs = {RuYds*.043) + 5.2

Example:

Adrian Peterson 2008: {1760 RuYds * .053} + 5.2 = 80.9 (Actual was 81)

Steve Slaton 2008: {1282 * .043} + 5.2 = 60.3 (Actual was 60)

You should be able to predict about 2/3rds of your RBs within 6 (+/-) using this formula (well, assuming you could accurately predict their yardage).

Couple questions?a) Did you just 2008 data for your linear curve fit?

b) How big was your data set (were all rushers involved or selected RBs)?

I used the top 30 in rushing attempts over the last 3 years. Yes, the assumption is linear.

 

[cobalt_27]

If you have a prediction as to the number of yards each RB will rush for, you could plug in this formula and get a pretty reliable estimate of 1st Downs:



1st Downs = {RuYds*.043) + 5.2

Example:

Adrian Peterson 2008: {1760 RuYds * .053} + 5.2 = 80.9 (Actual was 81)

Steve Slaton 2008: {1282 * .043} + 5.2 = 60.3 (Actual was 60)

You should be able to predict about 2/3rds of your RBs within 6 (+/-) using this formula (well, assuming you could accurately predict their yardage).

This might work but would completely ignore the short yardage / goal line type RB.This formula seems to want to make a linear fit to the top RBs that produced first downs.

I think the purpose of the OP's question is to how to find RBs that will produce 1sts, and probably at a higher rate than the norm. If I'm in a league like that, that's what I would want to know - who gives me above-average production?

Based on rushing alone, I took that formula and looked at the Top 50 players for rushing first downs in 2008. (Minimum was 24).

Then I flagged ones that were either 10% above or 15% below that "best fit" line.

Player Team Pos Efficiency

T.J. Duckett SEA RB 206%

David Garrard JAC QB 163%

Tim Hightower ARI RB 152%

Pierre Thomas NO RB 134%

Michael Pittman DEN RB 132%

Marion Barber DAL RB 132%

Ryan Fitzpatrick CIN QB 131%

Sammy Morris NE RB 123%

Peyton Hillis DEN RB 120%

Joseph Addai IND RB 119%

Mewelde Moore PIT RB 118%

Steven Jackson STL RB 116%

Le'Ron McClain BAL FB 116%

Maurice Jones-Drew JAC RB 116%

Fred Jackson BUF RB 114%

LenDale White TEN RB 112%

Brandon Jacobs NYG RB 110%

Clinton Portis WAS RB 108%

Brian Westbrook PHI RB 108%

Matt Forte CHI RB 108%

Ronnie Brown MIA RB 108%

Derrick Ward NYG RB 108%

Michael Turner ATL RB 107%

Marshawn Lynch BUF RB 107%

Ricky Williams MIA RB 104%

Kevin Smith DET RB 104%

Larry Johnson KC RB 103%

Thomas Jones NYJ RB 102%

Jonathan Stewart CAR RB 102%

Adrian Peterson MIN RB 100%

Willis McGahee BAL RB 100%

Steve Slaton HOU RB 99%

LaDainian Tomlinson SD RB 98%

Tashard Choice DAL RB 98%

Cedric Benson CIN RB 96%

Dominic Rhodes IND RB 95%

DeAngelo Williams CAR RB 94%

Ryan Grant GB RB 93%

Chris Johnson TEN RB 93%

Maurice Morris SEA RB 90%

Darren McFadden OAK RB 90%

Kevin Faulk NE RB 89%

Frank Gore SF RB 88%

Earnest Graham TB RB 88%

Edgerrin James ARI RB 88%

Jamal Lewis CLE RB 87%

Julius Jones SEA RB 82%

Warrick Dunn TB RB 79%

Justin Fargas OAK RB 76%

Willie Parker PIT RB 74%

Contrastingly, the players at the bottom often got yanked in those short yardage situations or on third downs on a regular basis.

This is really good to see. Adds a real nice qualitative interpretation to the data. I think you're absolutely right...add about 1-2 SD on top of the short yardage backs. As you can see, the Top 17 were over 110%, or 10% better than that formula. Many of those RBs would be considered 3rd down / short yardage guys.

 

[Chase Stuart]

If you have a prediction as to the number of yards each RB will rush for, you could plug in this formula and get a pretty reliable estimate of 1st Downs:



1st Downs = {RuYds*.043) + 5.2

Example:

Adrian Peterson 2008: {1760 RuYds * .053} + 5.2 = 80.9 (Actual was 81)

Steve Slaton 2008: {1282 * .043} + 5.2 = 60.3 (Actual was 60)

You should be able to predict about 2/3rds of your RBs within 6 (+/-) using this formula (well, assuming you could accurately predict their yardage).

You could improve on this by using carries, too.I.e., 1200 yards on 300 carries will produce a few more first downs than 1200 yards on 250 carries.

Actually, statistically, carries adds nothing to the equation. Once you factor in yards, the effect of # rushes is wiped out (on regression).

I don't believe that to be true.

 

[cobalt_27]

If you have a prediction as to the number of yards each RB will rush for, you could plug in this formula and get a pretty reliable estimate of 1st Downs:



1st Downs = {RuYds*.043) + 5.2

Example:

Adrian Peterson 2008: {1760 RuYds * .053} + 5.2 = 80.9 (Actual was 81)

Steve Slaton 2008: {1282 * .043} + 5.2 = 60.3 (Actual was 60)

You should be able to predict about 2/3rds of your RBs within 6 (+/-) using this formula (well, assuming you could accurately predict their yardage).

You could improve on this by using carries, too.I.e., 1200 yards on 300 carries will produce a few more first downs than 1200 yards on 250 carries.

Actually, statistically, carries adds nothing to the equation. Once you factor in yards, the effect of # rushes is wiped out (on regression).

I don't believe that to be true.

Well.Would you like me to send you the data output with the regression analysis that shows, when you insert # of rushes into the equation, it is non-significant.

 

[Chase Stuart]

If you have a prediction as to the number of yards each RB will rush for, you could plug in this formula and get a pretty reliable estimate of 1st Downs:



1st Downs = {RuYds*.043) + 5.2

Example:

Adrian Peterson 2008: {1760 RuYds * .053} + 5.2 = 80.9 (Actual was 81)

Steve Slaton 2008: {1282 * .043} + 5.2 = 60.3 (Actual was 60)

You should be able to predict about 2/3rds of your RBs within 6 (+/-) using this formula (well, assuming you could accurately predict their yardage).

You could improve on this by using carries, too.I.e., 1200 yards on 300 carries will produce a few more first downs than 1200 yards on 250 carries.

Actually, statistically, carries adds nothing to the equation. Once you factor in yards, the effect of # rushes is wiped out (on regression).

I don't believe that to be true.

Well.Would you like me to send you the data output with the regression analysis that shows, when you insert # of rushes into the equation, it is non-significant.

When I get home I'll try and look at the issue with my data. I would also try doing this for several years. I believe I have researched this question before and found carries to be significant.

 

[cobalt_27]

When I get home I'll try and look at the issue with my data. I would also try doing this for several years. I believe I have researched this question before and found carries to be significant.

3 years (N=150) should be plenty. Could definitely add more, but I don't suspect it will change much.Also, just so we're clear...when you analyze # of rushes separately, they're significant. But, when you add yardage in the equation, it eats up most of the variance explained by rushes. Plus, as you would expect, # of rushes are so highly correlated with yardage to the extent that you violate the principle of multicollinearity...a no-no in regression analyses. Therefore, you have to drop one of the variables.ETA, I just looked and it was the top 50 in RuAtt from 06-08.

 

[dgreen]

I think the purpose of the OP's question is to how to find RBs that will produce 1sts, and probably at a higher rate than the norm.

Right. Why would anyone want one formula to predict 1st downs for all RBs? We don't predict yards or TDs that way, so why 1st downs?

The formula is useful as a BASELINE to determine what you should be looking for in a player providing an above-average level of performance.Knowing that the NFL averages 4.0 yards per carry is only useful if you are looking for guys that beat it.

Good point. I guess I didn't think of it that way since I consider 4.0 YPC to be common knowledge and I could easily identify that as a baseline without a formula.

 

[Chase Stuart]

When I get home I'll try and look at the issue with my data. I would also try doing this for several years. I believe I have researched this question before and found carries to be significant.

3 years (N=150) should be plenty. Could definitely add more, but I don't suspect it will change much.Also, just so we're clear...when you analyze # of rushes separately, they're significant. But, when you add yardage in the equation, it eats up most of the variance explained by rushes. Plus, as you would expect, # of rushes are so highly correlated with yardage to the extent that you violate the principle of multicollinearity...a no-no in regression analyses. Therefore, you have to drop one of the variables.ETA, I just looked and it was the top 50 in RuAtt from 06-08.

You raise good points; unfortunately there was no getting home for me last night (or tonight it looks like) so I'm going to have to table this to the weekend. Let me be clear on my argument, though: If RB A has 300 carries for 1200 yards, and RB B has 250 carries for 1200 yards, RB A will have (on average) more rushing first downs than RB B.

 

[Chase Stuart]

I looked at the 142 RBs from '02 to '08 with at least 225 carries. I performed a regression using carries and rushing yards as the inputs and rushing first downs as the output. This does not address the legitimate multicollinearity question, but the best fit formula was:

-5.9 + 0.06*CAR + 0.04*RYD.

So 300 carries for 1200 yards would be 60 first downs. 250 carries for 1200 yards would be 57 first downs. Obviously not a significant difference, but FWIW, the p-value on the carries variable was 0.02.

 

[gheemony]

Wow, great stuff. And I couldn't get anyone to answer my question about a formula for performance peak and decline based on age:

 

[cobalt_27]

Chase, I've got your dataset of 142 (2002-2008) RBs all with 225+ CAR.

I still didn't think CAR added anything, so I ran some numbers. From the regression equation you put up there (same as what I got using those data), you can calculate and expected number of 1st Downs for each player and subtract that from the actual value each player got in a given year. Take the absolute value of that number for all 142 players, and the mean deviation from 0 (the ideal) was 5.7 (SD=5.0).

Now, if you run the regression using only yards (my method), you get EXP1 = {.047*YDS} + 1.65. Do the same thing, get an expected number of 1st Downs for each player and subtract that value from the actual number each player got in a given year. The mean deviation from 0 across all 142 players was 5.9 (SD=5.0).

At first glance, I was gong to say this difference isn't likely to be significant. However, when I looked at all 142 backs (post below), it is worth nothing that Chase's method more closely predicted the # of First Downs on 76 of the 142 RBs (my method was closer on 66 of the 142).

So, I'll reverse course here...my feeling is go with Chase and his equation on this one.

 

[cobalt_27]

The insomnia's strong tonight.

Below is a list of all 142 RBs in the dataset with

1. Actual First Downs

2. EXP1 = Expected 1st Downs Using Chase's CAR and YDS equation

3. DIF1 = Absolute value of the difference between Actual and EXP1-->Chase's Method

4. EXP2 = Expected 1st Downs Using only the YDS equation

5. DIF2 = Absolute value of the difference between Actual and EXP2-->My Method

Clearly Edgerrin James and Priest Holmes never got the memo about this (see below...way down the list).

Year PLAYER ATT RUYDS 1DWNs EXP1 DIF(1) EXP2 DIF(2)

2006 Ronnie Brown 241 1008 49 48.22 0.78 49.03 0.03

2007 Clinton Portis 325 1262 61 63.25 2.25 60.97 0.03

2002 Tiki Barber 304 1387 67 66.84 0.16 66.84 0.16

2006 Willis McGahee 259 990 48 48.62 0.62 48.18 0.18

2002 Ricky Williams 383 1853 89 89.84 0.84 88.74 0.26

2008 Thomas Jones 290 1312 63 63.07 0.07 63.32 0.32

2003 Clinton Portis 290 1591 76 73.95 2.05 76.43 0.43

2005 Reuben Droughns 309 1232 60 61.1 1.1 59.56 0.44

2002 Emmitt Smith 254 975 48 47.73 0.27 47.48 0.52

2003 Stephen Davis 318 1444 69 69.92 0.92 69.52 0.52

2006 Steven Jackson 346 1528 74 74.91 0.91 73.47 0.53

2005 LaDainian Tomlinson 339 1462 71 71.9 0.9 70.37 0.63

2007 Thomas Jones 310 1119 55 56.76 1.76 54.25 0.75

2005 Steven Jackson 254 1046 50 50.5 0.5 50.82 0.82

2007 DeShaun Foster 247 876 42 43.44 1.44 42.83 0.83

2004 Deuce McAllister 269 1074 53 52.5 0.5 52.13 0.87

2002 Curtis Martin 261 1094 54 52.79 1.21 53.07 0.93

2002 Jamal Lewis 308 1327 63 64.75 1.75 64.02 1.02

2005 Clinton Portis 352 1516 74 74.8 0.8 72.91 1.1

2006 Brian Westbrook 240 1217 60 56.31 3.69 58.85 1.15

2004 Ahman Green 259 1163 55 55.36 0.36 56.31 1.31

2005 Julius Jones 257 993 47 48.61 1.61 48.32 1.32

2004 Edgerrin James 334 1548 73 74.95 1.95 74.41 1.41

2003 Fred Taylor 345 1572 77 76.56 0.44 75.54 1.46

2008 Kevin Smith 238 976 49 46.79 2.21 47.53 1.48

2007 Warrick Dunn 227 720 34 36.13 2.13 35.49 1.49

2004 Domanick Williams 302 1188 56 58.96 2.96 57.49 1.49

2004 Clinton Portis 343 1315 65 66.42 1.42 63.46 1.54

2006 Tiki Barber 327 1662 78 78.97 0.97 79.77 1.77

2008 LaDainian Tomlinson 292 1110 52 55.31 3.31 53.82 1.82

2003 Ricky Williams 392 1372 68 71.63 3.63 66.14 1.86

2006 Willie Parker 337 1494 70 73.03 3.03 71.87 1.87

2003 Eddie George 312 1031 52 53.45 1.45 50.11 1.89

2002 Deuce McAllister 325 1388 65 68.16 3.16 66.89 1.89

2008 Steve Slaton 268 1282 60 60.55 0.55 61.91 1.91

2004 Jamal Lewis 235 1006 47 47.78 0.78 48.94 1.94

2003 LaDainian Tomlinson 313 1645 81 77.46 3.55 78.97 2.03

2004 Kevin Jones 241 1133 57 53.1 3.91 54.9 2.1

2007 Willis McGahee 294 1207 56 59.21 3.21 58.38 2.38

2006 Thomas Jones 296 1210 61 59.45 1.55 58.52 2.48

2008 Michael Turner 376 1699 84 83.4 0.05 81.51 2.49

2004 Corey Dillon 345 1635 81 79.02 1.98 78.5 2.5

2008 Marshawn Lynch 250 1036 53 49.86 3.14 50.35 2.66

2006 Deuce McAllister 244 1057 54 50.31 3.69 51.33 2.67

2007 Adrian Peterson 238 1341 62 61.02 0.98 64.68 2.68

2002 Travis Henry 325 1438 72 70.11 1.89 69.24 2.76

2004 Thomas Jones 240 948 49 45.82 3.18 46.21 2.79

2006 Edgerrin James 337 1159 59 59.97 0.97 56.13 2.87

2004 Rudi Johnson 361 1454 73 72.93 0.07 69.99 3.01

2005 Cadillac Williams 290 1178 54 57.84 3.84 57.02 3.02

2007 Marshawn Lynch 280 1115 51 54.77 3.77 54.06 3.06

2003 Domanick Williams 238 1031 47 48.93 1.93 50.11 3.11

2006 Ladell Betts 245 1154 59 54.16 4.84 55.89 3.11

2008 Matt Forte 316 1238 63 61.77 1.24 59.84 3.16

2007 LenDale White 303 1110 57 55.98 1.02 53.82 3.18

2005 Warrick Dunn 280 1416 65 66.51 1.51 68.21 3.21

2003 Jerome Bettis 246 811 43 40.84 2.16 39.77 3.23

2008 Brian Westbrook 233 936 49 44.92 4.08 45.65 3.36

2008 Adrian Peterson 363 1760 81 84.99 3.99 84.37 3.37

2005 Thomas Jones 314 1335 61 65.43 4.43 64.4 3.4

2008 Clinton Portis 342 1487 75 73.06 1.94 71.54 3.46

2006 Ahman Green 266 1059 55 51.73 3.27 51.43 3.57

2004 LaDainian Tomlinson 339 1335 68 66.95 1.05 64.4 3.6

2002 James Stewart 231 1021 46 48.12 2.12 49.64 3.64

2003 Marcel Shipp 228 830 37 40.49 3.49 40.66 3.66

2004 Kevan Barlow 244 822 44 41.15 2.85 40.29 3.71

2005 Jamal Lewis 269 906 48 45.95 2.05 44.24 3.77

2003 Anthony Thomas 244 1024 46 49.03 3.03 49.78 3.78

2006 Rudi Johnson 341 1309 67 66.06 0.94 63.18 3.82

2004 Tiki Barber 322 1518 77 73.05 3.95 73 4

2007 LaDainian Tomlinson 315 1474 75 70.91 4.09 70.93 4.07

2002 Shaun Alexander 295 1175 61 58.03 2.97 56.88 4.12

2004 Jerome Bettis 250 941 50 46.16 3.84 45.88 4.12

2004 Shaun Alexander 353 1696 77 81.88 4.88 81.37 4.37

2002 Eddie George 343 1165 61 60.57 0.44 56.41 4.59

2004 Emmitt Smith 267 937 41 47.04 6.04 45.69 4.69

2007 Jamal Lewis 298 1304 58 63.24 5.24 62.94 4.94

2006 DeShaun Foster 227 897 49 43.04 5.96 43.81 5.19

2008 Ryan Grant 312 1203 53 60.16 7.16 58.19 5.19

2006 Larry Johnson 416 1789 91 89.35 1.65 85.74 5.26

2002 William Green 243 887 38 43.62 5.62 43.34 5.34

2008 Chris Johnson 251 1228 54 57.41 3.41 59.37 5.37

2002 Corey Dillon 314 1311 69 64.49 4.51 63.27 5.73

2006 Chester Taylor 303 1216 53 60.11 7.11 58.81 5.81

2003 Ahman Green 355 1883 96 89.3 6.7 90.15 5.85

2002 Ahman Green 286 1240 54 60.01 6.01 59.93 5.93

2006 LaDainian Tomlinson 348 1815 81 86.22 5.22 86.96 5.96

2002 Duce Staley 269 1029 56 50.75 5.25 50.02 5.98

2002 Warrick Dunn 230 927 39 44.39 5.39 45.22 6.22

2008 Frank Gore 240 1036 44 49.25 5.25 50.35 6.35

2002 Clinton Portis 273 1508 79 69.67 9.33 72.53 6.47

2008 Jamal Lewis 279 1002 42 50.3 8.3 48.75 6.75

2006 Tatum Bell 233 1025 43 48.4 5.4 49.83 6.83

2008 DeAngelo Williams 273 1515 66 69.95 3.95 72.86 6.86

2008 Le'Ron McClain 232 902 51 43.54 7.46 44.05 6.95

2007 Edgerrin James 324 1222 52 61.63 9.63 59.09 7.09

2006 Cadillac Williams 225 798 32 39.05 7.05 39.16 7.16

2003 Tiki Barber 278 1216 66 58.59 7.41 58.81 7.2

2006 Warrick Dunn 286 1140 48 56.11 8.11 55.23 7.23

2008 Steven Jackson 253 1042 58 50.28 7.72 50.63 7.37

2004 Warrick Dunn 265 1106 46 53.51 7.51 53.64 7.64

2002 Edgerrin James 277 989 56 49.68 6.33 48.14 7.86

2002 Antowain Smith 252 982 56 47.88 8.12 47.81 8.19

2005 Domanick Williams 230 976 39 46.3 7.3 47.53 8.53

2004 Curtis Martin 371 1697 90 83.02 6.98 81.41 8.59

2007 Brian Westbrook 278 1333 73 63.15 9.85 64.3 8.7

2005 Rudi Johnson 337 1458 79 71.63 7.37 70.18 8.82

2006 Jamal Lewis 314 1132 46 57.51 11.51 54.86 8.86

2004 Reuben Droughns 275 1240 69 59.34 9.66 59.93 9.07

2006 Shaun Alexander 252 896 53 44.52 8.48 43.77 9.24

2002 LaDainian Tomlinson 372 1683 90 82.54 7.46 80.75 9.25

2004 Willis McGahee 284 1128 64 55.52 8.48 54.67 9.33

2006 Fred Taylor 231 1146 46 52.99 6.99 55.52 9.52

2006 Julius Jones 267 1084 43 52.77 9.77 52.6 9.6

2003 Troy Hambrick 275 972 57 48.89 8.11 47.34 9.66

2003 Deuce McAllister 351 1641 69 79.62 10.62 78.78 9.78

2003 Shaun Alexander 326 1435 79 70.06 8.94 69.1 9.9

2003 Curtis Martin 323 1308 53 64.92 11.92 63.13 10.13

2006 Travis Henry 270 1211 48 57.91 9.91 58.57 10.57

2003 Travis Henry 331 1356 76 67.28 8.72 65.39 10.62

2005 Willis McGahee 325 1247 71 62.67 8.34 60.26 10.74

2005 LaMont Jordan 272 1025 61 50.77 10.23 49.83 11.17

2004 Fred Taylor 260 1224 48 57.8 9.8 59.18 11.18

2007 Frank Gore 260 1102 42 53.05 11.05 53.45 11.45

2005 Willie Parker 255 1202 46 56.64 10.64 58.15 12.15

2005 Mike Anderson 239 1014 62 48.33 13.67 49.31 12.69

2007 Steven Jackson 237 1002 36 47.74 11.74 48.75 12.75

2007 Joseph Addai 261 1072 65 51.94 13.06 52.04 12.96

2005 Larry Johnson 336 1750 97 82.95 14.05 83.9 13.1

2006 Frank Gore 312 1695 68 79.34 11.34 81.32 13.32

2006 Joseph Addai 226 1081 66 50.15 15.85 52.46 13.54

2008 Marion Barber 238 885 57 43.24 13.76 43.25 13.75

2007 Willie Parker 321 1316 49 65.11 16.11 63.51 14.51

2002 Michael Bennett 255 1296 47 60.31 13.31 62.57 15.57

2003 Jamal Lewis 387 2066 83 98.39 15.39 98.76 15.76

2002 Fred Taylor 287 1314 47 62.96 15.96 63.41 16.41

2005 Shaun Alexander 370 1880 107 90.1 16.9 90.01 16.99

2005 Tiki Barber 357 1860 72 88.52 16.52 89.07 17.07

2003 Priest Holmes 320 1420 88 69.11 18.89 68.39 19.61

2005 Edgerrin James 360 1506 94 74.9 19.1 72.44 21.57

2003 Edgerrin James 310 1259 83 62.22 20.78 60.83 22.17

2002 Priest Holmes 313 1615 100 76.29 23.72 77.56 22.44

 

[JKL]

Great work, guys. I play in a first down league as well, and this will help.

In light of Jeff Pasquino's posting of the players who over or underperform the projection, I'm wondering if adding a variable that serves as a proxy for "short yardage specialist" would improve the projections from Chase's formula, something like touchdown% (since most short yardage guys are also the goalline guys) or size, mass, etc,, since short yardage guys tend to be bulkier as a group.

 

[Dinsy Ejotuz]

I wonder if adding a weight/size measure into the equation would help?

Or adding 1/0 variables for guys over 220 (225?) and under 210 (215?)?

 

[cobalt_27]

Great work, guys. I play in a first down league as well, and this will help. In light of Jeff Pasquino's posting of the players who over or underperform the projection, I'm wondering if adding a variable that serves as a proxy for "short yardage specialist" would improve the projections from Chase's formula, something like touchdown% (since most short yardage guys are also the goalline guys) or size, mass, etc,, since short yardage guys tend to be bulkier as a group.

So, do you want this to predict for ALL RBs or only those with reasonably good workloads (e.g., 225+ ATT)?I can tell you that, for those who rush over 225 times in a season, the TDPCT variable adds a lot. But, you're then trying to warp one equation to fit for 2 very different types of RBs.