Towards an Objective Measure of Talent in RBs (1 Viewer)

[Dinsy Ejotuz]

Did you see where I've actually got Chris Henry rated now?

Fair enough. We'll see how things go from here.

 

[Dinsy Ejotuz]

Can you explain (to an idiot like me) what the number in the table mean? You mention Andre Bown as probably being the BEST overall RB pick for an NFL team with a score of 1.32 and yet you go on to say that Lesean McCoy will disappoint after his first season and yet he's got a score of 1.34, so what do the number mean? What's a "good score"?

Heya Boone... if you read back a few posts you'll see that I grouped the ratings with some subjective terms.The parentheses mean (negative)... so McCoy is minus 1.34.

 

[EBF]

EBF, wdcrob, I remember going round and round about BMI last year. What new thoughts or theories are you trying for this year's evaluation? How are you adjusting etc?

I still think BMI is relevant. If you look at the players who have achieved sustained success at the RB position, almost all of them have high BMI scores. I don't think that's a coincidence. I'm always going to favor stocky backs over slender backs. That said, I think the main thing I learned from last year's group is that any talented player has a chance to succeed if he's given an opportunity. I'm not convinced that Slaton, Forte, and Johnson are the most talented backs from their draft class, but they all had the good luck to land in situations where they were the most talented back on the roster. This allowed them to rack up some nice production and increase their FF values significantly. Could Felix Jones, Jonathan Stewart, and Rashard Mendenhall have accomplished the same thing if they had landed in plum spots without a talented incumbent? I think so. The main difference between my approach last year and my approach this year is that I'm keeping more of an open mind. You always hope to find a perfect prospect like LaDainian Tomlinson or Jonathan Stewart, but those guys are rare. In most draft classes there will only be 1-2 guys who have everything you look for. You can't dismiss everyone else just because they have a flaw or two. If they land in a good situation, you have to consider the possibility that they'll succeed. I don't particularly like Chris Wells or Knowshon Moreno, but if they land in good spots then they can produce immediately. It doesn't matter that they have some flaws. If they get an opportunity, they're talented enough to excel in the short term.

 

[EBF]

EBF... did you actually look where I've got Henry rated? And I'll take your bet on Green. What are the stakes?

Brown is moving up my board. He's Mr. Popular on the visit circuit and his size/speed combo will get him picked earlier than where the public projects him. I think he has a chance to be successful, so I'm not really interested in betting against him.However, I'll bet that Greene will be a successful NFL RB. My guess is that two variables are keeping him down in your system: his advanced age and his slightly less than ideal 40 time. Those things don't scare me because his talent is apparent on the football field. I'm confident that he can be a successful grinder in the mold of Rudi Johnson or Michael Turner. The problem with an "objective" system is that it doesn't have the ability to draw distinctions between players who look similar on paper. Reggie Wayne and Reggie Brown might look similar on paper. On gameday? Not so much. If there's one trend I've noticed with your system it's that it consistently overrates workout warriors who were mediocre college players (Chris Henry, Brandon Jackson, Andre Brown). Not all players with X size and Y speed are equal.

 

[madd futher]

Just an early guess:There are probably only 4 backs that fit wdcrob's model this year for 'premium' backs, and two others who might hit if they are selected in roughly the top 4 rounds:'Big back' category (2): Premium back - Wells .........Marginal back - Andre Brown'Small back category (4): Premium Backs - McCoy (assuming you use his normal weight - not his 'flu' weight at the combine); Moreno, Donald Brown; Marginal back: PeermanI'm guessing that Shon Greene and Jennings are both misses, as will be every other 'small back'.

And since at least half the people here think I'm nuts, I'll go ahead and post the ratings for the top eight or ten 2009 RBs:

Andre Brown	 1.32 Knowshon Moreno	 1.21 Chris Wells	 1.09 Cedric Pearman	 1.01 Donald Brown	 0.16 Shonn Greene	 (1.20)Lesean McCoy	 (1.34)Javon Ringer	 (1.38)Rashad Jennings	 (1.99)

I think the folks taking McCoy, Green, Jennings and Ringer are going to be very disappointed after the first year or two. And Andre Brown is the best value in the draft IMO. There are a couple later-drafted guys that look like they can play to me, but like I said above I'm not sure how to account for players taken in the back half of the draft yet. It may not matter, but it might.And I'm very confident of all these ratings but Pearmen - there just aren't very many guys like him and he's a bit of an outlier. And weight changes won't matter much for anyone except Brown. He'd be helped a lot by gaining a few pounds instead of losing them.

I put enough stock in your theories that I went out and bought Pro Football Prospectus 2008 to learn more about your WR "4 Square" theory.My prediction on RBs (and I also only looked at RBs I expect to be drafted in the 1st 4 rounds) were completely on target except for LeSean McCoy. Obviously there is an element I am missing here. Would you shed a little more light?

 

[Chase Stuart]

But let's get to the heart of the matter -- why don't NFL GMs factor in the wdcrob factors into the decision making when they draft? Why won't NFL GMs, surely knowledgeable of the data you have examined, take Andre Brown early in the draft?

Because the NFL draft is horribly inefficient and a significant number of teams don't know what they're doing.

What's the basis for this remark? What makes it inefficient? Why do you think the market, which has people who have more time than you, a bigger interest than you, and access to much more information than you, along with the benefit of the wisdom of crowds, is worse at this than you? Is there any evidence? Because I've seen lots of evidence to the contrary.

 

[Chase Stuart]

I know a lot of people don't like my answer to Chase above, and don't believe it's possible to outpredict the draft.I'm not interested in a pissing match about it - that's why I've posted my predictions to watch. I'll live with those from here on in... any additional tinkering would be minor since there's not much else I could include.

It's not a pissing match; the theory is either statistically supported or it isn't. Judging whether your theory will work going forward will take at least 5 years to figure out. Why wait that long when we can figure out now whether it's legitimate or not?

 

[Dinsy Ejotuz]

Chase, why was baseball inefficient until Bill James?

 

[Dinsy Ejotuz]

What's the basis for this remark? What makes it inefficient? Why do you think the market, which has people who have more time than you, a bigger interest than you, and access to much more information than you, along with the benefit of the wisdom of crowds, is worse at this than you? Is there any evidence?

Because there are conceptually simple, easily-calculated metrics that when combined provide explanations for busts like Benson, Maroney and Perry and those reasons are perfect mirror images of the reasons later round guys like Ahman Green, Dom Davis and Steve Slaton are successful.And if the market was efficient no team would have touched Perry or Benson until well after the 1st round.

 

[EBF]

You keep trying to draw parallels between your system and what people like Bill James did in baseball, but if I'm not mistaken all of that Moneyball stuff was based on performance and not on physical traits. In some respects, you're actually doing the opposite by trying to suggest that a football player is defined by the sum of his physical parts.

 

[Dinsy Ejotuz]

You keep trying to draw parallels between your system and what people like Bill James did in baseball, but if I'm not mistaken all of that Moneyball stuff was based on performance and not on physical traits. In some respects, you're actually doing the opposite by trying to suggest that a football player is defined by the sum of his physical parts.

Who said there's nothing including performance in there? Just you as far as I can tell.I'm not sure how saying this the 10th time will help more than the 9th, 8th, etc... but what I'm doing now is almost unrelated to where I started.

 

[EBF]

You keep trying to draw parallels between your system and what people like Bill James did in baseball, but if I'm not mistaken all of that Moneyball stuff was based on performance and not on physical traits. In some respects, you're actually doing the opposite by trying to suggest that a football player is defined by the sum of his physical parts.

Who said there's nothing including performance in there? Just you as far as I can tell.I'm not sure how saying this the 10th time will help more than the 9th, 8th, etc... but what I'm doing now is almost unrelated to where I started.

It might help if you explained what you're doing. Every season you come out with a new mysterious method that produces a new list of players to target. When it fails you come back the next season with another "improved" mysterious method that produces another new list of players to target. I actually appreciate the spirit behind what you're doing, but it's kind of hard for people to get on board with this when all you do is show us a list of names without explaining where those predictions are coming from and why we're supposed to believe them.

 

[Dinsy Ejotuz]

I actually appreciate the spirit behind what you're doing, but it's kind of hard for people to get on board with this when all you do is show us a list of names without explaining where those predictions are coming from and why we're supposed to believe them.

Thanks EBF - I do understand that. I seriously thought about not making the predictions public since I'm not going to explain how I got these results. Maybe I shouldn't have - I know I can't convince anyone, or fight with the entire Shark Pool elite. And I realize that having done this last year with less than half the final work done makes it hard to accept I may have moved the ball forward now. So take these with as big a grain of salt as you like - I'm making these predictions public only so there's a track record. If people want to say that being wrong previously kills the value of them, fair enough. Nothing I can say to that. But the predictions are out there.In addition... I've never tossed everything into SAS or the like and just seen what came out. Most of the effort was in looking at the early-drafted busts and the late-drafted successes and tried to reverse engineer why they were bad or good. In Excel. Eyeballing the data and moving it around by hand. Some of the patterns I found proved to be overfitting - but some of them weren't, and across two years I eventually whittled away the stuff I thought was junk and kept looking for a way to make a single simplified process for all backs. I think I eventually got to something very useful. Time will tell.

 

[Chase Stuart]

Chase, why was baseball inefficient until Bill James?

I have no idea. I'm not a baseball expert.

What's the basis for this remark? What makes it inefficient? Why do you think the market, which has people who have more time than you, a bigger interest than you, and access to much more information than you, along with the benefit of the wisdom of crowds, is worse at this than you? Is there any evidence?

Because there are conceptually simple, easily-calculated metrics that when combined provide explanations for busts like Benson, Maroney and Perry and those reasons are perfect mirror images of the reasons later round guys like Ahman Green, Dom Davis and Steve Slaton are successful.And if the market was efficient no team would have touched Perry or Benson until well after the 1st round.

Do you really think it's arguable that Benson wasn't a first round talent? He was a top five pick with top notch pedigree. Why exactly shouldn't he have been a first round pick?If these concepts are simple and easy to calculate, why aren't teams using them?

 

[Chase Stuart]

I actually appreciate the spirit behind what you're doing, but it's kind of hard for people to get on board with this when all you do is show us a list of names without explaining where those predictions are coming from and why we're supposed to believe them.

Thanks EBF - I do understand that. I seriously thought about not making the predictions public since I'm not going to explain how I got these results. Maybe I shouldn't have - I know I can't convince anyone, or fight with the entire Shark Pool elite. And I realize that having done this last year with less than half the final work done makes it hard to accept I may have moved the ball forward now. So take these with as big a grain of salt as you like - I'm making these predictions public only so there's a track record. If people want to say that being wrong previously kills the value of them, fair enough. Nothing I can say to that. But the predictions are out there.

In addition... I've never tossed everything into SAS or the like and just seen what came out. Most of the effort was in looking at the early-drafted busts and the late-drafted successes and tried to reverse engineer why they were bad or good. In Excel. Eyeballing the data and moving it around by hand.

Some of the patterns I found proved to be overfitting - but some of them weren't, and across two years I eventually whittled away the stuff I thought was junk and kept looking for a way to make a single simplified process for all backs. I think I eventually got to something very useful. Time will tell.

That's about all I need to know.

 

[az_prof]

Time to start an '09 thread for the '09 analysis?

Need to know for '05-'08 has passed on by.

His system is partly based on draft position so you can't do an analysis until after the NFL draft.

 

[Dinsy Ejotuz]

I'm not really trying to convince you Chase. The predictions are there, and we'll see how it goes.

 

[Chase Stuart]

I'm not really trying to convince you Chase. The predictions are there, and we'll see how it goes.

Predictions based on no substance, are meaningless. We could have any random person make eight different user names, and pick three guys he thinks are really over or underrated. At the end of the year, one of his user names would look like a genius, but that wouldn't make him or his system a good predictor of future success.

 

[Dinsy Ejotuz]

Predictions based on no substance, are meaningless. We could have any random person make eight different user names, and pick three guys he thinks are really over or underrated. At the end of the year, one of his user names would look like a genius, but that wouldn't make him or his system a good predictor of future success.

 

[BuckeyeArt]

I'm not really trying to convince you Chase. The predictions are there, and we'll see how it goes.

Predictions based on no substance, are meaningless. We could have any random person make eight different user names, and pick three guys he thinks are really over or underrated. At the end of the year, one of his user names would look like a genius, but that wouldn't make him or his system a good predictor of future success.

Why would you say that? Just because he hasn't told us what the 'substance' is doesn't mean that it doesn't exist. He has already said that he tested on the early years but it's easy to see that there just isn't enough data available to fully test it the way you would want it tested. You act like his predictions are completely random and unless you think he is lying, they clearly are not.I'd much rather see someone try to find something based on something objective rather than listen to the endless number of 'scouts' on the board who seem to rely on the 'eyeball test'. You dismiss his his approach because "GM's would already factor it in" yet seem to accept the hundreds of posters who use the 'eyeball test'. Why would their views be any better? Don't you think the GM's have already seen that as well?

 

[Tick]

If these concepts are simple and easy to calculate, why aren't teams using them?

Chase, you know this is dangerous thinking. Breakthroughs are made all the time. Of course it's more likely that they'll come from the paid staff of an NFL team, but there are a ton of smart people out there looking carefully at the draft and trying to figure things out. You can dismiss them easily and know that you'll be right 99% of the time, but it's going to happen... someone not from the establishment is going to figure something out that will give them better predictions in some area than NFL teams have.Football Outsiders, Baseball Prospectus, the blossoming statistical revolution in basketball... it's easier with individual sports (which includes baseball), but useful ideas are coming from outside the pro leagues.

It'll be interesting to see whether wdcrob's predictions bear out, and it'll be a lot more interesting if he sticks with one formula for five years (having shared the formula with at least one other reliable person in advance) and shows a consistently better prediction rate than NFL draft position.

 

[Dinsy Ejotuz]

It'll be interesting to see whether wdcrob's predictions bear out, and it'll be a lot more interesting if he sticks with one formula for five years (having shared the formula with at least one other reliable person in advance) and shows a consistently better prediction rate than NFL draft position.

That's exactly what I hope to do. Unless I get access to additional information I'm not sure there's a lot more I could look at in any event.

 

[Dinsy Ejotuz]

Geez, I'm an idiot. (Something Chase can finally agree with.)

These ratings use only info available at the time of the combine. There's absolutely no NFL performance data used to calculate the ratings, except that I'm using actual listed playing weight instead of combine weights. That does clean things up a bit, but it's only truly relevant for a handful of backs IIRC.

I believe that means you could split the ratings any way you want and run them against actual NFL performance if you wanted to test them.

 

[Dinsy Ejotuz]

And for anyone who wants to take a crack at it... here are the 1998-2009 1st to 4th rounders I have in my database currently (estimated draft for 2009), all in one place:

Reggie Bush	 6.06 Felix Jones	 4.86 Rashard Mendenh	 4.79 LenDale White	 4.08 Willis McGahee	 3.83 LaDainian Tomlin 3.77 Ahman Green	 3.75 Fred Taylor	 3.64 J.J. Arrington	 3.58 Edgerrin James	 3.32 Marshawn Lynch	 3.24 Brian Westbrook	 2.84 Ryan Moats	 2.69 Larry Johnson	 2.68 Kevin Jones	 2.65 Deuce McAllister 2.58 Maurice Jones-Dr 2.49 Jonathan Stewart 2.40 Jamaal Charles	 2.18 Steve Slaton	 2.02 Chris Johnson	 1.96 Clinton Portis	 1.77 Frank Gore	 1.73 Darren McFadden	 1.54 Jamal Lewis	 1.48 Andre Brown	 1.32 Steven Jackson   1.24 Knowshon Moreno	 1.21 Chris Wells	 1.09 Cedric Pearman	 1.01 Domanick Willia  0.74 LaMont Jordan	 0.73 Musa Smith	 0.72 Chris Brown	 0.67 Adrian Peterson	 0.65 Correll Buckhalt 0.64 Ronnie Brown	 0.63 DeAngelo William 0.57 Ray Rice	 0.50 Travis Henry	 0.28 Marion Barber	 0.26 Joseph Addai	 0.25 Shaun Alexander	 0.22 Matt Forte	 0.19 Donald Brown	 0.16 Thomas Jones	 0.02 Onterrio Smith	 0.01 Brandon Jackson	 (0.17)Chris Henry	 (0.28)Lee Suggs	 (0.30)Amos Zereoue	 (0.31)Maurice Morris	 (0.40)Ciatrick Fason	 (0.42)Laurence Maroney (0.45)Quentin Griffin	 (0.52)Lamar Gordon	 (0.53)Carnell Williams (0.58)Rudi Johnson	 (0.60)Travis Minor	 (0.72)J.R. Redmond	 (0.75)Cedric Benson	 (0.78)Kevan Barlow	 (0.84)Antonio Pittman	 (1.00)Michael Bennett	 (1.01)Reuben Droughns	 (1.02)Brian Calhoun	 (1.14)Shonn Greene	 (1.20)Julius Jones	 (1.27)Chris Perry	 (1.30)Lesean McCoy	 (1.34)LaBrandon Toef	 (1.36)Javon Ringer	 (1.38)DeShaun Foster	 (1.60)Jerious Norwood	 (1.61)Kevin Smith	 (1.71)Tony Hunt	 (1.87)Justin Fargas	 (1.96)Rashad Jennings	 (1.99)Brian Leonard	 (2.10)Alvin Pearman	 (2.17)Tashard Choice	 (2.23)Dwayne Wright	 (2.66)Mewelde Moore	 (2.68)Kenny Irons	 (3.13)Cedric Cobbs	 (4.64)P.J. Daniels	 (5.42)

 

[DeCleater]

So your using Onterrio Smith as your baseline. Brilliant!

 

[switz]

So your using Onterrio Smith as your baseline. Brilliant!

Considering from a talent perspective Smith was far better than average. Over 5 YPC for his career. Unless you know how to project attitude issues, Smith should have a higher rating than a number of backs higher on that list.