Writers of Pro Football Prospectus 2008

Most Recent FO Features


» 2018 Free Agency Cost-Benefit Analysis

Is Kirk Cousins the best free-agent quarterback in recent memory? Should Trumaine Johnson or Malcolm Butler have gotten the larger contract? And what makes a free-agent contract good or bad, anyway?

23 Apr 2011

Varsity Numbers: Save Your Pick

by Bill Connelly

Last week we reviewed this year's Adjusted POE data, which led to an interesting debate in comments about opponent and offensive line adjustments. I feel I should further specify the components of Adj. POE:

1. The "Expected" piece of Points Over Expected comes from the collection of opponents against whom a runner ran. It basically says, "You had 13 carries against Opponent A, eight against Opponent B, 10 against Opponent C and seven against Opponent D." Those 38 carries should have generated 16.3 EqPts based on those opponents' per-carry averages allowed. You generated 18.1 EqPts in those carries, therefore your Points Over Expected (POE) is 1.8. That is the basis for the raw POE figure.

2. The "Adjusted" portion of Adj. POE comes from the runner's offensive line, particularly the team's Adj. Line Yards ratings. A multiplier is created from that measure and applied to a runner's POE total, in essence crediting a line with a certain portion of the raw POE total. If your offensive line was better than average, then your Adj. POE total is going to be worse than your raw POE; if your offensive line was below average, then your Adj. POE total is going to be higher than your raw POE.

Judging by some of the comments, I'm not sure I had spelled that out clearly, and I wanted to make sure everybody was on the same page.

Now that we've clarified what goes into Adj. POE, let's revisit what we can do with it.

Last year we tied POE and Highlight Yards to draftability, using a set of pulleys and levers. Based on the Adj. POE total and Highlight Yardage rate for a runner's last season in college (career stats were actually less useful), and the Speed Score figure often discussed at Football Outsiders, we were able to separate runners into different categories and probabilities of success. Quoting from last year's piece:

Here are the pre-draft Highlight Yards, Adj. POE and Speed Score averages for players in both categories:

  • Successful: 2.64 Hlt Yds/carry, +12.7 Adj. POE, 108.5 Speed Score
  • Not Successful: 1.89 Hlt Yds/carry, +3.6 Adj. POE, 99.9 Speed Score

So, to analyze a sample of almost 100 rushers, we are going to set up a series of arbitrary points.

  • +1 point: more than 2.64 Hlt Yds/carry, +12.7 Adj. POE or more, 108.5 speed score or higher
  • -1 point: less than 1.89 Hlt Yds/carry, +3.6 Adj. POE or fewer, 99.9 speed score or lower
  • 0 points: everything else

Using this broad system, a player can end up with somewhere between 3 and -3 points. Broken into the four tiers below, you can see defined distance from one tier to another.

The idea was simple enough -- look for the good and bad outliers. Each Draft-eligible rusher was given either one, zero or minus-one point for each of the three statistical categories, meaning each would compile between minus-three and three points.

Rookie seasons are not, in and of themselves, amazingly telling when it comes to predicting a full career's success or failure, but let's see what how last year's Draft class did.

2010 Draft Status
2010 NFL Results
"Plus-3" Backs Ryan Mathews
Jahvid Best
Round 1, Pick 12
Round 1, Pick 30
64 DYAR, 1.9% DVOA
-62 DYAR, -18.0% DVOA
"Plus-2" Backs C.J. Spiller
Jonathan Dwyer
LeGarrette Blount
Round 1, Pick 9
Round 6, Pick 188
-6 DYAR, -10.5% DVOA
-12 DYAR, -42.2% DVOA
98 DYAR, 4.0% DVOA
"Plus-1" Backs Dexter McCluster
Toby Gerhart
Brandon Minor
Round 2, Pick 36
Round 2, Pick 51
21 receptions as wide receiver
-20 DYAR, -26.4% DVOA
practice squad warrior
"Minus-1" Backs Ben Tate
Joe McKnight
Anthony Dixon
Javarris James
Keith Toston
Round 2, Pick 58
Round 4, Pick 112
Round 6, Pick 173
missed season with broken ankle
34 DYAR, 13.2% DVOA
-18 DYAR, -14.4% DVOA
27 DYAR, 4.3% DVOA
-23 DYAR, -35.1% DVOA
"Minus-2" Backs Montario Hardesty
Charles Scott
Keiland Williams
Andre Dixon
Chris Brown
Shawnbrey McNeal
Round 2, Pick 59
Round 6, Pick 200
missed season with torn ACL
practice squad warrior
62 DYAR, 14.8% DVOA
-18 DYAR, -14.4% DVOA
practice squad warrior

This scoring method and NFL general managers agreed on quite a few of last year's backs, at least at the top. The three backs taken in the first round were all "Plus-2" backs or better. There were some interesting disagreements, however. Georgia Tech's Jonathan Dwyer and Oregon's LeGarrette Blount were both "Plus-2" backs who were either late draftees or non-draftees. (Blount was a character issue, to the point where his "Plus-2" status was based on his 2008 performance, not his suspension-laden 2009 season.) Blount thrived, posting perhaps the best numbers in the rookie class. Dwyer, meanwhile, only got a few garbage-time carries late in the year.

Perhaps the most interesting test case of this class, second-rounder and "Minus-1" back Ben Tate, missed the year with a broken ankle. Tate's Speed Score of 114.2 was outstanding, but he just didn't produce much. He managed just a minus-3.5 Adj. POE and 1.82 Highlight Yards per carry. We will have to wait another season to gauge his potential. The same goes for second-rounder Montario Hardesty, who posted a strong speed score (105.8) and virtual nothingness (plus-2.4 Adj. POE, 1.52 Highlight Yards per carry), then missed all of his rookie season to injury.

The 2011 Draftables

While we sit back and wait for more data to come in on the 2010 (and earlier) running backs, let's take a look at the class of 2011. The top "points" guys went first in 2010; that is virtually guaranteed not to happen again this weekend.

Below are 30 Draft-eligible backs, listed in order of their common pre-Draft rankings. This is already seen as a weak class (Mark Ingram is the only likely first-rounder); these numbers will not do anything to change that notion.

2011 Draft Prospects and FO Measurables
(ordered by common pre-draft rankings)
Player School Year 2010
Mark Ingram Alabama Junior 2.23 +8.3 97.7 -1
Ryan Williams Virginia Tech Soph. 1.90 -0.4 98.9 -2
Kendall Hunter Oklahoma State Senior 2.09 +6.5 94.5 -1
Mikel Leshoure Illinois Junior 2.80 +0.5 105.0 +0
Daniel Thomas Kansas State Senior 2.06 +19.5 N/A +1
Johnny White North Carolina Senior 2.70 +3.9 96.7 +0
Jacquizz Rodgers Oregon State Junior 1.71 +1.7 84.6 -3
Jordan Todman Connecticut Junior 1.96 +4.1 108.3 +0
Demarco Murray Oklahoma Senior 1.27 +6.8 112.6 +0
Dion Lewis Pittsburgh Soph. 1.70 +3.5 88.5 -3
Player School Year 2010
Shane Vereen California Junior 1.88 -3.7 102.4 -2
Delone Carter Syracuse Senior 1.87 -4.9 102.7 -2
Bilal Powell Louisville Senior 3.01 +11.6 N/A +1
Derrick Locke Kentucky Senior 2.14 +3.1 100.3 -1
Roy Helu, Jr. Nebraska Senior 3.19 +11.2 114.8 +2
Alex Green Hawaii Senior 4.74 +55.0 106.9 +2
Stevan Ridley LSU Junior 1.52 +0.7 94.6 -3
Anthony Allen Georgia Tech Senior 1.79 -2.7 N/A -2
Graig Cooper Miami Senior 1.59 -2.1 89.2 -3
John Clay Wisconsin Junior 1.72 +2.0 88.9 -3
Player School Year 2010
Da'Rel Scott Maryland Senior 2.57 +6.0 118.9 +1
Mario Fannin Auburn Senior 2.49 +1.2 125.5 +0
Jamie Harper Clemson Junior 1.33 -4.6 105.9 -2
Darren Evans Virginia Tech Junior 2.24 +7.1 103.2 +0
Noel Devine West Virginia Senior 1.85 -11.7 N/A -2
Allen Bradford USC Senior 3.32 +1.7 110.0 +1
Vai Taua Nevada Senior 2.02 -6.1 93.5 -2
Evan Royster Penn State Senior 1.53 -9.1 93.9 -3
Damien Berry Miami Senior 1.49 -13.6 95.1 -3
Armando Allen Notre Dame Senior 1.56 -1.6 N/A -2

There are no Plus-3's on this list, but that isn't necessarily odd; there have been only five going back to the 2006 draft (Ryan Mathews and Jahvid Best in 2010, Chris Johnson and Jamaal Charles in 2008, and Jerious Norwood in 2006). It's the lack of Plus-2s and Plus-1s that is a bit startling.

(One note: Four backs regressed from strong 2009 numbers to decent-at-best 2010 numbers. Mark Ingram, Dion Lewis, and Jacquizz "Pocket Herculizz" Rodgers would have been Plus-0 backs using their 2009 stats, and even with no 40 time; Noel Devine would have been Plus-2. As mentioned above, career numbers haven't correlated any stronger to success than final-season numbers have. With these interesting cases, we'll have to revisit that next year.)

Category Player Proj. Round
Plus-3 Backs None N/A
Plus-2 Backs Alex Green
Roy Helu, Jr.
Plus-1 Backs Daniel Thomas
Bilal Powell
Da'Rel Scott
Allen Bradford
Plus-0 Backs Mikel Leshoure
Johnny White
Demarco Murray
Jordan Todman
Mario Fannin
Darren Evans
Minus-1 Backs Mark Ingram
Kendall Hunter
Derrick Locke
Minus-2 Backs Ryan Williams
Shane Vereen
Delone Carter
Jamie Harper
Noel Devine
Vai Taua
Anthony Allen
Armando Allen
Minus-3 Backs Dion Lewis
Jacquizz Rodgers
Stevan Ridley
Graig Cooper
John Clay
Evan Royster
Damien Berry

Two backs stand out quite a bit on this list, and neither tends to rank very highly on Draft experts' lists. Both Alex Green and Roy Helu Jr. have struggled with fumbles at times, which probably hurts their stock a bit, but since fumbles simply count as zero-point carries in this system, they are possibly not punished as much as they could be. Each has their own other negative traits as well (Helu is injury-prone, and Green runs high), but their numbers are fantastic.

Green's Highlight Yardage and Adj. POE totals were off-the-charts great in 2010 after general mediocrity as a first-year junior college transfer in 2009 (2.15 Hlt. Yds. per carry, minus-4.8 Adj. POE); his speed score is solid but unspectacular. He just needs to hold onto the damn ball. Meanwhile, Roy Helu had both injury and fumbleitis issues at times in his career, but his upside is significant. Just ask Missouri. He has great speed for his size, and he makes great things happen when he gets to the second level. His Adj. POE hints at inconsistency, but he is one of my favorite backs in this Draft, and I'm glad the numbers support me.

As for the backs likely to go first ... there is plenty of potential in players like Mark Ingram (whom I really like) and Mikel Leshoure. But what the numbers seem to be suggesting here is that there is equal potential in guys like Helu, Green and potentially Da'Rel Scott or Allen Bradford, players you can get much later on in the proceedings.

Then there are the three players the 2010 numbers absolutely hate: Devine, Lewis, and Rodgers. Lewis' quickness is outstanding (I loved watching him run the drills during the Combine -- he is strong and decisive), and Devine and Rodgers made an infinite number of plays based on theirs. But while one of them might work out, it's unlikely that two or three would. All three will likely be taken a lot higher than the numbers suggest they should. It will be interesting to see how they do, simply because their final years in college were just not nearly as successful as previous campaigns. They will either prove that using just the final year for these numbers is telling, or whether another approach is necessary.

Posted by: Bill Connelly on 23 Apr 2011

4 comments, Last at 27 Apr 2011, 9:39am by Frosty


by DA (not verified) :: Sat, 04/23/2011 - 6:22pm

Much of Helu's success last year was in part due to Taylor Martinez's rise as a spread-option QB. As soon as his ankle problem (that he got I think against Texas) became obvious for the competition, Helu's numbers dropped significantly.

I think that a better predictor for Helu's performance in the NFL would be his 2009 season, in which he did more than OK in a mediocre offense without an established QB.

This is to say: shouldn't we adjust Adj POE and HLTY/C according to the offensive system in which they played? Same way one would like to correct Lewin's projection system to prevent it from over-praising Air Raid QBs from Texas Tech.

by Kibbles :: Sun, 04/24/2011 - 4:48am

It's interesting to me that 4 of the 5 "+3 backs" all essentially fit the same mold- undersized speedsters with huge per-touch numbers who many people project as CoP backs at the next level. It almost makes me wonder whether this analysis is good at identifying underrated RBs, or if this analysis is good at identifying RBs who fit a very specific mold, and that particular mold just happens to be underrated. In other words, perhaps it's not that big of a deal that Ingram was a "+0 back", because that's only a comment on his archetype as a runner and not on his viability at the next level.

by Aaron Brook's Good Twin (not verified) :: Sun, 04/24/2011 - 11:41pm

Injuries are also an issue. Ryan Williams put up mediocre 2010 numbers, but his draft position is based on his 2009 productive, when he didn't have strained hamstring.

He's basically a faster, shiftier Darren Evans, regardless of 2010 comparison.

by Frosty (not verified) :: Wed, 04/27/2011 - 9:39am

The reason that Jaquizz Rodgers' production dropped so much in 2010 was because of offensive line problems along with breaking in a new QB. The opposing defenses simply lined up to stop Quizz and dared Katz (the QB) to beat them. This was particularly evident once the main WR threat, James Rodgers, went down against Arizona.