Writers of Pro Football Prospectus 2008

Most Recent FO Features

BarrettJT16.jpg

» SDA: Rivalry Showdowns

Rivalry week has significant conference and Playoff ramifications. Should Alabama, Mississippi State, Oregon, or Florida State be worried about getting upset by their rivals?

19 Apr 2010

Introducing SackSEER

Guest Column by Nate Forster

Just as the names Dwight Freeney and DeMarcus Ware strike fear into the hearts of opposing quarterbacks, the names Vernon Gholston, Erik Flowers, and Jamal Reynolds strike fear into the hearts of general managers. All three of those players were hailed for their raw physical talent and ability to sack college quarterbacks. All three have been NFL flops. When general manages swing and miss on high-end pass rushers, failure looks particularly bad on the stat sheet. Flowers only recorded five sacks over the course of his career, Reynolds only recorded three, and so far, Gholston has exactly zero sacks. Yet, in the same years that Flowers and Reynolds were selected by soon-to-be disappointed teams, savvy clubs snatched up players like Aaron Schobel and Kabeer Gbaja-Biamila later in the draft. Is there a way for teams to know before the draft that players such as Schobel, Gbaja-Biamila, Ware, and Freeney will succeed while players such as Gholston, Flowers, and Reynolds will fall flat?

I have poured through college statistics, biographies, and pre-draft workout data to try and shed light on what is one of the great mysteries of the NFL Draft. The result is SackSEER, a regression model that projects the professional sack totals of college edge rushers selected in the NFL Draft. Edge rusher is defined as a player who has spent a significant part of his career lining up as either a defensive end in a 4-3 alignment or an outside linebacker in a 3-4 alignment. The research to produce the method considered all edge rushers drafted in the first two rounds of the NFL Draft between 1999 and 2008.

There are four main factors that correlate to sack success in the NFL: vertical leap, short shuttle time, sacks per game in college (with some playing time adjustments), and how many eligible games worth of NCAA football the player missed for any reason (except early entry into the NFL Draft). SackSEER projects each prospect's total sacks through five years, which is roughly the average length of the rookie contract received by a first- or second-round pick. Although the individual trends are small, when considered together, they project sack production approximately three times more accurately than a player's draft position within the first two rounds. Overall, SackSEER accounts for approximately 40 percent of the historical variation among these players' accumulated five-year sack totals.

Like Playmaker Score, SackSEER is more accurate at identifying busts than it is at singling out potential stars. The best SackSEER projections include some of the most productive edge rushers in the NFL, but they also include some disappointing players such as Jason Babin and Bryan Thomas. By contrast, if many of the players in the worst SackSEER projections seem unfamiliar, nobody could fault you. With the exception of Tony Bryant, none of the players with the worst SackSEER projections have more than five career NFL sacks.

In the tables below, players with less than five years experience are listed with projected five-year totals based on draft position and sacks so far. These estimates are marked with asterisks.

Best SackSEER Projections, 1999-2009
Name College Year Team Pick SackSEER Sacks
(Yrs 1-5)
Aaron Schobel Texas Christian 2001 BUF 46 40.1 46.5
Mario Williams N.C. State 2007 HOU 1 38.6 39.5 (48.5*)
Andre Carter California 2001 SF 7 38.4 32.0
Shawne Merriman Maryland 2005 SD 12 37.9 43.5
Bryan Thomas Alabama-Birmingham 2002 NYJ 22 36.6 15.0
Connor Barwin Cincinnati 2009 HOU 45 36.5 4.5 (25.5*)
DeMarcus Ware Troy State 2005 DAL 11 35.8 64.5
Terrell Suggs Arizona State 2003 BAL 10 35.3 45.0
Courtney Brown Penn State 2000 CLE 1 34.8 17.0
Jason Babin Western Michigan 2004 HOU 27 34.2 15.0
Manny Lawson N.C. State 2006 SF 22 34.2 12.0 (16.0*)
Julius Peppers North Carolina 2002 CAR 2 33.6 53.5
Worst SackSEER Projections, 1999-2009
Name College Year Team Pick SackSEER Sacks
(Yrs 1-5)
Paul Toviessi Marshall 2001 DEN 51 12.5 0.0
Erik Flowers Arizona State 2000 BUF 26 10.9 5.0
David Veikune Utah 2009 CLE 52 10.9 0.0 (12.0*)
Ikaika Alama-Francis Hawaii 2007 DET 58 10.9 1.0 (7.0*)
Erasmus James Wisconsin 2005 MIN 18 10.7 5.0
Anton Palepoi Nevada 2002 SEA 60 10.5 4.0
Robert Ayers Tennessee 2009 DEN 18 9.4 0.0 (12.0*)
Tony Bryant Florida State 1999 OAK 40 8.2 17.5
Dan Cody Oklahoma 2005 BAL 53 7.9 0.0
Jarvis Moss Florida 2007 DEN 17 7.6 3.5 (10.5*)
Jerome McDougle Miami 2003 PHI 15 6.8 3.0
Michael Boireau Miami 2000 MIN 56 2.3 0.0

The trends that SackSEER identifies for edge rushers drafted in the first two rounds persist with later-round edge rushers. For instance, SackSEER would have identified Robert Mathis and Adalius Thomas as top edge rushers. Unsurprisingly, however, SackSEER is not quite as accurate when projecting the edge rusher taken deep in the draft. For now, we are only using the system for players selected in the first two rounds of the NFL Draft. (This will be further discussed in Football Outsiders Almanac 2010.)

Let's look at the four elements in SackSEER and why they help indicate a player's ability to successfully rush the passer in the NFL:

The vertical leap's importance is based on simple physics. If a 270-pound defensive end has the leg strength to jump 40 inches in the air from a standing position, it is very likely that he will be able to employ that same functional strength to burst quickly and powerfully off the line of scrimmage.

The short shuttle run measures change of direction speed, burst, and hip flexibility. DeMarcus Ware had a jaw-dropping 4.07 second short shuttle, and Jevon Kearse ran the short shuttle twice with an average time of 4.12 seconds. No elite edge rusher has emerged from any round of the NFL Draft since at least 1999 with a short shuttle slower than 4.42 seconds.

The third variable in SackSEER is a metric called SRAM, which stands for "Sack Rate as Modified." This measures sacks per game with a few important adjustments. First, sack rates are adjusted to compensate for the fact that college edge rushers as a whole become more productive as they progress through their college careers. A three-year starter who comes out as a junior will be a better prospect than a three-year starter with the same sack rate who is coming out after his senior year. SRAM also attempts to fill in the gaps for players who spent part of their college careers at positions that are less conducive to pass-rushing success, such as defensive tackle (Tamba Hali), 4-3 linebacker (Clay Matthews), or tight end (Ebenezer Ekuban).

The final metric is both the strongest factor and the least intuitive: missed games worth of NCAA eligibility. SackSEER suggests that a college edge rusher who misses numerous games for any reason other than early declaration for the NFL Draft has little chance of succeeding as a professional. This includes players who miss games due to injury, suspensions, academic standards, or sickness. Medical redshirts are included, although standard freshman redshirts are not. Players with health issues in college tend to have health issues in the NFL (Erasmus James, for example). Missing games for other reasons is also indicative of failure at the NFL level. Scroll down to the comment on Jason Pierre-Paul for the remarkable list of failed prospects who spent time at junior college.

The research behind SackSEER will be discussed in more detail in Football Outsiders Almanac 2010, as will the specifics behind the adjustments to sack rate that create SRAM. For now, we wanted to preview this year's draft with a look at how SackSEER evaluates the top edge rusher prospects of 2010. (Please note that the exact projections listed here may differ from those listed in FOA 2010 because of future refinements to the system.)

2010 Draft Prospects

Jerry Hughes, Texas Christian University

Vertical: 34.5", Short Shuttle: 4.15, SRAM: 0.55, Missed Games: 3
Projection: 27.7 Sacks through Year 5

Jerry Hughes separates himself from the rest of the pack by virtue of his elite 4.15-second short shuttle run at the Combine. Not only is Hughes' short shuttle time the best amongst defensive linemen in 2010, but it is also better than any shuttle time run by any edge rusher at the Combine in 2009 or 2008.

The other interesting factor with Hughes is a huge jump in SRAM between his sophomore and junior seasons. Hughes recorded only a 0.09 SRAM for his first two years but recorded a 1.0 SRAM during his dominant junior and senior years. This type of improvement compares favorably to other top edge rushers of past drafts. Patrick Kerney, Elvis Dumervil, and Dwight Freeney had 1.0, 2.0., and 3.5 sacks in their first two years, respectively, but subsequently exploded once they became full-time starters in their junior and senior years. Hughes' career path is similar. He was stuck for two years behind Chase Ortiz and Tommy Blake at TCU.

However, Hughes also bears a certain similarity to another prospect who was not quite as successful as Freeney and company: Jason Babin. Babin, like Hughes, recorded a lot of sacks once becoming a full-time starter at a small school, registered a lightning quick shuttle at the Combine and a mediocre vertical leap. Babin, along with Bryan Thomas, is the type of prospect that SackSEER occasionally misses on: a quick, productive, small-school edge rusher who lacks elite explosion. On the other hand, there are plenty of success stories who share Hughes' particular profile, such as Terrell Suggs, Robert Mathis, and Jared Allen. It all adds up to Hughes being a good, but not great, edge rusher prospect.

Derrick Morgan, Georgia Tech

Vertical: 34", Short Shuttle: 4.43, SRAM: 0.59, Missed Games: 1
Projection: 23.3 Sacks through Year 5

As a prospect, Derrick Morgan is the polar opposite of former teammate Michael Johnson (now with Cincinnati). Morgan was much more productive in college -- building off of a solid sophomore campaign by recording 12.5 sacks in 13 games as a junior -- but lacks Johnson's athleticism, registering mediocre vertical and short shuttle numbers at the Combine. Conventional wisdom says that Morgan is the "safest" edge rusher in the draft, but SackSEER considers Morgan as a middle-of-the-road talent, basically a 50-50 shot.

Everson Griffen, University of Southern California

Vertical: 34", Short Shuttle: 4.36, SRAM: 0.52, Missed Games: 3
Projection: 22.8 Sacks through Year 5

Very few followers of the NFL Draft are particularly excited about Everson Griffen. Unlike Hughes, Griffen has never had a "breakout" season. He has never recorded fewer than 4.5 sacks and never more than eight. Although Griffen's year-to-year sack numbers have been bizarrely consistent at USC, he suffered from consistency problems from game to game, and he was benched more than once for poor play. Giffen's benching only led to him missing three games, so SackSEER is at least lukewarm on his prospects for transitioning to the next level.

Brandon Graham, University of Michigan

Vertical: 31.5", Short Shuttle: 4.25, SRAM: 0.62, Missed Games: 4
Projection: 22.1 Sacks through Year 5

The comparison between Brandon Graham and fellow Michigan alum LaMarr Woodley has been beaten to death by pundits. SackSEER, however, distinguishes Woodley and Graham both by virtue of their prospects for success and by their athleticism. Woodley was an extremely explosive but somewhat "stiff" prospect, recording an excellent 38.5-inch vertical leap and a below-average short shuttle at 4.42 seconds, which added up to a solid 28.2 sack SackSEER projection. Graham, on the other hand, demonstrates good quickness with his 4.25-second short shuttle time, but his 31.5-inch vertical leap raises a major red flag. A more apt comparison for Graham would be Tamba Hali, who shared both Graham’s hustle and lack of jumping ability coming out of college.

Sergio Kindle, University of Texas

Vertical: 36.5", Short Shuttle: 4.53, SRAM: 0.48, Missed Games: 7
Projection: 18.8 Sacks through Year 5

Kindle had a standout performance with two sacks in the National Championship game against Alabama. Very few people have noticed that Kindle had only four sacks in the Longhorns' other 13 games.

Kindle played linebacker for his first two years at Texas, so he gets the full benefit of SRAM's adjustment for playing "out of position." That experience leads many to believe that Kindle is an ideal fit as a 3-4 outside linebacker, but SackSEER is markedly apprehensive due to Kindle's relatively low productivity and missed games. Kindle has suffered multiple minor knee and ankle injuries and, although it is not as fresh in the minds of the football-loving public as Dunlap's DUI, Kindle also served a three-game suspension because of a drunk driving incident.

Kindle's good 36.5-inch vertical leap improves his projection somewhat, but it is more or less canceled out by his poor shuttle run. The most successful edge rushers drafted between 1999 and 2008 after running a 4.5-second shuttle or slower in were Ray Edwards and Tony Bryant. These players had their best seasons as strong-side defensive ends in a 4-3, often lining up opposite slower-footed right tackles. As a classic 'tweener, whose size limits him to rushing the passer and dropping back in coverage, Kindle will probably not be able to fill a similar role.

Carlos Dunlap, University of Florida

Vertical: 31.5", Short Shuttle: 4.61, SRAM: 0.62, Missed Games: 2
Projection: 16.1 Sacks through Year 5

Carlos Dunlap may be a victim of his own hype. At Florida, Dunlap was billed as the next in the line of big men -- such as Mario Williams and Julius Peppers -- with the rare combination of size to be stout against the run and speed to rush the passer. However, many who have reviewed Dunlap's game tape have accused him of "taking plays off" and "not playing as fast as he times."

These critics may be half right, because Combine numbers suggest that Dunlap's total athleticism doesn't match his raw straight-line speed. Compare Dunlap's 31.5-inch vertical leap to those of Mario Williams and Julius Peppers, who leapt 40.5 inches and 36.5 inches, respectively. Dunlap's vertical leap is far more similar to those of the unsuccessful Florida Gators of recent past -- Derrick Harvey and Jarvis Moss. Dunlap also failed to impress with his short shuttle. Ironically, Dunlap's redeeming characteristic is his productivity on the field, where he has received the most criticism.

Jason Pierre-Paul, University of South Florida

Vertical: 30.5", Short Shuttle: 4.67, SRAM: 0.44, Missed Games: 26
Projection: 3.8 Sacks through Year 5

This projection is not a misprint. Although Pierre-Paul is widely regarded as one of the top edge rushers by draftniks, it is hard to conceive of a prospect that SackSEER would like less.

Is it fair to say that Pierre-Paul 26 missed games because he spent his first two years in junior college? Yes, because the few recent edge rusher prospects with junior college experience were among the worst defensive draft picks in recent memory. The illustrious list includes Anton Palepoi, Jerome McDougle, Tony Bryant, Lamar King, Erik Flowers, Michael Boireau and most recently, David Veikune, Cleveland's 2009 second-round pick, who did not record a sack as a rookie and was inactive for much of the season. Unless you count Julian Peterson, a 4-3 linebacker who doesn't fit this study, you have to go back to Leonard Little in 1998 to find a former junior college player who panned out in the NFL, and Little spent only one year at junior college. When Tony Bryant is your upside, you're in trouble.

So why do junior college edge rushers struggle so mightily when transitioning to the NFL? Most edge rusher prospects who play at the junior college level miss two years worth of their NCAA eligibility, and they are understandably "raw" when they enter the professional ranks. Although many coaches believe that such a player can be "coached up," the coaching staff can only devote so much of its valuable coaching resources to a single player. Moreover, many players who go to a junior college have significant or severe academic issues, and expecting them to digest a complicated NFL playbook while also "catching up" on their fundamentals may be unrealistic.

However, even if we removed the missed games adjustment, Pierre-Paul would still have the worst projection in this class. Although Pierre-Paul's "handflips" video has earned him the distinction as the 2010 NFL Draft's consummate physical "freak," he put up poor numbers in all of the workout metrics that matter to SackSEER. This is certainly not the first time that similar "gimmicky" athleticism has not translated into workout numbers. For instance, just last year, San Diego State defensive lineman Jarron Gilbert made waves (pardon the pun) with a widely circulated video depicting Gilbert jumping out of a swimming pool flat footed. Despite his aquatic jumping ability, Gilbert only registered a 35.5-inch vertical leap at the Combine.

The ability of Pierre-Paul to translate his particular brand of athleticism to pass rushing success is speculative at best. Overall, the general manager who pulls the trigger on Pierre-Paul better be very confident that he has something special -- so special that it will completely buck the historical trends.

SackSEER Sleepers

When it comes to sleepers, SackSEER rates two prospects above the rest: South Carolina linebacker Eric Norwood and Washington defensive end Daniel Te'O'Nesheim. Eric Norwood's Combine numbers are similar to those of Clay Matthews a year ago, and he had better college sack production. Te'O'Nesheim has been billed as a productive, high-energy player who lacks the athleticism to compete at a high level, but his 37-inch vertical leap and a 4.18-second short shuttle at the Combine are enough to convince SackSEER.

Nate Forster is an attorney in Boston whose hobbies include drinking Slurpees, wearing hooded sweatshirts, and delivering bone-crushing hits on opposing quarterbacks. Although never invited to the NFL Combine, he did hold a personal pro day where he amazed all in attendance by reading a transcript of a Todd McShay-Mel Kiper debate in the voices of Shaggy and Scooby Doo. Special thanks to my fiancée for putting up with me while I "was doing stats," and to the posters at the FootballOutsiders.com forums, especially AlanSP, for their thoughtful comments on an early version of this model.

Editor's Note: If you are interested in publishing a guest column at Football Outsiders, please e-mail your idea or rough draft to info-at-footballoutsiders.com. Come up with something really good, like SackSEER or the Lewin Career Forecast, and we may even ask to publish it in our book.

Posted by: Guest on 19 Apr 2010

116 comments, Last at 04 Nov 2012, 3:21pm by DCW

Comments

1
by Jimmy :: Mon, 04/19/2010 - 12:20pm

Nice article, I remember reading the thread you put up with as you were developing the system and it is good to see where it ended up.

It doesn't seem to be all that good a class of pass rushers if the best of the bunch can be expected to acrue 5.5 sacks a year.

25
by Rich Conley (not verified) :: Mon, 04/19/2010 - 3:44pm

Any metric expected to correlate to a group as a whole is going to have compressed ranges. IE, if you expect guys to average 5.5, there's usually going to be some who overperform, and some who underperform.

This gives you an idea who has the best chance of turning into a stud, and who has very little chance.

In reality, any of the 20+ projected guys could probably end up as studs. The rest of them, probably not.

39
by Jimmy :: Mon, 04/19/2010 - 7:28pm

I get that. According to this analysis it still doesn't make it a good class for pass rushers.

57
by bravehoptoad :: Tue, 04/20/2010 - 11:36am

Seriously, compared to last year when there were a few guys a good bit over 30.

edit -- woops, just saw post #48 below. So I should say, yes, seriously, considering that no one is this year's class comes within 5 of making the top ten.

2
by jklps :: Mon, 04/19/2010 - 12:27pm

What is the projection for Brian Orakpo?

15
by Jimmy :: Mon, 04/19/2010 - 2:00pm

The table was linked in the thread that was linked below.

His projection was 37.2 so even if that is his eventual total he would still be good for six sacks a year after his very good rookie season.

48
by SackSEER :: Tue, 04/20/2010 - 12:10am

Hello,

I am the SackSEER dude.

Since there is some linking going on to the earlier model, I just wanted to clarify about what the updates have to say about 2009. If Orakpo was in this draft, his projection would fall just about between Hughes and Morgan. Also, the adjustments that I have made downgrade Maybin considerably, because they better account for regression to the mean following his breakout sophomore season.

Last year's draft had a lot of quirky prospects with high-ish projections. You had two redshirt sophomores (which had never happened in the data set before), a senior defensive end that played tight end for three years, and a senior "elephant" backer who had played a more traditional 4-3 linebacker role in his first three seasons. I feel better projecting this class because we have all straight-up juniors and seniors, and lots of players with projections on the low side.

3
by alexbond :: Mon, 04/19/2010 - 12:36pm

I went to UW and have seen a whole lot of Daniel Te'O'Nesheim. He unfortunately has played at UW while the team was really really bad. He is even better than his sack stats suggest because he was doubleteamed basically every single play of his career and certainly every play for the past 2 or 3 years. In his time with us, he was the only pass rush threat on the team and certainly the only player on the D-line with any chance of making a play in either run or pass situations. And because the team lost so much, the other team was rarely in "must pass" situations and he couldn't pin his ears back and sell out to the pass rush. He is a pretty complete guy, capable of making plays in both the run and the pass game. I think he might be a little small for a 4-3 end, and I don't think he can cover well enough to be a 3-4 endbacker, but some team will get an excellent tweener in the 3rd or 4th round.

31
by Bobman :: Mon, 04/19/2010 - 4:57pm

Colts MUST pick Te'O'Nesheim in the 3rd. I was SUPER high on Triplett when he came out a decade ago as well and was thrilled he fell in the 2nd round to Indy. His career was mediocre, not entirely hi fault.

As an end, I think he's taller and heavier than Freeney, for example, but not as fast. A great article I read in some Pac-10 mag last fall quoted his HS coaches describing him as "the smallest Samoan I'd ever seen." That was as a 5-8 135 lb +/- freshman in HS.

36
by Anonymous Coward (not verified) :: Mon, 04/19/2010 - 6:26pm

Te'O'Nesheim is one of my favorite prospects in this draft. I have no idea why anyone would take Pierre Paul over him, yet Pierre Paul is typically lumpped in as a first rounder by all the draft sites and Te'O'Nesheim generally started as a 5th or 6th rounder on most of their boards but has been moving up. Somehow Pierre Paul with his 59 KEI and 13 backflips is labeled the elite athlete while Te'O'Nesheim who broke the 70 KEI threshold got labeled as the try hard but limited athlete type. No one has been able to explain that to me.

44
by Eric D (not verified) :: Mon, 04/19/2010 - 10:28pm

You ask why Pierre Paul gets labeled the elite athlete while Te'O-Nesheim is called the high-motor but limited athlete? Simple: Pierre Paul is black, Te'O-Nesheim isn't. I have long noticed how a non-black player can run a 4.4 with a 40" vertical, super shuttle and cone drill, etc., etc., and will always be described as "very intelligent" and "great motor" which makes up for "limited athleticism." Most people, even supposed experts, can't seem to see beyond race and just describe what the tale of the tape actually says.

89
by DaShontay (not verified) :: Wed, 04/21/2010 - 11:58pm

That's true. Toby Gerhart got screwed over also, do to racism. He should have won the Heisman, and he is getting dissed in the draft as well, for that one reason.

58
by bravehoptoad :: Tue, 04/20/2010 - 11:42am

Kirwan Explosion Index, for those of you (like me) who didn't recognize the acronym.

4
by Joseph :: Mon, 04/19/2010 - 12:51pm

Nice research. Having seen some different ideas with mock drafts, I hope my Saints at 32 can select Hughes from TCU. One of the Saints' beat writers actually selected him in his mock (published Sunday, iirc).
I'd be interested in the numbers for the other USF guy, George Selvie.

5
by Dave51 (not verified) :: Mon, 04/19/2010 - 12:53pm

Cool article

6
by Fan in Exile :: Mon, 04/19/2010 - 12:57pm

Wow, Ayers looks really bad right now. I'm hoping he's a fluke.

70
by AlanSP :: Tue, 04/20/2010 - 5:49pm

I frankly never got what anybody liked about Ayers in the first place. His college numbers were awful, his athletic "measurables" were lousy, his size/strength was nothing to get excited about either, and he had character concerns to complete the package. He apparently had a nice senior bowl, and he played the run pretty well as a senior, but I have no idea what made teams look at him and think "mid-first round pick"

71
by tuluse :: Tue, 04/20/2010 - 8:07pm

most TV personalities didn't like him. It was widely regarded as an odd choice when Denver drafted him.

84
by Eddo :: Wed, 04/21/2010 - 3:00pm

He *was* expected to be drafted in the teens, though. At least in the first round. Enough people outside of Denver must have liked him enough for that consensus to develop.

80
by Kibbles :: Wed, 04/21/2010 - 4:32am

Well, in McDaniels' defense, he only had 50 names on his draft board, so it's not like there were a lot of players for him to choose from at that spot.

Seeing all those Broncos busts on this list really brings me back. I never thought Paul Toviessi would get hurt and never play a down, but I was definitely bearish on both Moss and Ayers when they were drafted for much the same reasons why SackSEER was bearish on them.

96
by Phoenix138 :: Sat, 04/24/2010 - 9:34pm

Mike Mayock loved the guy. I think it was Ayers' handplay that Mayock thought would make him an effective player in the pros.

7
by Dean :: Mon, 04/19/2010 - 1:01pm

Interesting. It's hard to evaluate this without more of the rationale behind the scenes. Why did you decide those numbers work, and some others don't?

I'll be looking forward to filling in the gaps when PFA comes out.

13
by Jerry :: Mon, 04/19/2010 - 1:48pm
16
by Jimmy :: Mon, 04/19/2010 - 2:01pm

Jerry

Could you let me know how you embed links like that?

20
by Eddo :: Mon, 04/19/2010 - 2:50pm

<a href="URL">TEXT FOR LINK BODY</a>

33
by Jerry :: Mon, 04/19/2010 - 5:52pm

What Eddo said.

38
by Jimmy :: Mon, 04/19/2010 - 7:19pm

So you put the web address where the TEXT FOR LINK BODY is.

And cheers for the help.

41
by Lance :: Mon, 04/19/2010 - 8:05pm

No, no. You put the web address in the "URL" spot. The "Text for link body" is where you put a description. For instance:

<a href="http://www.espn.com">The Worldwide Leader</a>

Will end up looking like this:

The Worldwide Leader

Where clicking on "The Worldwide Leader" takes you to the ESPN website.

46
by Nathan :: Mon, 04/19/2010 - 11:06pm

wait, now i want to know how you got that link to show us as text and not a link

51
by Lance :: Tue, 04/20/2010 - 2:16am

The trick is to embed other tags within the tags. So for instance, to make this show up:

<a href="http://www.espn.com">The worldwide leader</a>

I've put an <em> and </em> between the "<" and the "a" and then again between the "<" and "/a". (Though any of the tags will do.) It looks like this:

<<em></em>a href="http://www.espn.com">The worldwide leader<<em></em>/a>

To actually get that to show up took even more embedded tags. But trial and error will eventually show you how it works.

52
by Dales :: Tue, 04/20/2010 - 4:50am

There is a different way to make 'tags' visible rather than having them get used. Use the ampersand key followed by the letters lt and a semicolon to make the < sign (full code &lt;), and the same but with gt to make the > sign (&gt;). Use those two four symbol codes in place of the < and > and it will get displayed rather than interpreted:

<A HREF="http://thisisafakeurl.com">This is a fake label for my fake url</a>

The above was made by using:
&lt;A HREF="http://thisisafakeurl.com"&gt;This is a fake label for my fake url&lt;/a&gt;

8
by geekosphere101 (not verified) :: Mon, 04/19/2010 - 1:03pm

Very interesting. SackSEER lines up pretty well with my opinions of this class. I like what little I've seen of Hughes, Morgan seems like a safe pick, especially considering his ability in run support, JPP has bust written all over him, etc.

Edge rusher seems to be one of the hardest positions to scout correctly. Because it is one of the top 3 in positional value, it will be interesting to see how this metric hold up.

9
by speedegg :: Mon, 04/19/2010 - 1:07pm

whoa, this is cool. I wonder how many late nights and algorithm crunching SackSEER took to produce. Great stuff.

10
by Karl Cuba :: Mon, 04/19/2010 - 1:13pm

So you now have Speed Score, Playmaker, the Lewin Career Forecast and now SackSEER, as well as your kicker thing, whose name I forget. How long until you have every position and are then able to statistically project the draft?

I really like the missed games metric, a lovely little abstraction that allows you to analyse the cumulative impact of several different factors that are often difficult to quantify. I would have thought that it could immediately be applied to other positions to achieve the same effect.

(some height/weight figures might be helpful to put players in context)

Top stuff.

11
by Elric :: Mon, 04/19/2010 - 1:21pm

I'd be interested to find out what Ndamukong Suh's expected sack rate would be... he's an outlier for your statistics - but people have claimed he could play anywhere on the line so it'd be an interesting exercise.

His stats:
Vertical: 35.5", Short Shuttle: 4.44, SRAM: ? (24 sacks in 54 games with 38 starts all at DT), Missed Games: 11 (medical red shirt)
Projection: ? Sacks through Year 5

18
by wr (not verified) :: Mon, 04/19/2010 - 2:16pm

It would be interesting to see if this projection worked as
well for DTs and DEs (was it tried on DTs?).

Captcha : barnyard time (scared)

12
by Monkey Business (not verified) :: Mon, 04/19/2010 - 1:26pm

Now if you guys could create studies for QBs and LTs, all someone would need to do to build a great pro franchise is subscribe to FO.

14
by drobviousso :: Mon, 04/19/2010 - 1:48pm

Cool post dude. Look forward to the expanded version in the book.

17
by Dales :: Mon, 04/19/2010 - 2:15pm

Looks like someone in Houston may have figured out something along these lines, while someone in Denver most decidedly has not.

22
by SackSEER :: Mon, 04/19/2010 - 2:58pm

I thought the same thing that you did when I was crunching the numbers.

Also, having the benefit of having the full data in front of me, I would also say that perhaps Al Davis has figured out something that SackSEER hasn't, because he has gotten moderate production out of a few players that SackSEER absolutely hated, which is why I have to say "except for Tony Bryant" after every other sentence. Also, how Matt Shaughnessy managed to fall on top of the quarterback four times last year is beyond me.

24
by Mr Shush :: Mon, 04/19/2010 - 3:33pm

Note that those three Texans were drafted by three different head coach/general manager combinations: Capers/Casserly for Babin, Kubiak/Casserly for Williams and Kubiak/Smith for Barwin. Casserly had final say in the Capers era, but lost it to Kubiak when Kubiak was hired (and by most accounts other than his own had no real say in the 2006 draft, immediately after which he was axed and replaced with Smith). I have no idea how many of the more junior front office staff survived the regime change, or if such a notion might have originated with one of them. To be fair, the Texans have largely drafted very well in general since Kubiak was hired, so it may simply be a case of good overall talent evaluation (coupled with an initial roster containing precisely no viable 4-3 DEs and hence a propensity to draft them high) leading to similar results to a statistical system, without such a system actually being used.

19
by Dave0 :: Mon, 04/19/2010 - 2:22pm

jarron gilbert went to san jose state, not san diego state

21
by Ron (not verified) :: Mon, 04/19/2010 - 2:52pm

Just curious how/why you used Conner Barwin as an example in your chart when he had 4.5 sacks on 369 offensive snaps while Cameron Wake had 5.5 sacks on less than half the offensive snaps(167 offensive snaps)

I would think Wake would project much higher with the numbers

23
by Mr Shush :: Mon, 04/19/2010 - 3:26pm

The lists above are the players who had the highest projections based on their college careers and combine performances, not their NFL performance. Also, only players drafted in the first two rounds were considered, as the system becomes less accurate for later drafted players. Barwin was a second round pick; Wake was undrafted. Finally, Wake left school in 2005, meaning those 5.5 sacks represent the totality of his NFL value through five years of being eligible to play pro football. He was a CFL stud, meaning his first live NFL action came at a point where he was 27 years old and had two years of prior pro pass rushing experience. Barwin saw his first NFL action at not quite 23, having rushed the passer for only one season in college (he was previously a tight end). Wake was a better player than Barwin in 2009, but I think it's very likely Barwin will have the more productive career when all's said and done.

87
by John (not verified) :: Wed, 04/21/2010 - 7:18pm

Not sure how many games he missed or how his sack numbers would translate, I know he moved around at Penn.

But Wake's other two numbers:

Shuttle: 4.13
Vertical: 45.5 (!)

He was the best pass rusher in the NFL last year, snap for snap. He had 7 sacks and 20 QB pressures in 167 snaps.

I wouldn't bet against him this year now that Porter is finally gone.

26
by Rich Conley (not verified) :: Mon, 04/19/2010 - 3:46pm

"Also, only players drafted in the first two rounds were considered, as the system becomes less accurate for later drafted players. "

Is anyone else kind of convinced that all of the metrics having this same caveat is an indication that there's some institutional inertia happening on teams? IE, lower picks not getting playing time despite being better players than the guys above them?

27
by Big Johnson (not verified) :: Mon, 04/19/2010 - 4:04pm

thats definately a possibility and is probably fairly common. Reminds me of one of my least favorite sports cliches and mostly used for quarterbacks.

We must start quarterback A over quarterback B because we have invested X amount of dollars in quarterback A.

56
by Mr Shush :: Tue, 04/20/2010 - 11:11am

That may be part of it, but I think it's mostly just about controlling for high production from non-NFL calibre players due to system, quality of competition or some combination of the two. That's probably more true of LCF than SackSEER, since SackSEER involves a substantial combine component which is not dependent on college performance.

68
by tuluse :: Tue, 04/20/2010 - 4:30pm

No, I think they don't play because they're not very good.

Especially at a position like defensive linemen where you can rotate so many players in.

28
by mattman :: Mon, 04/19/2010 - 4:10pm

I was just reading scouting reports on Jerry Hughes over the weekend, and thought he sounded like Trent Cole. So I was already hoping the Eagles would draft him at #24 before I read this report. Now I'm even more convinced.

Incidentally, Cole's numbers:

Vertical: 38
Shuttle: 4.22
Sacks: 19 (played a lot of linebacker)

Can't find games played or missed.

35
by justanothersteve :: Mon, 04/19/2010 - 6:10pm

Mike Vandermause of the Green Bay Press-Gazette also thinks the Packers may be wanting to draft Hughes at #23 to play opposite Clay Matthews. Wonder if we'll see a draft day trade up by someone to pick him.

64
by AlanSP :: Tue, 04/20/2010 - 3:26pm

Actually, Trent Cole didn't play LB, but he did play DT (at 220 lbs!), which similarly lowers sack totals. He played 38 games in college. Not sure about how many he missed

73
by SackSEER :: Wed, 04/21/2010 - 12:46am

I have Trent Cole down for 12 missed games. He was a Proposition 48 casualty his first year at Cincinnati. His individual metrics, however, are still strong enough that SackSEER (7 Round Version) thought that he was a solid value pick.

76
by Nathan :: Wed, 04/21/2010 - 1:08am

Trent Cole is pretty good. Cursory Wikipedia search shows some good pass rushers with short shuttle times in the 4.3-4.5 areas. Should SackSEER (good name) value short shuttle times b/c of acceleration? Or does it more estimate cutting ability (hard to change directions if you weigh 275, as a pass rusher you pick a move and go with it, rock paper scissors style and trust explosion)?

29
by Karl Mal0wned! (not verified) :: Mon, 04/19/2010 - 4:17pm

Haha, the picture accompanying this article was too distracting for me to proceed and actually read the article itself, rats.

30
by David Gardner :: Mon, 04/19/2010 - 4:31pm

Yeah, isn't it cool?

Your eyes are naturally drawn to images with strong diagonals. It was an added bonus that his shirt matched the green theme of the site.

83
by Karl Mal0wned! (not verified) :: Wed, 04/21/2010 - 2:11pm

Yeah, it was the strong diagonals, not the fresh look on his face or the weird thing he's doing with his hand :D

Dave, you sir are a multi-talented and well-informed individual, thank you.

32
by Dan :: Mon, 04/19/2010 - 5:48pm

Nice analysis.

Could you say something about what other factors you tested in your regressions? For instance, there's no mention here of height, weight, or arm length. Is that because you included those variables in a regression, they came out nonsignificant, and you concluded that they were not predictive? Or is it because you haven't tried including them (perhaps because you don't have those data)?

34
by Jason Lisk (not verified) :: Mon, 04/19/2010 - 6:08pm

I'm concerned that this involves a fair amount of data dredging or data snooping and should not be relied on yet to make the kind of evaluation you claim, similar to the LCF.

Did you form your hypotheses before looking at the data that those four factors were the most important, and then check the data set? Or did you look at numerous variables and find those within the data set that gave the best results?

Did you test it against an independent data set, such as picks from 1992-1998?

If you were looking for something by looking at a bunch of variables, you were bound to find something.

40
by Jeff M. (not verified) :: Mon, 04/19/2010 - 7:46pm

Except that the task of this research is self-consciously inductive. They're not doing any hypothesis testing.

55
by GnomeChumpsky (not verified) :: Tue, 04/20/2010 - 9:37am

1) Silly, that's what the next season or two is for (hypothesis testing)

2) This is building a predictive model, not trying to prove some universal truth (e.g. photons might have a great speed score, but would be terrible running backs). It has obvious limitations, most notably that it relies on a certain degree of self-selection on gestalt evaluations at the level of college (e.g. who plays where and against whom and who gets coached, etc.) and even in the draft (e.g. ZOMG this doesn't work for low draft picks). However, it currently accounts for the variance better than draft number (also suspect because we aren't just drafting pass rushers and the draft is about game theory and getting "dibs" more than getting the players to line up based of ability and worth) and, most importantly, it's FUN!

3) You come up with something better (not a put down, a request, I would like to see it [also not sarcastic, a better model would be awesome and it would create some academic drama])

79
by Bill Barnwell :: Wed, 04/21/2010 - 1:58am

I just want to point out here that the LCF isn't an example of data dredging.

It's a pretty standard application of regression. The research question, in the case of the LCF, is "Are there any variables that predict professional success for quarterbacks?", with the null hypothesis being "There are no variables that predict professional success for quarterbacks outside of Jon Gruden's good face."

Next, a variety of data was gathered on a large number of college quarterbacks. This is where I think you're getting confused -- gathering a bunch of relevant data and using it to measure causation isn't data dredging. There's enough of a sample size that using part of the data (say, the first half of the timeframe for which QB data was available) to retrodict the performance of the players in the second half of the timeframe yields an effective model. (Since I didn't do the research in question, I don't have the exact numbers in front of me.)

Then, after eliminating those variables that don't have any predicative power, you're left with the final regression, containing statistically significant variables that have been tested against parts of the dataset not included in the regression. The LCF formula was enough to reject the null hypothesis.

81
by jimbohead :: Wed, 04/21/2010 - 11:01am

Thanks for the explanation of the background on LCF. Very helpful.

The article that wikipedia cites is a really good and valuable read. As long as LCF was tested against an independent data set, we're all fine and dandy. I'd whine about this study (potentially) not having done that, but we've already discussed what an abusrdly tough thing it is to put together that database. And once you start trying to look up combine and pro-day numbers from the early 90s, I imagine that it may be an impossible task.

http://www.plosmedicine.org/article/info:doi/10.1371/journal.pmed.002012...

37
by Reinhard (not verified) :: Mon, 04/19/2010 - 6:46pm

There seems to be one variable missing, namely the players size/power.

For example Brandon Graham, he is THICK I am not surprised he has a bad vertical. He sort of looks like an overgrown MLB or a really fast DT, and he plays more with his leverage and hands than with his burst.

Similar to Speed Score, the vertical is much more/less impressive when considering a player's height. I'm not saying that this variable was a strong indicator in the analysis that you ran, just that it seems not to match up with my expectations.

49
by SackSEER :: Tue, 04/20/2010 - 12:29am

Thanks for the thoughts Reinhard,

There is no evidence in my data of a relationship between a player's height and their vertical, and there is only a tiny correlation between BMI and vertical. In any event, it certainly isn't enough to excuse a 31.5" that is well below average. Freeney was a very thick, short, player, and hit 36.5".

I will say that I am, too, a little disappointed in Graham's projection. I have seen a bit of him because I am a University of Michigan grad, and he is so smooth, and gets good leverage, that I would not be surprised if he beats his projection. I also had the impression watching him that his sack rate last year was a little understated by the numbers. But that may just be the homer in me talking.

----------
SackSEER

60
by Reinhard (not verified) :: Tue, 04/20/2010 - 1:13pm

Shoot.... I meant to say there at the end that the vertical is more or less impressive considering the players WEIGHT

74
by SackSEER :: Wed, 04/21/2010 - 12:51am

This is true. Heavy players have more trouble with the vertical. But I did try adjusting for weight and plain old vertical does work better.

Taking all of that into consideration, ZOMG Mario Williams! 40.5" at 295 lbs! Why did anybody ever think that passing on Reggie Bush for this guy was a bad idea?

77
by tuluse :: Wed, 04/21/2010 - 1:43am

Because too many people remember the 70s when an elite running back was all you needed to make the playoffs.

78
by Nathan :: Wed, 04/21/2010 - 1:54am

Cause Barry Sanders was BADASS.

61
by Karl Cuba :: Tue, 04/20/2010 - 2:17pm

Have you thought about including hurries and knockdowns? Doesn't FO research suggest that the cumulative total of those stats is more predictive of future success, or more consistent at least. Just a thought.

62
by jimbohead :: Tue, 04/20/2010 - 2:59pm

Does NCAA football track those? Honestly, what's probably most impressive to me about this study is the amount of effort it must have taken to put together the database behind the regression.

63
by AlanSP :: Tue, 04/20/2010 - 3:05pm

Hurries are tracked, but as far as I can tell, it's done differently from team to team, and some teams are far more liberal than others about what to count as a hurry, so the data end up being fairly useless.

And yes, having put together a similar database myself (but with all 7 rounds), I can tell you that it takes a huge amount of time and effort to piece together all of the data from different sources, especially for lesser-known players.

42
by UNFNole (not verified) :: Mon, 04/19/2010 - 8:41pm

I love it! As you said it is probably a better predictor for bust than successes, but will you explain to me the decision to leave out upper body strength? Vert and the Shuttle Run both account for lower-body explosion, but wouldn't it be fair to say that the ability to shed blocks or break the edge in the NFL also heavily relies on upper-body explosion? I would say that a huge part of the learning curve is the difference is strength from a College TE or T to an NFL one.

Would this effect the metric much or help it more accurately gauge future success? Just curious.

Once again though, great work!

75
by SackSEER :: Wed, 04/21/2010 - 1:01am

I would not necessarily disagree with you at all. But the problem is that we do not (at least not yet) have a good way to measure upper body strength that is indicative of success.

-----------
SackSEER

43
by CJ (not verified) :: Mon, 04/19/2010 - 10:03pm

Other potential "sleepers", although both have 2nd-3rd round grades:

Koa Misi, Utah- 38" VJ, 4.27 shuttle, started 36-38 games at Utah but only had 10.5 sacks

Jason Worilds, VT- 38" VJ, 4.29 shuttle, 14 sacks in 3 years

45
by RICKY3245 (not verified) :: Mon, 04/19/2010 - 11:06pm

Loved the read, and thanks. One correction though Gilbert when to San Jose State not San Diego State.

47
by Dice :: Mon, 04/19/2010 - 11:48pm

Very interesting. Definitely like to see the results after five years, like with the draft classes.

50
by Rick B. (not verified) :: Tue, 04/20/2010 - 1:13am

So the conclusion we should draw from this is:

There aren't any elite pass-rushing prospects in this year's class, but Jerry Hughes, Derrick Morgan, Daniel Te'o-Nesheim and Eric Norwood might be solid?

Hmm.

53
by bubqr :: Tue, 04/20/2010 - 5:18am

Can we get Gholston projection ?

59
by Led :: Tue, 04/20/2010 - 11:54am

Good question.

EDIT: Gholston had a 35.5 inch vertical and a 2.40 short shuttle time. He had 22.5 sacks in two years (14 as a senior), missed 10 games as a sophmore with a broken hand and didn't play much as a freshman. His so-so shuttle time and 10 missed games would hurt his SackSEER projection.

54
by Brendan Scolari :: Tue, 04/20/2010 - 7:30am

Very interesting stuff, it will be interesting to see how this metric fares as you get more data.

Jarron Gilbert actually went to San Jose State, Not San Diego State.

65
by AlanSP :: Tue, 04/20/2010 - 3:46pm

I remember discussing this with you last year, and I was working on something similar myself before getting distracted by real life, so I'm glad to see that you fleshed it out more. I've been touting Jerry Hughes to my friends for a while now based on this; he's a classic example of a guy who's underrated because he's slightly smaller than average.

As I recall (I haven't looked at the data in a while) a similar set of factors was useful for the middle rounds as well, although there are problems with noise there because the relative number of players who do anything at all in the NFL drops off as you get later in the draft. Still, players like Jared Allen, Elvis Dumervil, and Robert Mathis fare well here (Mathis's college numbers in particular were jaw-dropping).

What college stats did you look at as possibilities when you were creating this? I remember when I was running the regressions last year, the real head-scratcher was that interceptions kept coming up as a significant predictor. I think that was just because of a handful of guys like Julius Peppers, Jared Allen, Mathias Kiwanuka, etc. Forced fumbles often came up as well, but those are highly correlated with sacks.

67
by AlanSP :: Tue, 04/20/2010 - 4:09pm

Just noticed the paragraph where you said pretty much the same thing about the later-round picks.

I have a sudden urge to go back and look at all of this again, but I need to get working stats software on my computer :/

72
by SackSEER :: Wed, 04/21/2010 - 12:43am

Heya AlanSP,

Glad that you're still about and kicking. I agree with you on the interceptions. Julius Peppers has a really obscene number, and Allen has a few too. Those outliers seem to really skew the data and interceptions are such rare events that I just cannot believe that they really have much predictive power.

I looked at forced fumbles closely in in the first two rounds especially. They are very collinear with sacks. I even tried looking at FF/sack to see if there was some magic about sacks that come from the weakside and cause fumbles. No dice there.

Also, your comment in 63 about how difficult it is to compile these statistics is well-taken. This took me a huge amount of time to collect data. The most difficult to find are pro days, because proving whether or not a player had a pro day is essentially like trying to prove a negative. There were definitely days when I felt like shaking my fist at the sky and saying "why has the world forgotten Adrian Dingle? Why?!"

66
by AlanSP :: Tue, 04/20/2010 - 3:51pm

The games missed metric is really interesting. I had never even thought to look at that. Do years spent at community college count as games missed in this sense?

I'm also extremely interested in how you adjusted for time spent at other positions. This was one of the most frustrating things I encountered in trying to make sense of the data, particularly since it's often hard to even determine exactly how much time a player spent at each position.

69
by AlanSP :: Tue, 04/20/2010 - 5:18pm

And just noticed the comment about community college in the Pierre-Paul comment. I should have read this more thoroughly before responding

82
by Anonymous1212121212 (not verified) :: Wed, 04/21/2010 - 11:14am

"Yet, in the same years that Flowers and Reynolds were selected by soon-to-be disappointed teams, savvy clubs snatched up players like Aaron Schobel and Kabeer Gbaja-Biamila later in the draft."

Just a note, Erik Flowers was drafted by the Bills in 2000, while Aaron Schobel was drafted by the Bills in 2001.

85
by AlanSP :: Wed, 04/21/2010 - 6:19pm

More assorted questions/comments:

-How did you handle guys who played at both DE and LB in a 4-3 (e.g. Mathias Kiwanuka, Brian Orakpo) and guys who played at both 4-3 DE and 3-4 DE (Shaun Ellis and Richard Seymour come to mind, although I realize that both of them switched after their 1st 5 years)? I think there might also be some guys who get moved inside to DT (e.g. Anthony Hargrove, although again, that was after 5 years).

-What did you use as the dependent variable for guys who were drafted after 2005 and so haven't been in the league for 5 years yet? Was it that projected total shown in the chart?

-I'm impressed that you managed to track down pro day results for all of those guys, but do you worry about the reliability compared to the combine (which is more of an even playing field)? I wouldn't expect this to make much of a difference in something like a vertical jump, but differences in the surface players are running on could matter for something like the shuttle.

-Are games missed just the number of games that a player's team played that he didn't play in? i.e. if player A and player B both miss the entire year with an injury, but player A's team makes a conference championship/bowl game/I-AA playoffs and player B's team just plays their regular schedule, does player A get penalized for missing more games?

-One other stat I forgot about that came up in some of my regressions last year was Run TFL/game (i.e. total tackles for a loss minus sacks). I haven't looked at this that closely though, so I don't know if there's really anything to it. It's somewhat colinear with sacks, but not nearly to the degree that FF is, so it might be adding something different.

86
by tuluse :: Wed, 04/21/2010 - 7:16pm

Could you measure sack rate as sacks per opponent pass attempt, rather than per game, or is is too hard to get that data?

88
by Reality_Check (not verified) :: Wed, 04/21/2010 - 7:37pm

Jarron Gilbert played at San Jose State, not San Diego State...

90
by AlanSP :: Thu, 04/22/2010 - 7:10pm

Also, out of curiosity, what does SackSEER say about Arthur Moats? His shuttle was just okay at 4.37, but the 36.5" vertical is good, he never missed a game in college, and his sack profile is similar to Jerry Hughes.

92
by AlanSP :: Fri, 04/23/2010 - 2:31pm

Other couple other possible late-round guys who seem to profile reasonably well in terms of SackSEER's criteria:

Adrian Tracy: 35.5", 4.34, 0 games missed, 30 sacks in 47 games (all starts. SRAM 0.64 if there aren't any adjustments I'm not aware of). Pretty similar to Moats on all of the SackSEER measures.

Dexter Davis: 35", 4.30, 0 games missed, 31 sacks in 50 games (SRAM 0.62?). Ditto.

I might take the vertical jump numbers on those two with a grain of salt, though, because they're accompanied by extremely poor broad jumps.

For the sake of full disclosure, I should note that the guy we were talking about as a late-round sleeper last year was Philip Hunt.

91
by Bobman :: Fri, 04/23/2010 - 1:07am

With the 31st pick in the NFL draft, the Indianapolis Colts select Jerry Hughes FTW.

He can also play backup RB if need be.

My Giant-loving family back in Jersey just got sent this article and Pierre Paul's blurb. Not gonna be happy. Unless they like back-flips.

93
by chemical burn :: Fri, 04/23/2010 - 10:36pm

I know this article is old, but let me just say, as an Eagles fan, I'm so happy this article is up here on FO - I can check on their 1st and 3rd picks, players that I don't know anything about (aside from convetnional wisdom - i.e. unsubstantiated bullshit) because I don't follow college ball. Glad they got one of the steals and also intrigued about the discussion of Gibson in the comments (here's hoping his height/weight ratio is actually hurting his vert and throwing off his projection...)

Anyway, this is such great stuff anf you can't get anything as interesting and intelligent about football anywhere else on the web... great, great stuff - even if it ends up not prophetically accurate...

94
by Mansteel (not verified) :: Fri, 04/23/2010 - 11:08pm

I am a big Giants fan and therefore I submit that this whole "Sackseer" thing is sheer nonsense. JPP is the best player in the draft and destined for the HOF--the Giants beat writers said so...or at least implied so.

95
by Jerry :: Sat, 04/24/2010 - 1:47am

Nate,

Any chance of putting up numbers for the guys in the first two rounds who didn't make the article?

Thanks.

97
by purplerain (not verified) :: Tue, 04/27/2010 - 9:06am

re: Kindle. Now that we know he's a Raven, I wonder if he found the right team for him to succeed. Looks like he *will* play on the strong side, with Suggs on the other side taking on the LT. With wide bodies Mount Cody and Nata absorbing blockers, and the addition of Corey Redding to spell the aging Pryce, Kindle might be in the best situation for him, given the weaknesses SackSEER exposes. Also, slipping down to the second round hedged the risk factor -- he'll certainly cost a lot less than Jason Pierre-Paul and Derrick Morgan. Got to give props to Bill Polian -- apparently if the Ravens hadn't traded out of #25, they were going to take Jerry Hughes, and they tried to trade back into the first round to get him, and obviously failed.

98
by AlanSP :: Tue, 04/27/2010 - 6:50pm

Just to post the relevant numbers on players that weren't mentioned in the article who were taken in round 2:

Koa Misi: 38" vert, 4.27 shuttle, 14 games missed, 10.5 sacks in 38 games

Jermaine Cunningham: 35" vert (pro day), 10 games missed, 18.5 sacks in 45 games, didn't do the shuttle at the combine or his pro day

Jason Worilds: 38" vert, 4.29 shuttle, 15 sacks in 39 games*, 2 games missed*, junior

*Worilds played in 2 games his freshman year before being injured and was granted a medical redshirt. If you count those 2, the total is 41 games. Since the criterion is games of NCAA eligibility missed, those wouldn't count as missed games, and he gets extra credit for coming out as a junior. Personally, this seems like sort of a silly way of describing things. If Worilds had missed, say, 6 games with an injury his freshman year instead of 11, and then played the next 3 years exactly as he did, he'd count as a senior with 9 games missed rather than a junior with 2 games missed. In essence, his projection improves because he missed more time and got a redshirt (if I understand the description of SackSEER correctly)

I'd also note that, like Adrian Tracy and Dexter Davis, Worilds had a good vertical jump accompanied by a poor broad jump.

99
by SackSEER :: Wed, 04/28/2010 - 10:41pm

Hey AlanSP,

Just a quick hit on the Worilds comment. There is no extra credit for coming out as a Junior. The missed game metric is calculated based on games missed, not how much eligibility the player had left at the end of his college career. So, with Worilds, for example, as you note, he had a medical redshirt his first season after playing two games (since Virginia Tech had 13 games left, he accrues 11 missed games for his medical redshirt). He also missed two other games after coming back from the first injury, so he has a total of 13 missed games.

Also, I am not checking this thread regularly, so if you want to hash this out a little more, email me offline at sackseer at gmail dot com.

100
by Dave W (not verified) :: Tue, 11/30/2010 - 12:33am

Well, Jason Pierre Paul had two of his projected five career sacks yesterday. Moreover, the comparison to other JUCO DEs is poor logic- you've set this up: All JUCO DEs are failures, JPP is a JUCO DE, JPP will be a failure. Obviously that's not a sound argument at all, as there about 1000 other factors that go into success.

Anyways, nice try at coming up with a statistical projection model, but it just won't work when it comes to sacks (and in any event, since when is the sack the be-all, end-all stat for determining whether a player is successful or not? The system that an OLB plays in is the ultimate factor in sacks, and 3-4 DEs generally aren't asked to rush the passer.).

101
by Jack Neefus (not verified) :: Sun, 12/05/2010 - 7:39pm

To revisit Pierre-Paul once again, he just surpassed his 5-year sack projection of 3.8 in his 12th game -- playing part-time. And has three forced fumbles, a recovery, and an interception (called back) as well.

I suspect the model has problems in giving no credit for JUCO experience, but full credit for small-college playing time and seasons spent playing out of position.

Pierre-Paul was labeled a workout warrior, but his workout numbers were actually average. The Giants' scouting department has said that he was doing thing ON the football field they had never seen anyone else do.

The article may be correct in saying:

"Overall, the general manager who pulls the trigger on Pierre-Paul better be very confident that he has something special -- so special that it will completely buck the historical trends."

That is exactly what Jerry Reese did.

102
by Joe H (not verified) :: Tue, 12/07/2010 - 5:57am

Jason Pierre Paul just sacked SackSEER. Nice try, but SackSEER was just silly to begin with. And to make SackSEER look even more ridiculous, Bill Polian just came out and admitted drafting Jerry Hughes was a mistake. Yikes!

103
by Cr (not verified) :: Tue, 12/07/2010 - 2:09pm

Jason Pierre Paul has another thing going for him that no number crunching predator can account for... He has too accomPlished monsters in Osi and Tuck that will help his development tremendously ... They will also help open things up for him on every play...I'm a diehard Giants fan and I was upset when the giants drafted him at first but have come to learn not to question the giants when it comes to drafting pass rushers. Seeya sackseer!!

104
by Cr (not verified) :: Tue, 12/07/2010 - 2:11pm

Oops... Two

105
by Jim Z. (not verified) :: Sun, 03/06/2011 - 9:23am

I think that pronouncements of Jason Pierre-Paul "proving Sackseer wrong" are probably premature.

Did Pierre-Paul accumulate more sacks than he was projected to this past year? Yes. However, if you watch every one of his sacks closely, you find that the majority of them were achieved either on stunts or when he was unblocked. He only cleanly beat his man, I think, twice.

Also, you certainly have to factor in the other talent on that defensive line that Pierre-Paul has the benefit of playing with. Would he be successful on a defensive line lacking Justin Tuck, Barry Cofield, and Osi Umenyiora, all of whom demand extra attention from blockers?

The real test of the SackSEER system will be the 2011 season, if there is one. Another year will need to pass before we can truly give these DEs an opportunity to prove their worth. If Pierre-Paul is given a starting job and, say, accumulates 2-3 sacks in an unproductive campaign, I wouldnt' necessarily be surprised. Similarly, if Jerry Hughes or Brandon Graham explode in their sophomore season and put up double-digit sacks, I wouldn't be surprised either. Such results (produced after sufficient time in the NFL acclimating to the environment) would obviously do more to validate SackSEER than these DE's rookie seasons.

111
by NFLgotya (not verified) :: Wed, 12/21/2011 - 5:20pm

WOW, swing and a miss

106
by WDave (not verified) :: Tue, 09/13/2011 - 6:55pm

2 more sacks for JPP. Once again, the system needs revamping. Saying that JUCO guys missed games is the dumbest assertion I've ever heard- it's completely different from an injury.

107
by Bob Mrosko (not verified) :: Mon, 09/26/2011 - 10:48am

Why, it's Jason Pierre-Paul with 4.5! Who knew?

108
by jpp is a brats (not verified) :: Wed, 09/28/2011 - 6:58pm

This article is a joke, jerry hughes is a bust, along with derrick morganand brandon graham, while jason pierrepaul is a perennial probowl talent...allpro left tackles like jordan gross and jason peters had no chance when they had to block him

109
by Dookieface (not verified) :: Wed, 12/07/2011 - 9:10am

So much for SackSEER! BAHAHAHAHAHA.

110
by Mr Shush :: Wed, 12/07/2011 - 11:42am

Mr. Rosenhaus, your dedication to your client's interests is truly impressive, but I would like to gently point out that the draft took place some time ago and his next contract will be determined by his (unquestionably excellent) play on the field, not an obscure projection from his college statistics. Please stop trolling now.

112
by NFLgotya (not verified) :: Wed, 12/21/2011 - 5:21pm

He's not trolling, hes basking in the glory of reality.

113
by nath :: Tue, 01/17/2012 - 4:57pm

Yeah, so, SackSEER was almost literally as exactly wrong as it could have possibly been. Hughes can't get on the field and JPP is one of the best pass-rushers in the league already. Maybe it just plain doesn't work.

114
by Mr Shush :: Wed, 01/18/2012 - 7:08am

To be fair, Hughes came with a disclaimer ("the type of prospect SackSEER sometimes misses on") and was projected to be the best in a weakish year, not an actual great prospect. And the player he was compared to in making this disclaimer, Jason Babin, has of course gone in the intervening years from looking like a horrible SackSEER miss to a horrible waste of talent by his early-career coaches.

115
by Suckit (not verified) :: Thu, 02/23/2012 - 8:30pm

JPP > SackSEER > Idiots who came up with SackSEER

116
by DCW (not verified) :: Sun, 11/04/2012 - 3:21pm

Just curious- if JPP's formula is changed to get rid of the ridiculous assertion that he missed 26 games (newsflash- JUCO kids practice, injured kids don't. Big difference)?