2009 NFL Draft Defensive End/3-4 Rush Linebacker Projections

|

Although the quarterback position receives most of the notoriety for producing first round busts, the 4-3 defensive end/3-4 rush linebacker ("pass rusher") position has had as many, if not more, highly touted players who ultimately disappoint. For example, of all of the quarterbacks drafted in the first round since 2002, 28.5% have made the Pro Bowl, while only 25% of the pass rushers drafted in the first round within that time frame have earned the same distinction. However, with rushing the passer at a premium in the NFL, teams are still willing to roll the dice in the first round on a player who could just as easily be sitting on the bench as he could be sending the quarterback to the turf. Given the uncertainty surrounding the position, any metric that can shed light on the chances of success for pass rushing prospects in the NFL Draft would be welcome.

I combed through the college stats, combine, and pro day data for all of the 4-3 defensive ends/3-4 rush linebackers drafted in the first and second rounds since 2001. My ultimate result was a statistical regression that accounts for approximately 40% of the variation between first round prospects who end up terrorizing quarterbacks on Sundays and those who will end up doing their best Michael Haynes impressions. For instance, the model identified Dwight Freeney, DeMarcus Ware, Bryan Thomas, Julius Peppers, Mario Williams, Aaron Schobel, Shawne Merriman, Jason Babin, and Terrell Suggs as the most likely to become dominant pass rushers at the next level. Although Thomas and Babin have not produced as expected, in contrast, the most highly drafted players from the same time period are far less impressive: Mario Williams, Julius Peppers, Chris Long, Justin Smith, Gaines Adams, Vernon Gholston, Andre Carter, Jamaal Anderson, and Derrick Harvey.

The regression contains four independent variables. In order of importance, they are: the prospect's short shuttle time, the prospect's vertical leap, the prospect's draft position, and the prospect's per game sack production in college. I measured pro success by the total number of sacks that the player is projected to accumulate in his first seven pro seasons based on his production to date.

Although all of the factors in the regression are nearly of equal importance, the most important is the short shuttle. In my dataset, short shuttle seemed to function as a sort of gauge of "bust potential"--the top players often will have mediocre short shuttles, but the worst performers will have well below average short shuttles. Kenechi Udeze had great production in college and a good vertical leap, but he had the worst short shuttle of any first round pass rusher prospect with an absolutely molasses-like 4.73 seconds (as a little bonus for Lions fans, the second worst short shuttle recorded for pass rushers drafted in the first two rounds since 2002 is Ikaika Alama-Francis). As another example, Vernon Gholston, who after a sackless rookie year is widely believed to be a bust, had great workout numbers overall but posted a slightly below-average short shuttle at 4.4 seconds. The lesson from short shuttle seems to be that, no matter how productive a player is in college, no matter how much "vertical" explosion he can bring to bear on opposing offensive linemen, if he lacks the requisite level of agility to outmaneuver opposing offensive tackles after they "lock on," he will not be making Tom Brady or Peyton Manning's lives difficult anytime soon.

Nearly as important as short shuttle in the regression is the vertical leap. Of the top 5 first round pass rushers who have been drafted in the first round of the NFL Draft (Shawne Merriman, DeMarcus Ware, Dwight Freeney, Aaron Schobel, and Mario Williams), all achieved at least a 37" vertical leap at the NFL combine or at their pro day, which is well above average.

The final factor is a metric that I call "adjusted college sacks/game." Basically, it's calculated by dividing the prospect's sacks in college by games played without counting any sacks or games played during the prospect's freshman or redshirt freshman year. The adjustment is useful because some pass rushers early in their college careers receive a lot of token playing time and thus have their per game sack production in college unfairly reduced by a large "games played" denominator. Although straight college sacks/game also correlates to NFL success, the correlation of adjusted college sacks/game was much stronger, so I used that for my model. All in all, it's no surprise that the ability to sack the quarterback in college is a good indicator of ability to sack the quarterback in the NFL.

Anyway, enough with the boring stats stuff--on to the projections! I am currently using the ProFootballWeekly top 100 list for the "projected draft position" factor because it is the most accurate measure of player value that is currently and freely available. When Rick Gosselin of the Dallas Morning News releases his top 100 list, which is based on numerous interviews with NFL scouting departments, I will update my projections.

Aaron Maybin, DE/OLB Penn State
Projection: 45.85 Sacks by Year 7
Projected Draft Position: #13 Overall
Vertical Leap: 38" and 40"
Adjusted College Sacks/Game: 92.3%
Short Shuttle: 4.38 Seconds

Make no mistake, Aaron Maybin has a monster projection, which puts him nearly at Terrell Suggs and LaMarr Woodley levels. Aaron Maybin's projection is all the more impressive considering that the model takes into account that an otherwise good looking prospect can possibly go all "Vernon Gholston" once he enters the NFL. Maybin's second vertical leap is particularly laudable, only Mario Williams and Shawne Merriman were able to match or exceed it. Maybin's vertical leaps corroborates the praise that he has received for being the most explosive pass rusher with the best first step in the 2009 NFL Draft. Maybin's agility drills, although actually just a hair below average based on historical levels, were among the tops in his position at the combine, with only Connor Barwin, Ian Campbell, and Michael Johnson scoring better.

There is also much that the model suggests about Maybin's chances to break the curse attached to pass rushers drafted out of Penn State: as Courtney Brown, Michael Haynes, and most recently, Tamba Hali have achieved various levels of busthood. Although I did not look at Courtney Brown, the projections for Michael Haynes or Tamba Hali give little cause for concern. Michael Haynes had a bad vertical leap at 30.5", mediocre college production, and no short shuttle time. Assuming an average short shuttle time, the model would have projected Michael Haynes to severely underperform his draft status with 26.71 sacks in his seventh year, which he underperformed by approximately 20 sacks by logging 5.5 sacks. Tamba Hali, on the other hand, was drafted late in the first round, had an even worse vertical leap, poor college production, and an average short shuttle. The model would have projected Tamba Hali as a bust, with a projection of 22.19 sacks in his seventh year, and although he was not especially good on a defense that absolutely could not pressure the quarterback last year, he has actually been a pretty useful player and is on pace to register 42.76 sacks by his seventh year. So basically, one Penn State pass rusher has overperformed, one has underperformed, and another pass rusher's career was derailed by a series of devastating injuries before it really got started. There is no intrinsic reason that Penn State pass rushers should not be able to make the transition to the NFL; it's not like Penn State faces teams without good offensive lines.

By far the greatest mainstream concern with Aaron Maybin is his lack of games played (26) which tends to create a "sample size" issue for him, where essentially his outstanding Sophomore season sack numbers become his "adjusted sacks/game" factor. Although there is no significant correlation between games played and success in the NFL, there is not a great historical analogue in any of my data for Maybin's relative inexperience, as the only two first round pass rushers who started less than the 30 games had poor projections notwithstanding their lack of college playing time. Still, Maybin's short college career is a legitimate concern.

On the other hand, how many first round pass rushers amassed as many as 12.0 sacks in their second season in college? Three: Dwight Freeney, Julius Peppers, and Mathias Kiwanuka. That is not a bad list.

And concerns about Maybin's size? DeMarcus Ware weighed in at 251 lbs. at his pro day, 1 lb. less than Maybin weighed in at his. How did that work out for DeMarcus Ware?

Brian Orakpo, DE/OLB, Texas
Projection: 37.61 Sacks by Year 7
Projected Draft Position: #12 Overall
Vertical Leap: 39.5"
Adjusted College Sacks/Game: 62.34%
Short Shuttle: 4.45 Seconds

Back in January, I offered the pithy, but apparently unoriginal phrase "say no to Orakpo", based on my belief that his lack of college productivity would doom him to NFL failure. Orakpo's projection is only slightly above that expected of a pass rusher drafted in the first two rounds, which suggests that he may be slightly overvalued as a high first round pick.

There are significant concerns in Orakpo's data that bear mentioning. Orakpo's short shuttle time of 4.45 seconds is below average. Given concerns that Orakpo has a limited array of pass rushing moves, he may not have the agility necessary to beat offensive tackles that he cannot simply run around or bull rush. Orakpo's college production is also a concern: the best pass rusher in the first round with less than a 65% adjusted college sacks/game is Will Smith, who has been good, but not great. However, Orakpo's 39.5" vertical leap suggests rare explosion that could translate into healthy sack production at the next level.

Larry English, DE/OLB, Northern Illinois
Projection: 33.34 Sacks by Year 7
Projected Draft Position: #43 Overall
Vertical Leap: 36"
Adjusted College Sacks/Game: 80.26%
Short Shuttle: 4.38 Seconds

Although projected to be drafted much later than Brian Orakpo, the model believes that he is a much better value, as Orakpo has only a nominally better projection. His projection is buoyed by his solid all around computer numbers and his excellent production in college.

Interestingly, the model tends to over and under project players from non-BCS schools moreso than their Bowl Championship counterparts. This is probably because NFL scouts have difficulty properly grading pass rushers who face-off against less than elite offensive tackles in college. Thus, as a "boom or bust" prospect, Larry English would probably be a bad choice for a team like the Detroit Lions, who need to hit on all of their picks, and probably be a great second round pick for a team like New England, who can afford to "gamble on greatness."

Michael Johnson, DE/OLB, Georgia Tech
Projection: 30.47 Sacks by Year 7
Projected Draft Position: #39 Overall
Vertical Leap: 38.5"
Adjusted College Sacks/Game: 43.24%
Short Shuttle: 4.37 Seconds

I was a bit worried before I ran the numbers that Michael Johnson's gaudy combine workout would inflate his projection, while his well-reported lack of desire and work ethic would lead him to become an NFL bust. However, Michael Johnson's dearth of competitiveness showed up on the field where, despite his tools, he garnered an unimpressive 43.24% sack per game rate in his final three seasons. Michael Johnson's upside is probably Kamerion Wimbley, a player who posted good combine numbers and achieved some success in the NFL despite his gross lack of sack production in college.

Cody Brown, DE/OLB, Connecticut
Projection: 28.65 Sacks by Year 7
Projected Draft Position: #60 Overall
Vertical Leap: 36.5"
Adjusted Sacks/Game: 71.88%
Short Shuttle: 4.4 Seconds

No Everette Brown or Robert Ayers, yet? Oh my! Cody Brown, who is projected to be taken at the tail end of the second round, has a stronger projection than either, much more highly touted player. Cody Brown could be a great "under the radar" prospect with high character to go along with his good college production and passable short shuttle time.

Clint Sintim, OLB, Virginia
Projection: 28.29 Sacks by Year 7
Projected Draft Position: #31 Overall

Vertical Leap: 34.5"
Adjusted Sacks/Game: 64.86%
Short Shuttle: 4.42 Seconds

To his credit, Clint Sintim is the only high-level pass rushing prospect who actually played in a 3-4 defense in college, and he has been linked to the New England Patriots as a potential second round pick. However, Sintim is a similar prospect to Larry English, and scores just a little bit lower than he did on all of the relevant metrics besides draft position. Similar players include DeWayne White and Matt Roth.

Everette Brown, DE/OLB, Florida State
Projection: 23.71 Sacks by Year 7
Projected Draft Position: #20 Overall
Vertical Leap: 31.5" and 31"
Adjusted College Sacks/Game: 76.92%
Short Shuttle: 4.53 Seconds

Perhaps the most fun part of devising a system like this is predicting who is going to bust. Everette Brown was a productive pass rusher in college, but based on a poor short shuttle time and not one, but two very poor vertical leap scores, the model thinks that he will suck out loud in the pro's.

Certainly, there are pass rushers who were productive in college, had below average computer numbers, and succeeded. However, none of them quite fit Brown. Terrell Suggs had a poor vertical leap at 33", but Terrell Suggs also had a godlike adjusted college sacks/game rating of 136%, as well as an average short shuttle. There are other analogues as well, like Jared Allen with his 33" vertical leap or Mathias Kiwanuka with his 32" vertical leap. However, again, Jared Allen had better sack production in college and an average short shuttle time, and Mathias Kiwanuka also had better sack production in college and recorded a great short shuttle time. Thus, there is a clear road map for pass rushing prospects to succeed despite a poor vertical leap score: they have no problem with the short shuttle and they were great, not just good in college. Everette Brown simply just doesn't fit the bill.

In fact, there is not a lot of evidence that a pass rusher with a 31.5" vertical leap can be a better than an average sack artist. The best pass rushers with a vertical leap of 32" or less that I'm aware of are Mathias Kiwanuka, Tamba Hali, Charles Grant, Darren Howard, Clark Haggans, Patrick Kerney, and promising but young Tennessee Titan Jason Jones. Kerney is very good, and many of those players are very solid, but it is not a particularly impressive list of players considering that it represents Everette Brown's upside.

Even more damning is Everette Brown's short shuttle time. There is a short list of first and second round pass rushers who had a short shuttle time worse than 4.5 seconds: Jerome McDougle, Calvin Pace, Kenechi Udeze, Dan Cody, Ikaika Alama-Francis, and Dan Bazuin. That is a pretty sad group: these players have a total of only 37 career sacks among them.

Of course, Everette Brown could still beat the odds and set the NFL on fire with his "impressive array of pass rushing moves," but to do so, he has a lot of history to beat.

Lawrence Sidbury, DE/OLB, Richmond
Projection: 19.45 Sacks by Year 7
Projected Draft Position: #64 Overall
Vertical Leap: 35"
Adjusted College Sacks/Game: 48.57%
Short Shuttle: 4.43 Seconds

Regarding Sidbury, Mel Kiper, Jr. recently went on record as saying "who's to say that he won't be better than Brian Orakpo?" Well, I'll say it: barring something horrible happening to Orakpo, it is highly unlikely that Sidbury will out produce Orakpo. Sidbury is the classic prospect who gets overhyped as a "sleeper" and then "sleeps" through his NFL career. Sidbury was not productive at all until his senior season at college, despite a very low level of competition.

Robert Ayers, DE/OLB, Tennessee
Projection: 9.07 Sacks by Year 7
Projected Draft Position: #26 Overall
Vertical Leap: 29.5"
Adjusted College Sacks/Game: 20.51%
Short Shuttle: 4.49 Seconds

As anyone who knows me can attest, I love me some Mike Mayock, and 99% of the time his list of the top prospects in the NFL Draft is beyond reproach. However, the remaining 1% of the time he'll do something completely batty like putting Chris Houston as the number nine overall prospect in the Draft. Chris Houston over Patrick Willis? That didn't make a lick of sense at the time, and it makes even less sense now. There's a reason that that article isn't posted on NFL.com anymore and I have to link to it through archive.org.

In this tradition, the otherwise esteemed Mr. Mayock has anointed Robert Ayers as number 5 overall prospect in the NFL Draft. At the time he released this list, I was highly skeptical that a player who played 48 games in college and only recorded 9 sacks could merit a first round pick based on his pass rush ability. What was happening in those games? How can a defensive end with so few actual sacks grade out so well on tape? Did he just knock the opposing quarterback over fifteen times per game but just missed the sack all but nine times over the course of his career? Well, now that I have fired up my Excel Spreadsheet and given the web site NFL Draft Scout about a bagillion hits, I'm ready to call Mayock's bluff.

Where to start? Robert Ayers's college sack rate of 20.51% is the worst ever for a first-round pass rusher. The closest is Tyler Brayton at 32.43%. Moreover, his vertical leap is better than that of only one first round prospect: Derrick Harvey's 28.5". Robert Ayers's best data point is his short shuttle, but that doesn't mean that it is to his credit: Robert Ayers only beats out the luminaries that I listed in the Everette Brown analysis.

So how bad is Robert Ayers's projection? Of the 50 players that are included in the regression, only one player has a had a worse projection, which is impressive considering that players get a bump for being projected to go in the first round. The one player who projects worse is Detroit Lion Ikaika Alama-Francis, who projected to have 8.68 Sacks by Year 7. Ikaika Alama-Francis, so far in his young career, has a grand total of one sack, and as I have detailed more fully before, he has even had trouble staying on the active list.

Short shuttle, vertical leap, and college production also correlate to NFL success even in the later rounds of the draft. As a few examples, Giants star Justin Tuck scored well in all three metrics, Trent Cole had a good vertical leap score and shuttle, and Jared Allen and Shaun Phillips each posted great college production and good to great short shuttle times.

Conclusion

NFL teams would be well-served to pay much closer attention to each pass rushing prospect's short shuttle times, vertical leap scores, and per game production in college. Although this model is not perfect, and I fully expect at least one or two players discussed here to significantly over or under perform their projections, it can still be a useful tool in player evaluation. So the next year when you see one of the NFL's newest defensive ends sacking the quarterback, don't be surprised if his ability to move quickly in 10 yard bursts or his ability to jump very, very high had something to do with it.

4/21/09 Edit: AlanSP pointed out some additional successful players with low vertical leaps, so I added them to the Everette Brown section. Also, for some reason I thought that Jason Jones was a Houston Texan, as was pointed out in the comments.

5 comments:

Misfit said...

Tyson Jackson has been getting a lot of attention recently. How do you project him?

Andre WareAreTheyNow? said...

His Misfit,

Thanks for commenting! I actually did not look at Tyson Jackson because I see him as either a 3-4 DE or a 4-3 DT, and the system really isn't designed to project those guys (I took out guys like Seymour from the model, which I think is a better analogue for Jackson).

That said, if some team were to take him and thought that he could just come off the edge like Dwight Freeney, and were willing to bench/cut him afterwards, he would project to only 3.5 sacks by Year 7. Now, I would think that he will probably drastically outperform his projection, because usually guys who are drafted to rush the passer get shoved out the door with that production, but Jackson's value will be stopping the run, and he'll get enough playing time to accidentally stumble towards the quarterback a couple of times ;).

les steckel said...

Wouldn't sack projections be more accurate if you used sacks per pass attempt as opposed to sacks per game? A DE or OLB in a pass happy conference like the Big 12 would probably get more sacks per game than a similar player in the Big 10 or SEC.

Dales said...

http://www.profootballtalk.com/2009/04/20/the-first-mr-x-mock-draft/

Pro Football Talk has a 'new' mock draft up, made by some insider who supposedly consults with multiple NFL scouts.

They have Ayers going to... the Lions at pick 20. :-/

Andre WareAreTheyNow? said...

Les Steckles,

That could potentially be a more accurate way to do it, although finding box scores for all of those college games would be quite a task!

As far as the Big 12, interestingly, all of the Big 12 players in my model were overprojected except for Justin Smith. However, Smith was the only one who had a respectable adjusted sacks/game, so it is not like their projections were inflated because they racked up lots of sacks.

And Dales, I saw the Mr. X mock and it made me shiver. Just a little.

 

©2009 The "Goodbye, Ladies" Draft Report | Template Blue by TNB