SackSEER -- Let's Try This Again

Guest Column by Nathan Forster
In Football Outsiders Almanac 2010, I introduced SackSEER, a model that projects the sack totals of highly drafted 4-3 defensive ends and 3-4 outside linebackers in their first five years in the NFL. SackSEER is composed of four metrics: the prospect's vertical leap, short shuttle time, per-game sack productivity in college (with certain adjustments), and missed games of NCAA eligibility. This particular blend of the prospect's athleticism, production, and the injury concerns and/or off-the-field issues, represented by the missed games metric, would have identified edge rushing superstars such as Mario Williams and Shawne Merriman and busts such as Robert Ayers and Jarvis Moss.
Although it is too early to enter any definitive judgment on last year's draft class, there is no question that this was a rough year for SackSEER. SackSEER's most highly projected edge rusher, Jerry Hughes, played little in 2010; and SackSEER missed on whatever transformed Carlos Dunlap from a healthy scratch at midseason into the most productive rookie edge rusher by year's end.
Most notably, however, SackSEER's controversial 4.5 sack projection for Jason Pierre-Paul now looks silly. Pierre-Paul met his five-year projection in the space of just his rookie year as part of a heavy rotation with the prolific Justin Tuck and Osi Umenyiora. With 4.5 sacks as a rookie, Pierre-Paul is on pace to be slightly more productive than the average highly drafted edge rusher.
Although his present expected value is well short of a DeMarcus Ware or a Clay Matthews, even an average level of career production from Pierre-Paul is a significant surprise to SackSEER. It is still not clear exactly how good Pierre-Paul will be (SackSEER was similarly down on Tony Bryant, who also had 4.5 sacks his first year and ended up with 18.5 in his first five years), but if he develops into the player that the New York Giants apparently believe that he will, it might be a good opportunity to revisit the model to try to find out where SackSEER went wrong.
The real culprit in Pierre-Paul's projection was the huge disconnect between his much-hyped athleticism and his actual performance at the NFL Combine. Pierre-Paul's scores for his vertical leap and short shuttle run were both a full standard deviation below the mean performance for edge rushers. Although Pierre-Paul did perform better in the other Combine drills -- his broad jump and bench press were also below average, but his 40-yard dash and his 3-cone drills were slightly above average -- his workout performance fell afield of what one would expect from a player who was hyped as the next Jevon Kearse. Pierre-Paul appears to be a rare player who is not only extremely athletic, but extremely athletic in a way that does not translate to the Combine.
However, 2010 wasn't all bad news for SackSEER. The other six projections are still on pace to be solid, and the projections for Brandon Graham and Jermaine Cunningham are especially close -- both are on pace to be accurate within two sacks. Moreover, Jason Babin, a longtime SackSEER miss, finally became the player that SackSEER always thought that he would be, recording 12.5 sacks and securing a Pro Bowl nod. Finally, SackSEER would have identified Cameron Wake, who emerged from the obscurity of the Canadian Football League to contend for the NFL's sack crown, as a significant sleeper. Although he went undrafted, Wake had the best pre-draft SackSEER workout since at least 1999, posting a 4.19 shuttle and a 45.5-inch vertical leap.
SackSEER's read on last year's draft class was that the highly drafted edge rushers would be below average. This year, however, SackSEER believes that this class of edge rushers deserves the hype and includes an elite prospect. Remember that SackSEER does not apply to 3-4 defensive ends, so you will not see any projections for Allen Bailey, Adrian Clayborn, Cameron Heyward, Cameron Jordan, or J.J. Watt here.
Von Miller, Texas A&M
Vertical: 37.0", Short Shuttle: 4.06, SRAM: 0.76, Missed Games: 4
Projection: 36.4 Sacks by Year 5
SackSEER loves Von Miller, and it is easy to see why. Miller's vertical leap of 37 inches is good, but if anything, it understates his ability to explode. The 40-yard dash and the broad jump scores are, historically, closely correlated with vertical leap performance, and Miller's 4.53-second 40 and his 10-foot-6-inch broad jump actually exceed the mean performance of highly drafted edge rushers by a greater margin than his vertical. Miller's 4.06-second shuttle is amazing -- no edge rusher prospect has run a shuttle at the Combine that starts with a 4.0 since DeMarcus Ware.
But Miller's Combine performance distract from his outstanding production. Miller was not used as a hybrid linebacker/defensive end until his junior year, so the first half of his college career gets a considerable boost from SRAM's positional adjustments. Miller's four missed games are a result of the typical difficulties that freshmen players have breaking into the lineup and are likely a non-issue. Miller has played through all of the injuries that he has suffered in college and has no known off the field or character issues. He even has a cool looking pair of glasses, which is not necessarily predictive of sacks but is highly predictive of awesomeness.
A skeptic, however, would say that SackSEER doesn't tell us much about Miller that we don't already know. We already know that Miller is explosive, quick, productive, and clean off the field, but SackSEER does nothing to address concerns about Miller's size and his ability to hold the point against the run. Size at the edge rusher position has been tricky. Prospects with good size and good SackSEER projections rarely bust, and there have been plenty of players such as Aaron Maybin and Manny Lawson who end up playing down to their size despite impressive athleticism. However, some of the best edge rushers have been undersized, and often severely so. Most recently, Clay Matthews took the NFL by storm despite weighing only 240 pounds at the Combine, and Trent Cole and Robert Mathis have been outstanding despite being well south of the 240-pound mark on draft day.
Although an injury or struggles against the run could certainly derail Miller's career, Miller has the potential to become an elite player at his position.
Justin Houston, Georgia
Vertical: 36.5", Short Shuttle: 4.37, SRAM: 0.61, Missed Games: 3
Projection: 26.0 Sacks by Year 5
Houston has quietly built himself into a solid prospect. He has had steady production, demonstrating consistently better numbers each season on the way to recording more than 20 sacks in three years in the ultra-competitive SEC. Houston also played in nearly every game of his Georgia career, save for three games that he missed due to a team suspension. Houston's workout, much like his Georgia career, was steady and solid, with a good vertical and an average shuttle.
Houston's combine performance is all the more impressive in the context of Houston's 270-pound weigh-in, which was 12 pounds heavier than his listed weight at Georgia. Although Houston was one of the few edge rushers in this draft class invited to the Combine as a linebacker, rather than as a defensive lineman, he is probably better suited to play as a 4-3 defensive end than most of his contemporaries. Houston is a great value pick that SackSEER likes considerably more than most projected first-round picks.
Ryan Kerrigan, Purdue
Vertical: 33.5", Short Shuttle: 4.39, SRAM: 0.70, Missed Games: 1
Projection: 24.7 Sacks by Year 5
Kerrigan gave you about everything you could ask for on the field: He only missed one game during his college career and recorded 33.5 sacks. The question with Kerrigan has always been whether he has the athleticism to translate his college production in the NFL. SackSEER actually has a relatively complicated answer to this question. Although Kerrigan's 33.5-inch vertical leap is an inch below average for the position, there are a number of reasons to be more bullish on Kerrigan's athleticism than his SackSEER metrics suggest.
First, Kerrigan recorded a confirmed 35-inch vertical leap in high school, and it is doubtful that he has lost explosion after four years in a major college program. Second, Kerrigan is one of those uncommon players who underperforms on the vertical leap but overperforms on the broad jump. The broad jump is actually so closely correlated with the vertical leap that there is some uncertainty as to which is the more predictive metric. The vertical leap is the clear leader among highly drafted edge rushers, but the broad jump rallies strongly enough in the later rounds to leave open the possibility that the broad jump could prove to be a slightly more predictive metric for pass rushing success in the future. Kerrigan's short shuttle is only slightly below average, which is probably a win for him considering the questions concerning his flexibility and lateral agility leading up to the Combine.
Da'Quan Bowers, Clemson
Vertical: 34.5, Short Shuttle: 4.45, SRAM: 0.60, Missed Games: 2
Projection: 22.0 Sacks by Year 5
Just a month ago, Da'Quan Bowers was considered a genuine candidate to be the first overall pick. Bowers, however, has not fared well since being placed under the microscope. There are questions about his production, his athleticism, and his knee. Although SackSEER often breaks with conventional wisdom, SackSEER agrees that Bowers is a high-potential prospect with some serious question marks.
Bowers dominated the ACC with 15.5 sacks last year, but recorded only 4.0 sacks in his previous two. Unfortunately for Bowers, the best historical analogue for his sack pattern is Jamaal Anderson. Like Bowers, Anderson was a huge defensive end at 288 pounds, and like Bowers, Anderson had exactly four sacks his first two years before ripping off a 13.5-sack performance against quality competition.
Bowers' inconsistent production would be a little easier to swallow if his workout numbers jumped off of the page, but they are mediocre at best. Although his recent knee injury could provide a credible excuse for his performance, his knee injury is a legitimate concern in and of itself. Knee injuries put a quick end to the career of Andre Wadsworth, and other highly drafted edge rushers with some history of knee injuries include players such as Alonzo Jackson, Dan Bazuin, and Anton Palepoi, who each had short NFL careers.
Bowers should be an incredible prospect: He is a 280-pound end who played through injuries and led college football in sacks as a junior. However, due to inconsistent production and so-so workouts, Bowers has only an average projection, and he probably belongs in the mid-to-late first round rather than the Top 10.
Aldon Smith, Missouri
Vertical: 34.0", Short Shuttle: 4.50, SRAM: 0.62, Missed Games: 3
Projection: 20.0 Sacks by Year 5
Smith is certainly an odd prospect. After accomplishing the unprecedented feat of recording 11.5 sacks as a freshman, Smith declined as a sophomore, recording only 5.5 sacks during an injury-plagued season. Smith's production on the whole is good -- few prospects record 17 sacks in their first two years of college. However, add in a below average Combine performance, and SackSEER feels that Smith is a slightly below average project relative to his projected draft position.
There is an additional red flag for which SackSEER does not (at least not yet) account -- the short but infamous history of redshirt sophomore edge rushers selected in the Draft. There have been only two redshirt sophomore edge rushers selected in the Draft since at least 1999: Aaron Maybin and Paul Kruger, both in 2009. Maybin and Kruger have given their teams almost no production -- they have exactly one regular season sack between them -- and both dramatically underperformed their SackSEER projections. Of course, the sample size here is too small to tell if this is a legitimate trend or the happenstance of two disappointing players who share the distinction of entering the draft uncommonly early, but it is enough to give pause before spending a high pick on a player with inconsistent production and below average agility scores. The concerns associated with redshirt sophomores, consequently, apply equally to ...
Robert Quinn, North Carolina
Vertical: 34.0", Short Shuttle: 4.40, SRAM: 0.56, Missed Games: 13
Projection: 15.5 Sacks by Year 5
Robert Quinn, who missed his entire junior year due to a suspension by the NCAA for accepting benefits from an agent in violation of NCAA rules, is quite possibly the most inscrutable edge rusher prospect that SackSEER has ever seen.
Let's start with Quinn's missed games. In order to be consistent with SackSEER's missed game metric, we need to shave more than seven sacks from Quinn's projection. A season-long suspension is unprecedented -- there is not a single edge rusher prospect in my database who has been suspended for more than a handful of games. The closest analogue to Quinn is probably Trent Cole, who missed a season of eligibility, along with many others, after running afoul of the NCAA's then-controversial Proposition 48 rules for academic performance. Cole certainly turned out all right, so Quinn's missed games are probably not the detriments that SackSEER thinks they may be.
Nor is Quinn's production during college particularly illuminating. Quinn is essentially a one-year wonder, and even edge rushers who collect sacks at a high rate early in their careers can quickly become football versions of the Royal Tenenbaums. A great recent example is George Selvie, who had an even better sophomore campaign than Quinn, recording 14.5 sacks in 13 games. Selvie faded down the stretch, recording only 8.5 sacks in his last 25 games, and was ultimately selected as a seventh-round afterthought by the St. Louis Rams. On the other hand, Quinn has been renowned for his sophomore game tape and was credited with an absurd number of hurries during that season. Hurries are graded inconsistently from team to team, so it is impossible to tell if Quinn was more dominant as a sophomore than his sack numbers would indicate, or if the North Carolina scorekeepers were a bit overzealous when it came to Quinn.
Are you feeling ambivalent about Quinn yet? Quinn's workout numbers are equally confounding. He had a mediocre workout at the Combine, with an average 34-inch vertical and a below average 4.40 shuttle, which is 1.5 inches short from being the exact same workout posted by Vernon Gholston. However, because standing on his Combine numbers would be far too simple, Quinn also worked out at his Pro Day, where he recorded a 4.26-second shuttle but lost an inch off of his vertical leap. He also recorded an amazing 10-foot-5.5-inch broad jump, which is 7.5 inches farther than his broad jump at the Combine. The disparity between Quinn's Pro Day vertical and his broad jump is huge: It is larger than all but five out of the 250 edge rushers drafted since 1999. Historically, numbers for "redo" drills like Quinn's have not had any predictive value, but there is no particular reason to expect this trend to continue.
Given all of the asterisks for Quinn's projection, if we were setting betting lines for Vegas, we would probably take him off the table all together. You probably should not throw your copy of Football Outsiders Almanac 2010 at your TV if your favorite team drafts Quinn, but rather, hope against hope that your team conducted an extremely meticulous analysis of Quinn's game tape and background before taking the plunge. SackSEER grades Quinn as a below average prospect, but it remains to be seen if his projection is a sage warning for teams to be wary of a player with a short record of production and inconsistent workouts, or meaningless mathbabble.
Brooks Reed, Arizona
Vertical: 30.5", Short Shuttle: 4.28, SRAM: 0.34, Missed Games: 5
Projection: 15.1 Sacks by Year 5
Brooks Reed and Clay Matthews are both edge rushers with long-flowing blond locks, but Reed will likely prove to be markedly less Thor-like than the super-powered Matthews. Aside from the bench press and the 20-yard split, Reed failed to meet or exceed any of Matthews' excellent numbers from the 2009 Combine. Matthews and Reed do have similar SRAMs, but Matthews had a much better excuse for his low production. Matthews initially struggled to crack a talented lineup of linebackers at USC until the team finally deployed him as a hybrid defensive end/linebacker during his senior year. Reed, on the other hand, was anointed a starter as a full-time defensive end by his sophomore year, but failed to record more than eight sacks in any year. His two-sack junior campaign was particularly pedestrian.
Reed has received some hype for his 1.54-second 10-yard split, which was the fastest among edge rushers at the Combine. This year there has been a lot of pre-Draft chatter concerning the importance of the 10-yard split for edge rusher prospects. Historically, however, there is absolutely no relationship between a prospect's 10-yard split and his success rushing the passer in the NFL. The 10-yard split is probably more a function of the efficiency of the prospect's "sprinter's start" than the speed with which he can rush the line of scrimmage.
Jabaal Sheard, Pittsburgh
Vertical: 31.5", Short Shuttle: 4.65, SRAM: 0.39, Missed Games: 5
Projection: 10.6 Sacks by Year 5
Much like Pierre-Paul's overtaking of George Selvie a year ago, Jabaal Sheard has emerged as a potential first-round pick. And his previously highly regarded teammate, Greg Romeus, has plunged to late-round consideration after injuries limited him to two games in 2010. Sheard did not demonstrate much explosion at the Combine and struggled considerably at the agility drills during his Pro Day. Sheard also has not been productive despite an abundance of opportunity, recording only 19.5 sacks in four full years of playing time. Sheard's prospects for success certainly are not hopeless, but teams with low first-round picks might be well wise to look to other positions rather than drafting Sheard.
Sleeper
This year, SackSEER generally agrees with the consensus Internet pre-draft rankings for edge rusher prospects, and thus, expects that this draft will lack sleepers at the position. However, there are a couple of exceptions, starting with Dontay Moch.
It is hardly accurate to call Moch a sleeper given the headlines that he has made for his ability to run the 40-yard dash. SackSEER, of course, is more interested in Moch's 42-inch vertical leap, which leaves him explosion to spare. Add to the mix that Moch is a relatively productive player, with 30 career sacks to his credit, and you have a player with potential to bring serious heat off of the edge.
However, there are plenty of reasons to doubt Moch's transformation from workout warrior to NFL sack monster. Moch measures 6-foot-1 and 248 pounds, which is light for even a 3-4 outside linebacker without the potential to add much weight. Moch was productive at Nevada, but his production was hardly eye-popping, especially considering the low level of competition at the Western Athletic Conference.
Moch's upside, however, is tantalizing, and he is absolutely worth a third round pick.
Deep Underground Mining Sleeper
Marc Schiechl hails from the Colorado School of Mines. Schiechl recorded 46.0 sacks for the Orediggers, which is good for a .97 SRAM, higher than every edge rusher in the data set save for Terrell Suggs and Robert Mathis. Although he had only average workout numbers from his pro day -— he had a 35-inch vertical and a 4.50 shuttle -- his freakish production is hard to ignore, even though it came at the Division II level. Although Schiechl is a long shot who may not even get drafted, his profile is similar to Division I-AA prospects such as Mathis and Jared Allen, whose stellar production at small schools foreshadowed similar success at the NFL level.
Deep, Deep Below the Canadian Permafrost Sleeper
If you're looking for a deep sleeper, look no further than the Philadelphia Eagles' relatively unheralded signing of Canadian Football League edge rusher Phillip Hunt. Hunt's career arc is eerily similar to breakout star Cameron Wake's. Hunt was a standout sack artist for the Houston Cougars, notching more 30 sacks, but was not invited to the Combine. Apparently undeterred, Hunt performed admirably at his Pro Day, registering a 41.5-inch vertical leap and a 4.22-second shuttle. As was true with Wake, despite outstanding workouts, Hunt went undrafted due to lack of size.
Like Wake, Hunt sought refuge in the CFL. And again, like Wake, Hunt dominated. After a three-sack rookie season, Hunt recorded 16 sacks for the Winnipeg Blue Bombers. His reward was a contract with the Philadelphia Eagles, adding to a high-potential stable of young edge rushers including Brandon Graham and Daniel Te'o-Nesheim.
When I was developing SackSEER, reader Alan Plotzker (who developed a similar database) and I questioned why a player with such an impressive blend of athleticism and production did not even merit a seventh-round pick. If anything, Hunt was a stronger prospect coming out of college than Wake. Although Wake had better workout numbers, Hunt had much better production and did not have Wake's injury history. So can lightening strike twice? It is unlikely that Hunt will match Wake's success, but if he somehow makes his way into the Eagles' starting lineup, watch out.
Comments
56 comments, Last at 05 Nov 2011, 12:06pm
#1 by lionsbob // Apr 11, 2011 - 1:11pm
Always interesting stuff...and much like the Lewin forecast-it can never be totally accurate...
I remember when you first posted something about this on the message board, since then I have been looking at short shuttle and vertical leaps with a keener eye.
One guy I like as a late round pass-rusher is Ricky Elmore(Reed's teammate at Arizona). He had a 4.32 short shuttle run (4.41 at his Pro Day), a 31.5 inch vertical (34.5 at his pro day), 24.5 sacks as a three year starter (21.5 the last 2 seasons).
#2 by Podge (not verified) // Apr 11, 2011 - 1:14pm
I'm a bit confused about the Quinn write-up. Does the shaving 7 sacks off thing mean that if he'd say declared for the draft last year (I don't know if he could have) the projection would be 22ish sacks?
I'm also curious about Greg Romeus as a prospect. Last year there was talk in mock drafts before he decided to stay in school that the Rams could take him in the 2nd round. Is his injury that bad that he's now nothing more than a late round prospect?
#16 by Nathan Forster // Apr 11, 2011 - 4:23pm
I don't think that Quinn was a redshirt Sophomore, so he would not have been eligible to declare for the draft last year. Although you propose an interesting thought experiment: what to do if a player is suspended and responds by declaring for the draft (or vice versa)? I think that as a practical matter this would not or would be unlikely to happen, but if it did I would have to think hard about how to treat it. Reminds me a little of "cause in fact" and "proximate cause" from law school, if that means anything to anybody.
I understand why Greg Romeus would drop like a rock. Greg Hardy was a much more productive player that Romeus, and he dropped to the fifth round, even though he was at least healthy enough to do pre-draft workouts. There is no 40, vertical, broad, shuttle, 3 cone, or bench press on file for Romeus, who was hardly a consensus first-rounder beforehand. I can definitely understand the upside by taking him with a late third rounder, though.
-----------
Sorry JPP!
#25 by Podge (not verified) // Apr 12, 2011 - 5:16am
I just think it's a bit harsh to knock Quinn's projection for missing games for thinking he was at Auburn (ding!). I think you're right, he wasn't a redshit sophomore (otherwise he'd have come out I suspect). If his missed games was 0 I suspect his project would be a bit higher. But probably still not elite.
Random question that I was going to pose a little later in this wall of text, but what do you consider to be an "average" projection (not adjusting for draft position)? About 20 sacks over 5 years?
Another question, why can't think apply to 3-4 DEs? I know that the position is a different one, in that the expectation of rushing the passer is less than that of a 4-3 DE or 3-4 OLB, and the player himself is less likely to be suited to it (bigger, so likely to be slower and less able to jump miles in the air). That being said, I'm curious if there is much correlation between the SACKSEER metric for college 3-4 DEs and their sack production in the pros? I can see a couple of problems with that, particularly:
1.) You'd need to compare apples with apples, so the average projection would be lower than 20 sacks (if that is the average) and you couldn't compare them to 4-3 DEs or 3-4 OLBs in the same way.
2.) You'd need to apply the eyeball test to look at how they play the run. A 4-3 DE who can't play the run well you can get by with, if he's a good enough pass rusher, but the same doesn't really apply to a 3-4 DE.
I dunno, I'm sure you've probably thought of this already!
#27 by speedegg // Apr 12, 2011 - 2:09pm
Don't think SackSEER model will project to 3-4 DEs because how 3-4 DE's are used. Most 3-4 schemes have their D-linemen cover 2 gaps, occupying the O-linemen while leaving the linebackers to hit the running back or QB. Only the Cowboys, Chargers, and (maybe?) the Texans run a 3-4 defense with a 1-gap system, allowing the D-linemen to attack one gap and hit ball carrier. Might be better to say they use a 4-3 alignment with the personnel of a 3-4 defense. Even then, most sacks come from their linebackers.
#35 by Mr Shush // Apr 14, 2011 - 12:12pm
Every indication is that the Texans are going to be running the typical Phillips 1-gap 3-4 with strong 4-3 elements especially on passing downs, yes.
#36 by AlanSP // Apr 14, 2011 - 2:42pm
There are a few problems with 3-4 DEs. One is that they produce so few sacks at the NFL level as a group, in large part because they act more like DTs than edge rushers in most schemes. There are simply very few 3-4 DEs that have been successful pass rushers in recent history; offhand, I can think of Aaron Smith, whose college production was elite, and Richard Seymour (a college DT that's bounced between DE and DT for much of his career.
Seymour illustrates one of the other big problems, which is that a great deal of them were DTs in college, so it's really not an apples-to-apples comparison with college DEs.
#3 by JJohnson (not verified) // Apr 11, 2011 - 1:14pm
I was hoping to see something about Bruce Miller from UCF. Would you be able to post his projections in the comments section?
#17 by Nathan Forster // Apr 11, 2011 - 4:30pm
Bruce Miller is highly productive, but really has not demonstrated much athleticism. He doesn't get a number, because I think it is unlikely that he gets drafted in the first two rounds, but he's probably about "Aldon Smith"-level. Also, if I had to pick a third sleeper, it would be Miller--that's how few intriguing late round guys there are this year, unless there is some other guy from a mining school that I'm missing.
-----------
Yes, SackSEER has become self-aware.
#4 by Larry House (not verified) // Apr 11, 2011 - 1:15pm
It appears that SackSEER does not take in to account that a majority of Miller's and Quinn's college sacks came at the hands of weak sisters. And Brooks Reed paltry junior year was the result of a high ankle sprain. If Reed would've been able to have a better junior year then he likely would be a first RD draft pick. Moch has as good if not better Combine measures and on field performance and stats than Von Miller.
#7 by lionsbob // Apr 11, 2011 - 1:38pm
Von Miller senior year: 10.5 sacks-against Arkansas, Missouri, Texas Tech, Oklahoma, Nebraska, Texas, Baylor, and LSU (none against Stephen F. Austin, Louisiana Tech, Florida International, Oklahoma State, and Kansas).
Miller junior year: 18 sacks: against New Mexico, Utah State, and UAB (8 combined), Arkansas, Kansas State, Texas Tech, Iowa State, Colorado, and Texas (none against Oklahoma State, Oklahoma, Georgia, and Baylor).
#18 by Nathan Forster // Apr 11, 2011 - 4:41pm
If you have access to game by game information, I'd love to take a peak at it. My own experience has been that game by game sack numbers for older prospects is hard, if not impossible, to come by, and getting it would probably involve harassing people at athletic departments around the country. That said, the big highlight everyone played over and over before the draft was Vernon Gholston beating Jake Long off the edge. It's hard to think of better competition than that.
I'm not sure how you come to the conclusion that Moch is more productive than Miller. Moch has less sacks than Miller in more games, and except for splitting some time at linebacker his freshman year, he was consistently rostered as a defensive lineman, increasing his chances for sacks. Moch did have a better vertical than Miller, but his shuttle was actually a little below average, while Miller's was elite. Moch is a fine prospect who certainly has a chance to meet or exceed Miller's performance, but I can't see any reason to rank him above Miller.
-----------
Sorry JPP!
#37 by AlanSP // Apr 14, 2011 - 3:47pm
SackSEER doesn't account for the quality of competition because there's no real evidence that it matters. Ware, Umenyiora, Mathis, Allen, Strahan, and many others played against weak competion. Why it doesn't matter is an interesting question. My guess is that regardless of your conference, you'll face very few tackles that are good enough to start in the NFL, let alone talented enough to be an NFL star. It does have predictive value to look at how edge rushers do against weaker competition, because when you're talking about early draft picks, essentially everyone they face is "weaker competition."
#5 by Anon (not verified) // Apr 11, 2011 - 1:22pm
don't be too hard on yourself regarding JPP. I watched the games and a couple of his sacks were OL tripping over themselves. I won't be surprised if he has less the 4.5 sacks next year.
Is it possible to put up JJ Watt's numbers. He could be end up as a 4-3 end and his physical measurables were nuts.
#10 by rk (not verified) // Apr 11, 2011 - 2:29pm
Even if you disregard a couple of sacks, the projection is way off. SackSEER pegged him as a guy who would get less than a sack per year, in other words, a guy who isn't ever on the field enough to take advantage of a tripping OL.
#12 by Key19 // Apr 11, 2011 - 2:44pm
Agreed on the Watt point. However, be careful with your numbers: he began college as a TE and only has two years as a DE I believe. Using his total games will therefore be a poor decision, you have to only judge him by his final two years of work.
Spoiler: BloggingTheBoys did a SackSEER article earlier in the offseason, and Watt's numbers when applied correctly are RIDICULOUS. Watt is the only DE I'd be happy about the Cowboys taking this year, and if they take Aldon Smith I may have to buy some old Jerry Jones Arkansas jerseys on eBay just so I can burn them and defecate on the ashes. Yes, I'd be that upset.
#13 by thendcomes // Apr 11, 2011 - 3:31pm
I'm with you here. JPP had 4.5 sacks, but I recall only 1 sack (against McNabb) that demonstrated individual talent. In particular, on 2 sacks he ran through the line untouched.
SackSEER did not predict the Giants, who coming into the season already had the deepest DL in the league, would take JPP. Accounting for surrounding talent, along with other factors, could better explain why the predictions failed, rather than weaknesses in the algorithm. Any of the higher draft picks would have flourished when surrounded by as much talent as the Giants had last season, even after losing Kiwanuka.
#20 by Nathan Forster // Apr 11, 2011 - 4:54pm
J.J. Watt definitely has some potential as a 4-3 end. A 37" vertical and a 4.21 shuttle at 290 pounds is excellent. I've heard some Internet rumblings knocking Watt for his burst in spite of his combine numbers, but from what little I've seen of Watt, he looks plenty explosive to me. You would have liked to have seen more production, however, even in the short space of his time as a defensive player. I like to think of Watt as a poor man's Mario Williams, but that can still be a very good thing.
-----------
Sorry JPP!
#46 by JK (not verified) // Sep 25, 2011 - 9:22pm
Then you would be as bad as SackSeer. He now has 7 sacks and doesn't have Osi playing with him. People should just admit when they are wrong. JPP is a freak that destroyed SackSEER from ever being taken seriously.
#47 by Mr Shush // Sep 28, 2011 - 8:33am
SackSEER is clearly wrong about Pierre-Paul, but that doesn't mean it isn't a valuable tool. You're over-reacting to a case where even when the projection was made it was acknowledged that this player might be a difficult one for the system to peg. He's not the first player it's been wrong about, he won't be the last, but that doesn't mean it's worthless.
#49 by Anon (not verified) // Oct 17, 2011 - 4:19am
JPP now has 7.5 sacks in 6 games this year, and 12 in 22 games total. That's a whole lot of Lineman out there tripping on themselves!!
#6 by Dean // Apr 11, 2011 - 1:29pm
In what universe was George Selvie a possible first round pick a year ago?
#9 by BlueStarDude // Apr 11, 2011 - 1:52pm
It wasn't last year but Selvie was draft-eligible after his sophomore year and, at least in those early stages, was considered a potential first rounder had he declared for the draft. Obviously his stock diminished after his poor junior year and completely tanked after his equally bad senior year.
#8 by jklps // Apr 11, 2011 - 1:47pm
I think 2 of JPP's sacks last year were against the Redskins without Trent Williams in one game.
I would hope JPP would be a better DE than Stephon Heyer is a LT....
#11 by JasonK // Apr 11, 2011 - 2:38pm
That doesn't change much. Every DE gets to play against weak or injury-decimated opponents a couple times per season. Strip out the sacks that players get against backup linemen or deer-in-headlights QBs, and 10+ sack seasons would be as rare as 2000-yd rushers.
#14 by PurpleJesus28 (not verified) // Apr 11, 2011 - 4:12pm
Out of curiosity, what was Moch's projection?
As for Romeus and Bruce Carter, they did not perform the drills at the Combine.
#15 by Kwame (not verified) // Apr 11, 2011 - 4:20pm
I charted maybe 9 Giants games this year for FO, JPP actually could have had more sacks. For a rookie he consistently pressured the QB. I think what sackseer couldn't account for was the Giants defense actually featuring him. They ran plays to get him freed up to the QB. He basically played Mathais Kiwi's position in the nickel once Kiwi went down. Some of it was scheme some was JPP's really good first step and quickness.
#19 by Thomas_beardown // Apr 11, 2011 - 4:54pm
I don't have much to add, just want to say this is really cool research you've done. Thanks.
#21 by cisforcookie (not verified) // Apr 11, 2011 - 7:53pm
for a 3-4 ROLB, is there any evidence that size is really all that important? obviously you'd have some trouble at 190 lbs, but is 230 or 240 lbs too light in any meaningful way?
#22 by Anonymouss (not verified) // Apr 11, 2011 - 9:19pm
Yes I would think so, Elvis Dumerville was effective because had he been 6'2, his weight would be listed at 270. not all 270 men have the ability to move in space like a LB has too do, but I always would like to see Dwight Freeney as an 3-4 LB.
An interesting take isn't the 3-4 as much as I would point to Pittsburg using a 3-5 with polamalu. He was exposed in the superbowl but he shouldn't be obviously 1-1 with greg jennings, but otherwise the LB has to crash the line and the is a physical limit that nature sets all on her own
#23 by Aaron Brook's … (not verified) // Apr 11, 2011 - 9:30pm
A balky Achilles probably had a lot of do with Polamalu's relative disappearance in the playoffs.
#24 by Moderate Mouse (not verified) // Apr 12, 2011 - 1:41am
Sweet! Sack seer projections for this year's draft. I was wondering when this would come out.
#26 by thesteve (not verified) // Apr 12, 2011 - 1:29pm
How does SackSEER feel about Adrian Clayborne? He posted a pretty freakish 4.13 short shuttle and improved his VJ from 33.5 to 35.5 at Iowa's pro day, and never missed a game. I don't know if his sack rate would be good enough to rate highly, though.
#28 by Drunkmonkey // Apr 12, 2011 - 4:29pm
First of all, thanks for putting in Sheard. My father is a Pitt alum, and his buddies have all been asking the same question about Sheard: Why does anybody have him rated anywhere near the first round? I mean, he may look the part, but neither his combine measurables along with his production (or lack thereof) seem to add up to the first round.
Secondly, what exactly is SRAM? I mean, I don't think it's fully explained in the article, and I don't remember it from the previous SackSEER article. Maybe I'm failing in memory, but could somebody tell me what it either stands for or represents?
Last, I'm hesitant to ask, since it seems everybody wants to know what SackSEER has to say about this linemen or that, but what about penetrating DT's? Like Fairly or Liuget? I realize that this is intended to look at bigtime pass rushers, but what about the possiblity that these guys make an instant impact in the pass rush?
#31 by Drunkmonkey // Apr 12, 2011 - 8:15pm
So I decided to not be lazy for once, and looked this up during my break from studying, and found out that SRAM does stand for something: Sack Rate as Modified.
http://www.footballoutsiders.com/nfl-draft/2010/introducing-sackseer
#29 by Jim Z. (not verified) // Apr 12, 2011 - 7:40pm
DrunkMonkey: For the record, I don't have access to the SackSEER "formula" but I took a look at the combine workout numbers for last year's crop, and found that the DTs who had the most impact as rookies (Suh, Gerald McCoy, Geno Atkins, and Lamarr Houston) were the ones with the best vertical leaps and short shuttle times, interestingly enough.
Ndamakung Suh - 35.5" VJ, 4.44 SS
Gerald McCoy - 30.5" VJ, 4.48 SS
Lamarr Houston - 33.5" VJ, 4.71 SS
Geno Atkins - 33" VJ, 4.43 SS
If you consider that most DTs in any given year struggle to break 30" in the Vertical Jump, and average > 4.6s in the Short Shuttle, it makes sense that the four DTs who did the most so far are the ones with above-average numbers in those workouts.
This year's crop of DTs:
Nick Fairley - 31" VJ, 4.56 SS
Marcell Dareus - 27" VJ, 4.62 SS
Corey Liuget - 27.5" VJ, 4.68 SS
Muhammed Wilkerson - 26" VJ, 4.59 SS
Marvin Austin - 30.5" VJ, 4.40 SS
Drake Nevis - 30.5" VJ, 4.65 SS
As you can see, only three of the top DT prospects even break 30" in the vertical leap, and none of them come close to Suh's elite 35" from last year. Only one DT prospect this year was timed at less than 4.5 s in the Short Shuttle and that was Marvin Austin.
Outside of Austin and maybe Fairley, I don't think this draft class is all that great in terms of pass-rushing defensive tackles, and both of these guys also have significant character issues that could easily derail their careers regardless of on-the-field ability, unlike the complete lack of character concerns for last year's top prospects (Suh and McCoy).
#30 by Drunkmonkey // Apr 12, 2011 - 8:13pm
Thanks for the work on this. I wasn't really expecting a response, but this is pretty good, especially because I'm pretty sure my team will be drafting a penetrating DT, not necessarily with their first pick, but before the end of the 3rd round.
#32 by Jim Z. (not verified) // Apr 12, 2011 - 9:27pm
If I'm a GM of a team which needs a penetrating DT, I would seriously think long and hard about Marvin Austin in the bottom of the 1st round or top of the 2nd round. He'll slip into that region because of character concerns and the UNC suspension fiasco, but he's easily the most physically talented DT of this entire draft class. If he takes the NFL seriously, he could be the steal of the draft.
Conversely, I would stay away from Liuget and even Dareus in the first round if I'm a 4-3 team needing a pass rushing DT. These guys, based on college production and combine workouts, project as nothing more than solid all-around run-stuffing DTs. Valuable to have, but their ceiling is very limited in terms of pass-rushing production and I wouldn't spend a premium pick on one of them.
#33 by 2G1D (not verified) // Apr 12, 2011 - 9:45pm
How does Akeem Ayers look in SackSEER?I have watched some tape on him that is not impressive at all, but I've also seen some that makes him look explosive. I know he started out as a mid-first rounder, but he seems to have dropped? Suggs comp?
#34 by jim Z. (not verified) // Apr 13, 2011 - 1:03pm
I don't know about his Pro Day numbers, but his Combine numbers project him as nothing more than an average prospect: 31" Vert, 9'8" Broad, 4.28 SS, 7.49 3C... the only number that isn't below average is the SS, which is just slightly above average. And he did this all at 254lbs, so it's less impressive than 270+ lb guys posting the same numbers.
I don't think Ayers will be much of a player in the NFL, and not only because of his combine numbers. Not only is he a poor athlete, but he lacks the football IQ unathletic LBs sometimes can use to overcome their athletic limitations.
#38 by AlanSP // Apr 14, 2011 - 4:15pm
Thanks for the shout-out about Hunt.
Speaking of guys like Hunt, I'd really like to see this expand beyond the earliest rounds of the draft. I always thought that one of the big positives of this is the ability to identify guys like Mathis and Allen. One advantage of this is that the factors that matter later on are largely the same as those that matter early in the draft. Standard multiple regression might not be the best method for later rounds because there are just too many zeroes, since most later picks don't do anything. My impression is that the negative predictive value is better than the positive predictive value. That is, having a good forecast doesn't necessarily portend greatness, but having a bad one usually takes "dominant pass rusher" off the table (Dunlap notwithstanding).
I'm thinking a later-round version would be something like a formalized version of your "sleepers" section; i.e. the guys who have a real chance to be good (better than most of the late-rounders taken in the same range). Maybe the prediction could be a probability of reaching a certain level of production rather than a raw projected output. Just sort of brainstorming.
#39 by eagler7 // Apr 15, 2011 - 5:27pm
I think that would be a great idea. For the later round guys, it would have a lot more to do with short shuttle and vertical leap than on the field production because the late round gems are usually backups on their college team (unless the "on the field production" part of sackSEER involves the sacks:downs played ratio rather than the sacks:game ratio). I think that would be a fantastic idea but would take a lot longer to create.
#40 by Nathan Forster // Apr 15, 2011 - 11:31pm
That's an interesting idea. Here's a "quick and dirty" model that estimates the probability of a player drafted after the first two rounds recording twenty or more sacks in his first five years in the NFL. I wouldn't take this too seriously, but it might give a decent sense of how strong some of these "sleepers" may be in relation to one another. These are the guys that I have in my database as likely to be drafted this year:
Dontay Moch 58.6%
Marc Schiechl 54.1%
Bruce Miller 22.9%
Chris Carter 21.7%
Ricky Elmore 18.1%
Sam Acho 14.1%
Pierre Allen 13.6%
Cliff Matthews 12.1%
Ryan Winterswyk 11.9%
Jeremy Beal 11.9%
Cheta Ozougwu 11.6%
Eddie Jones 10.5%
Zane Parr 9.2%
D'Aundre Reed 9.1%
Greg Romeus 8.9%
DeMarcus Dobbs 8.4%
Kentrell Lockett7.0%
Brandon Bair 4.8%
Ugo Chinasa 4.7%
Gabe Miller 4.0%
Markus White 3.3%
Wayne Daniels 2.7%
Karl Klug 2.0%
Pernell McPhee 0.5%
-----------
Sorry JPP!
#41 by AlanSP // Apr 16, 2011 - 9:00pm
You'd be surprised how much production matters for the mid-late round guys. Guys like Jared Allen, Robert Mathis, Elvis Dumervil, and Shaun Phillips had elite production. Other late-round success stories like Trent Cole and Tully Banta-Cain had good production as well (especially if you adjust for the time Cole spent at DT).
There are actually extremely few mid-late round edge rushing gems who were mainly backups in college (none that I can think of offhand, though I don't have my database in front of me). The best bets tend to be guys who were dominant in college, like Schiechl in this year's draft. Yes, there are also guys like Elton Patterson who go on to do nothing, but late in the draft, that's not a huge risk to take.
#42 by eagler7 // Apr 17, 2011 - 2:27pm
Your right, I was just assuming that a lot of late-round rushing gems might be backups from division 1 schools, because of a few late round successful guys at other positions (like Tom Brady, for example). It is a better bet to go with guys that were dominant in College, like you said. Also, Trent Cole doesn't have elite production like the other guys you mentioned? I do have to admit that I am a little biased towards Trent Cole though.
#43 by AlanSP // Apr 18, 2011 - 10:05am
To clarify, I meant elite production in college, not the NFL. Cole's production with the Eagles would certainly qualify as elite. In college, he was very good, but didn't dominate the way that others on that list did.
#44 by Nathan Forster // Apr 19, 2011 - 7:38pm
I think that pretty much all of the true sleepers at the edge rusher position fall into one of two categories:
1) Superb production (Mathis, Dumervil, Gbaja-Biamila, Porter, Phillips)
2) Production uncertain due to position switches (Cole, Avril, Okeafor, and to a lesser extent, Kampman and Mark Anderson)
I think the second category really addresses why we do not see backups at the position in college excel at the NFL level.
I am not sure that the comparison works with quarterbacks for this reason. Even if you are an NFL-quality pocket passer quarterback in college, you're not guaranteed to get on the field. There is only "one" quarterback position in the starting lineup, and if your team has an over hyped Drew Henson or a Matt Leinart, you are likely to be out of luck. That's why, although it is an exception, you sometimes get the Matt Cassells.
The dynamic is much different on the other side of the ball. First, there are two defensive end positions, so if you happen to be on the same team as Mario Williams, coach can just stick you at the other end. And even if your team also has Manny Lawson, if you are truly an NFL quality talent, you can probably upgrade the defense significantly at defensive tackle or linebacker. Heck, some of these guys were even tight ends. Thus, historically, it has been much more likely that these players will fall into category two above then be true "college backups."
-----------
Yes, SackSEER has become self-aware.
#45 by Padre29 (not verified) // May 28, 2011 - 2:49pm
I'm a CFL fan and Hunt had an interesting journey to the Blue Bombers, he was not drafted BUT he was in the Browns camp all of that Preseason, when it became clear he was not going to make the squad he signed in WPG and had 3 sacks in 4 games, then blew up in 2010, he did however have a penchant for drawing offside flags and late hit flags.
Like Wake toured the Broncos after the season ended in 2008 and they only offered a Practice Squad deal, Hunt toured several camps and gave the Browns another shot at an offer and only the Eagles gave him a shot as he felt he was better equipped to be a 4-3 DE then a 3-4 OLB and he wanted to work with Dline Guru Jim Washburn, who also turned SackSeer's "bust" margin to "success" with Jason Babbin.
#48 by Allyn (not verified) // Oct 04, 2011 - 5:48pm
Why hasn't arm length been included in sackSEER's formala? Doesn't this value have any predictive value in terms of sack production? It seems to crop up in analysis of offensive tackles all the time. Wouldn't that make it equally important for defensive ends? What about 10-yard splits? You even mention it when you discuss Brooks Reed. Doesn't that have any predictive value in terms of burst off the line?
#50 by Dave J (not verified) // Oct 22, 2011 - 3:18pm
Jason Pierre-Paul and Jerry Hughes have sacked SackSEER. JPP for being a stud and star already despite SackSEER damning him as a bust. Hughes for being a bust despite SackSEER predicting greatness.
It is a crime that Pierre-Paul's doubters used Nathan Forster's faulty work to damage his potential NFL career. Hopefully we won't here about SacSEER now that is has been proven flimflam. Right now there is no formula to predict future NFL pass elite rushers. Nice try.
#51 by Mr Shush // Nov 03, 2011 - 12:46pm
Um, seriously? Are you his agent, or something? His mother?
Am I right in thinking you agree with both the following propositions?
1. NFL teams altered their drafting decisions based on SackSEER, resulting in Pierre-Paul being drafted later than he otherwise would have been.
2. Statistical projection systems are worthless unless they are 100% accurate.
If not, how exactly do you think "Pierre-Paul's doubters used Nathan Forster's faulty work to damage his potential NFL career"? And do you think no-one should ever try to improve their decision-making process unless they believe they can make it perfect?
#52 by Anon (not verified) // Nov 04, 2011 - 5:47am
I see this "no system is 100% accurate" argument alot. No, it doesn't have to be 100% accurate but the sheer inaccuracy on JPP was so comical as to question the entire validity of the system. It basically projected him to be the biggest DE bust in a decade. In fact, quite the opposite is true as JPP could go down as one of the all time greats. And no, this is not hyperbole has 13 sacks before his 22nd birthday can attest to.
#53 by Mr Shush // Nov 04, 2011 - 9:01am
I'm inclined to disagree: Pierre-Paul's college career was so unusual that it should have been obvious that he was a player SackSEER would be likely to have problems with, and that's before considering how unrepresentative his workout numbers were of his true athleticism. It doesn't really bother me any more than Brian Westbrook's lousy speed score. Statistical systems require an intelligent human interpreter to say, where appropriate, "hang on a minute, we probably shouldn't put too much credence on this particular result because x". One massive miss bothers me far less than more generalised inaccuracy would, especially when it's so easy to understand how it happened.
#54 by Anon (not verified) // Nov 04, 2011 - 3:33pm
I tend to agree with you level-headed analysis of the situation. However, for better or worse it seemed like SackSeer wanted to make a big splash in its debut by declaring a high profile pick to be the biggest bust in decades. Unfortunately, the exact opposite effect has happened as now SackSeer will always be associated with their incredible miss on JPP. With each sack this monster will racks up in his career, it will make SackSeer look more and more silly in the general public's eye. It doesn't help that commenters above still try to discount JPP's play (Offensive Linemen tripping over themselves - really?)and SackSeer still stubbornly projects JPP to only be an "slightly better" than an average DE only and won't approach a Demarcus Ware or Clay Matthews? He's got 13 sacks at the age of 22 in basically half a season and is still incredibly raw. 10 years from now when he is racking up his 100th career sack, people will still be commenting on how this guy projected him for 3 sacks in his whole career, and will be all SackSeer will be remembered for.
#55 by Anon (not verified) // Nov 04, 2011 - 4:00pm
I also don't understand this statement:
"Pierre-Paul's college career was so unusual that it should have been obvious that he was a player SackSEER would be likely to have problems with, and that's before considering how unrepresentative his workout numbers were of his true athleticism"
Go and re-read the 2010 report. According to SackSeer JPP's college career wasn't unusual at all as they were sure that anyone coming out of a junior college was an almost guaranteed bust. And how can you blow off the workout numbers as being unrepresentative of his true athleticism...isn't the whole system predicated on using those workout numbers to predict a pass-rushers success? It seems like SackSeer was exactly designed to pick up "busts" like JPP who went to junior colleges and have poor workout numbers - a human interpreter said wait a minute, this guy looks like an athletic freakand is projected to be good but he went to a JUCO (so he must be too dumb too pick up the playbook!) and his athleticism is "gimmicky" and won't translate into the NFL. This interpretation was pretty much wrong on every level and reveals major flaws in the system.
#56 by Mr Shush // Nov 05, 2011 - 12:06pm
Wow. I hadn't read that write-up in quite a while, and it's a hell of a hostage to fortune. I'm always pretty dubious of comparing players who NFL scouts rate very differently on a purely statistical basis, and I think that it would have been prudent to note that not only was that group of Juco transfers a decidedly small sample but that most of them were drafted significantly later than Pierre-Paul would be. In fact, the only man on the list drafted as highly, Jerome McDougle, was a bust entirely through injury, and should probably be disregarded entirely in developing a model (and certainly in looking for comparables for one particular player).
I do think FO sometimes have a tendency to argue the case for unusual finding in their stats reflecting reality, rather than trying to analyse whether they do and considering why in this particular case they might not. See "Randy Moss: just a cog".