Writers of Pro Football Prospectus 2008

Most Recent FO Features

ReplayIns14.jpg

» Scramble for the Ball: Getting it Right?

Instant replay review is one of the cornerstones of the modern NFL. The process and its myriad special rules have been internalized and constantly debated. Mike Kurtz wonders: is it worth it?

15 Apr 2013

SackSEER 2013

Guest Column by Nathan Forster

In Football Outsiders Almanac 2012, we unveiled SackSEER 2.0, an improved version of our system for projecting the pass rushing success of college edge rushers (4-3 defensive ends and 3-4 outside linebackers) selected in the NFL Draft. SackSEER 2.0 contained the following elements:

  • An "explosion index" that measures the prospect’s scores in the forty-yard dash, the vertical jump, and the broad jump in pre-draft workouts;
  • A metric called "SRAM" which stands for "sack rate as modified." SRAM measures the prospect’s per game sack productivity, but with adjustments for factors such as early entry in the NFL Draft and position switches during college;
  • The prospect’s college passes defensed divided by college games played; and
  • A metric that measures the prospect’s missed games of NCAA eligibility. This metric includes games missed for reasons such as time spent in Junior College, academic problems, injuries, benchings, and suspensions.

SackSEER 2.0 expresses its thoughts on each drafted edge rusher through two outputs: SackSEER projection and SackSEER rating. SackSEER projection projects the number of regular season sacks that a prospect will record in his first five seasons in the NFL. Unlike SackSEER rating, SackSEER projection incorporates as an additional element the prospect’s projected round drafted from NFL Draft Scout. SackSEER rating provides a historical percentile rating on the college edge rusher’s prospects for success as compared to the other prospects in SackSEER’s database, irrespective of projected draft position. For instance, SackSEER currently has 299 edge rushers in its database, so a prospect in this year’s draft who is stronger than 200 of those prospects on the historical trends identified by SackSEER would have a SackSEER rating of 66.9 percent [200/299]. So, if you want to see how the prospects stack up based on SackSEER’s trends alone, you can look at SackSEER rating. And if you want to see how the prospects stack up based on SackSEER’s trends when balanced against conventional wisdom, you can look at SackSEER projection.

Last year, SackSEER 2.0 was somewhat bullish on USC defensive end turned outside linebacker Nick Perry, and was highly critical of two edge rushers who were thought to be sure first-round picks: Melvin Ingram and Courtney Upshaw. (Upshaw was actually drafted in the second round.) It is hard to get a read on Perry, who recorded 2.0 sacks in six games before landing on injured reserve. Ingram and Upshaw, however, struggled at compiling sacks in their first years. Ingram recorded only a single sack. Upshaw recorded only a sack and a half during the regular season and was held sackless during the Ravens’ playoff run.

The one player that SackSEER 2.0 seems to have really missed on is Bruce Irvin. Irvin had a fine first year for the Seahawks, recording 8.0 sacks to lead all rookies. Since we are always looking to improve SackSEER, we have two more tweaks to make to the model, both of which would have improved SackSEER’s take on Irvin significantly.

As we mentioned in Football Outsiders Almanac 2012, there is a correlation between pass-rushing success and the agility drills, but those drills were not included in the model because although they approached statistical significance, they were not statistically significant. Well, adding the prospects from the 2011 NFL Draft to the database provided a big push in favor of the three-cone drill. Sam Acho and Von Miller, both drafted in 2011, have the two quickest edge rusher three-cone times in the database by a large margin (Acho and Miller recorded times of 6.69 and 6.70 respectively, which each beat the previous record by more than a full tenth of a second). Both Acho and Miller turned out to be successful players, so needless to say they exert considerable influence on the refreshed version of the model. Irvin, who is not yet included in the model (only edge rushers two NFL seasons removed from their draft are included), matched Miller’s amazing three-cone time of 6.70 seconds. Accordingly, SackSEER now includes a separate metric for the three-cone drill.

Moreover, we have sought to replace the missed games metric with something that does not penalize players with backgrounds in junior college. When SackSEER was first rolled out in 2010, the list of highly-touted edge rushers with junior college experience was truly horrendous. Since then, however, the only two junior college edge rushers of any repute, Jason Pierre-Paul and Irvin, have been quite good.

We have replaced the missed games element of SackSEER with two new metrics. First, SackSEER includes a factor for whether or not the player received or was eligible for a medical redshirt during his college career. (A player is eligible for a medical redshirt if he appears in fewer than 30 percent of his team’s competitions and then suffers a season-ending injury.) This is to help capture players with serious injury concerns. Second, SackSEER now includes an adjustment for Division II and Division III prospects. There have actually been a large number of successful edge rushers from smaller schools, however, they have come exclusively from non-AQ Division I and FCS schools. Prospects from Division II and Division III schools have historically underperformed. The key driver for this metric is Andy Studebaker, who recorded dominant statistics during college -— but for tiny Wheaton College. Moreover, a disproportionate number of drafted edge rushers from junior college were also Division II/III players, so that could have been part of what the original junior college adjustment was picking up on.

Note that the Division II/III adjustment is part of the SackSEER rating, but is not included in SackSEER projection because NFL Draft Scout does a good job of baking level of competition into its rankings. Also, the medical redshirt and the Division II/III adjustment are the "weakest" of SackSEER’s metrics, so they probably deserve a little more skepticism than the others.

Turning our eyes to this year's draft, SackSEER thinks that this is a fairly strong class. There is plenty of top-end talent and the class is deep, with some intriguing sleeper prospects. If this class has a weakness, however, it is that there is no prospect who marries good athleticism with strong college sack production.

One other quirk of adding the 2011 edge rushers to the database is that they were so good that they raised the expectations for the average drafted edge rusher. Accordingly, SackSEER’s 2013 projections are higher than last year’s projections, both because the level of prospect is a little better and because the influence of the 2011 class causes SackSEER to produce higher projections.

Here are the SackSEER projections for the top edge rushers available in the 2013 NFL Draft:

Barkevious Mingo, LSU

Proj. Round 1
Explosion Index +1.64
SRAM 0.43
PD Rate 0.28
Three-Cone Drill 6.84
SackSEER projection 34.5
SackSEER Rating 94.6%

Last week, Matt Waldman covered both Mingo and Dion Jordan in his Futures column. Based on his scouting, he noted that Jordan was close to a sure thing with all-around talent at the outside linebacker position, while Mingo was a bit more of a high-risk, high-reward draft pick because his on-field technique doesn't quite match his athleticism and explosiveness.

Ironically, SackSEER has the exact opposite opinion. When it comes strictly to sack production, SackSEER sees Jordan as a high-risk, high-reward draft pick whose on-field production in college never quite matched his measurables, while Mingo is SackSEER's number-one edge rusher prospect for 2013. Mingo outrates Jordan in every single one of the elements in SackSEER, both the Combine drills and the college production measures.

Mingo is somewhat analogous to Jevon Kearse. Like Kearse coming out of Florida, Mingo has outstanding athleticism and a lot of passes defensed. Mingo’s workouts were consistently excellent -— his forty-yard dash, vertical jump, broad jump, and three cone were all well above average for the position. As was also the case with Kearse, Mingo has only lukewarm sack production. Some contend that Mingo’s production dropped this year because he was asked to rush the passer much less than in seasons past. Edge rushers with less production than Mingo have gone on to be stars at the position, so a team selecting Mingo -— even with a fairly high draft pick -— is making a solid gamble.

Ezekiel Ansah, BYU

Proj. Round 1
Explosion Index +0.77
SRAM 0.22
PD Rate 0.34
Three-Cone Drill 7.11
SackSEER Projection 30.3
SackSEER Rating 82.3%

Why is SackSEER so high on a prospect with only 4.5 career sacks? Because as a senior, Ansah was a dominant pass defender. Ansah had an interception and nine passes defensed —- which is more passes defensed than all but one of BYU’s defensive backs logged in 2012. Ansah has been lauded for his excellent size and athleticism. Interestingly, Ansah performed better on his forty-yard dash than on his vertical jump and broad jump. Typically, players who are large for the edge rusher position struggle more with the forty, and have an easier time with the jumps (see, for example, Mario Williams’s average-ish 4.70 forty-yard dash and his amazing 40.5-inch vertical leap).

Bjorn Werner, FSU

Proj. Round 1
Explosion Index -0.66
SRAM 0.58
PD Rate 0.37
Three-Cone Drill 7.30
SackSEER projection 28.5
SackSEER Rating 77.3%

Werner is yet another reference point in the debate over on-field production versus workouts. Werner did everything you could ask him to on the field: recording a strong 0.58 SRAM and defensing 0.37 passes per game. On the other hand, Werner’s below-average Combine gives some reason for pause. Werner’s profile bears some similarities to Darren Howard, who had lots of sacks and passes defensed in college, but performed poorly (indeed, much more poorly than Werner) in pre-draft workouts.

For more on Werner, check out this Futures column by Matt Waldman.

Jamie Collins, Southern Miss

Proj. Round 2-3
Explosion Index +2.5
SRAM 0.48
PD Rate 0.35
Three-Cone Drill 7.10
SackSEER projection 27.2
SackSEER Rating 97.7%

SackSEER likes Jamie Collins a lot. Collins has a SackSEER Rating of 97.7 percent, which means that, when not controlling for draft position, Collins is SackSEER’s eighth-favorite prospect ever. Even after controlling for projected draft position, SackSEER favors Collins to collect more sacks than probable first-round picks Jarvis Jones and Dion Jordan, as well as all of the edge rushers who were available in the 2012 NFL Draft.

Collins had a fantastic Combine workout -— a 4.64 forty, a 41.5-inch vertical leap, and an edge rusher-record 139-inch broad jump. Collins, however, is not just a workout warrior. Collins’ projection is also buoyed by his passes defensed -— Collins had 18 passes defensed in 52 games, including three interceptions.

The best case against Collins is his mediocre SRAM against a relatively low level of competition in the Conference USA. Collins is eerily similar in this respect to a recent prospect who isn’t working out so well: Dontay Moch. Moch and Collins share the same undersized 240-pound frame, the same unreal Combine numbers, and the same average-to-mediocre sack production against low-end Division I competition.

There are, however, reasons to believe that the arc of Collins’ career could bend differently than Moch’s. For one, Collins has better excuses for failing to record big sack numbers. Unlike Moch, Collins was bounced from position to position at Southern Mississippi. As a freshman, he was a defensive back. As a sophomore, he was a linebacker. Moreover, as a senior, the Southern Mississippi Golden Eagles were truly terrible. The team went 0-12, and as is often the case when a team is perpetually behind, Southern Mississippi’s opponents barely passed (338 total attempts). Only six Division I teams faced fewer passes than the Golden Eagles. It is hard to sack the quarterback when your opponent spends most of the game running the ball to protect a lead, but Collins still produced his best season, recording 10.0 sacks and five passes defensed in 12 games.

Also, Moch’s career has had a certain snakebitten quality. Moch broke his foot in his first preseason game, barely played during the rest of his rookie year, lost the first four games of his second year to a suspension for banned substances, and ended his second year on injured reserve for recurring migraines. It would be unreasonable to assume that Collins would suffer through a similar odd series of circumstances just because he and Moch played in non-AQ conferences, are small for their position, and have mind-bending athleticism.

Of course, Collins is far from a guarantee, but he presents a lot of upside (especially for a 3-4 defense) at an important position for a relatively low price.

Dion Jordan, Oregon

Proj. Round 1
Explosion Index +0.88
SRAM 0.31
PD Rate 0.05
Three-Cone Drill 7.02
SackSEER projection 24.6
SackSEER Rating 52.3%

A decision to turn the card in for Dion Jordan is definitely a decision to bet on measurables. Jordan is a long, fast athlete at 6-foot-7 with 33 7/8-inch arms and a 4.60 forty time. The problem with Jordan is that he has tons of game experience but little production to show for it. In 39 games at Oregon, Jordan only recorded 14.5 sacks and two passes defensed. It is curious that Mingo absorbs so much criticism for his production against LSU, while Jordan receives little, when Mingo has outperformed Jordan on all fronts and against stronger competition. Jordan’s lack of passes defensed is especially concerning, given that scouting reports on Jordan note (positively) that he was regularly asked to drop into coverage, and thus likely had plenty of opportunities to make plays on the football. Jordan’s best-case scenario is probably to become Kamerion Wimbley -- another player with great athleticism and little sack production in college.

Jarvis Jones, Georgia

Proj. Round 1
Explosion Index -0.91
SRAM 0.84
PD Rate 0.18
Three-Cone Drill 7.46
SackSEER Projection 23.7
SackSEER Rating 58.7%

This is what happens when you bomb your pro day. Jones’s 4.88 forty, 30.5-inch vertical jump, and 9-foot-4 broad jump are poor numbers -- especially for a smaller edge rusher like Jones, who tips the scales at only 245 pounds. Here is a complete list of edge rushers drafted since 1998 that had a 4.80 forty or worse at under 250 pounds: Casey Dailey, Bryant McNeal, Cheta Ozougwu, and Kroy Biermann. Biermann has been a solid contributor, but he is the only one on that list with a career sack. Jones’s production is good, but not quite good enough to outweigh bad pre-draft workouts. Compare Jones to a player like Terrell Suggs, who had a disappointing 4.84 forty at the Combine but "are you sure that’s not a typo?"-level production in college. Suggs had 44 sacks in 36 games, compared to Jones’ 28.5 in 34 games.

Damontre Moore, Texas A&M

Proj. Round 1-2
Explosion Index +0.25
SRAM 0.72
PD Rate 0.13
Three-Cone Drill 7.08
SackSEER projection 23.5
SackSEER Rating 80.1%

Damontre Moore had 26.5 sacks and five passes defensed in 38 games with the Aggies. Moore is taking some flak for his 4.95 forty-yard dash, but the talking heads are ignoring Moore’s good jumps -— a 35.5-inch vertical and 10-foot-2 broad -— which are just as important. Moore also recorded a good three-cone time at his pro day.

For more on Moore, check out this Futures column by Matt Waldman.

Margus Hunt, SMU

Proj. Round 1-2
Explosion Index +1.03
SRAM 0.30
PD Rate 0.13
Three-Cone Drill 7.07
SackSEER Projection 22.4
SackSEER Rating 64.5%

Margus Hunt has had an interesting path to the NFL Draft. Hunt was originally a discus thrower from Estonia who matriculated with the football team after SMU disbanded its men’s track program. Much has been made of Hunt’s rawness, but he actually has a pretty decent amount of experience: 53 college football games worth of it. One would have expected Hunt to blossom by now if he was going to do so at all, but through those 53 games and the Senior Bowl, Hunt has shown little ability to make an impact on the field outside of blocking kicks. Hunt’s major plus is his great workout numbers, which is why his SackSEER Rating is a bit higher than average. Historically, Combine performance has correlated more closely to NFL success than sack production, so SackSEER thinks Hunt is worth a flyer at some point in the draft, although probably not the low first/high second round.

2013's Best Sleepers

David Bass Jr., Missouri Western State (Div II)

Proj. Round 5-6
Explosion Index -0.50
SRAM 0.83
PD Rate 0.55
Three-Cone Drill 7.07
SackSEER Projection 20.1
SackSEER Rating 87.3%

David Bass Jr. enjoyed a tremendous career at Missouri Western State, recording 40.5 sacks, four interceptions, and a mind-boggling 23 passes defensed.

There’s no doubt that averaging over half a pass defensed per game is a tremendous accomplishment for a defensive lineman at any level. Here is a complete list of all edge rushers drafted since 1998 who averaged at least 0.5 passes defensed per game in college: Jared Allen, Connor Barwin, Jevon Kearse, Paul Kruger, Julius Peppers ... and ... well, Andy Studebaker. Overall, that’s a strong group that averaged 34.2 sacks through their first five years in the NFL.

We were ready to announce Bass as the greatest SackSEER sleeper prospect ever, but a sub-standard Combine dampened his prospects considerably. Still, there have been edge rushers that have overcome worse Combine performances than Bass's, and few can top the destruction that Bass has wrought upon those who sought to pass the football against the Missouri Western State Griffons.

There is some evidence that NFL decision-makers tend to fetishize Combine performance for prospects at low levels of competition, and sometimes unfairly so. (For a great example from another position, compare Brent Grimes’ workouts with his college production.) That could be a mistake with Bass, whose production suggests unique upside for a player expected to last deep into day three of the NFL Draft.

Devin Taylor, South Carolina

Proj. Round 6
Explosion Index +1.11
SRAM 0.35
PD Rate 0.35
Three-Cone Drill 6.89
SackSEER projection 16.0
SackSEER Rating 93.3%

Devin Taylor had the distinction of playing next to Jadeveon Clowney, a defensive end who was such a beast as a sophomore that some sportswriters suggested that he should sit out a year to avoid injury. Taylor is an odd prospect who SackSEER likes because he has strong workout numbers and nearly as many PD’s as he had sacks.

(Don't ask us what SackSEER would say about Clowney, who is the favorite to be chosen first overall in the 2014 draft. Obviously, we don't have any combine drill stats for him yet. He has 21 sacks and three passes defensed through two seasons.)

Here is a chart of the rest of the SackSEER projections for edge rushers who received an invite to this year’s NFL Combine:

Name College Proj. Round Explosion
Index
SRAM PD Rate Three-Cone
Drill
SackSEER
Projection
SackSEER
Rating
Cornellius Carradine Florida State 1-2 +0.06 0.49 0.00 7.29 16.4 28.8%
Sam Montgomery LSU 2 -0.11 0.58 0.06 7.18 16.3 48.3%
Trevardo Williams Connecticut 3 +1.38 0.63 0.06 7.24 15.5 79.4%
John Simon Ohio State 3–4 +0.72 0.41 0.20 7.10 15.5 76.3%
Michael Buchanan Illinois 4 -0.08 0.27 0.23 6.91 13.5 61.0%
Alex Okafor Texas 2 -0.90 0.48 0.07 7.26 13.3 20.2%
Corey Lemonier Auburn 3 +0.70 0.49 0.03 7.14 12.8 53.7%
Chase Thomas Stanford 3–4 -0.64 0.53 0.17 7.17 11.8 49.3%
LaVar Edwards LSU 3–4 +0.14 0.20 0.13 7.03 11.6 37.8%
Ty Powell Harding 5 +1.09 0.52 0.07 6.98 11.2 29.1%
Name College Proj. Round Explosion
Index
SRAM PD Rate Three-Cone
Drill
SackSEER
Projection
SackSEER
Rating
William Gholston Michigan State 5–6 -1.29 0.34 0.36 7.20 8.9 48.3%
Walter Stewart Cincinnati UDFA +0.93 0.48 0.33 7.28 8.0 83.9%
Cornelius Washington Georgia 4–5 +1.81 0.21 0.06 7.47 7.7 48.2%
Mallicah Goodman Clemson 4 -0.36 0.23 0.07 7.10 7.1 17.4%
Armonty Bryant East Central UDFA -0.14 0.77 0.23 7.22 6.8 33.8%
Joe Kruger Utah 6–7 -0.07 0.36 0.11 7.17 4.3 36.5%
Qaunterus Smith Western Kentucky 7–UDFA +0.11e 0.53 0.07 7.28e 2.5 43.8%
Tourek Williams Florida International 7 -0.14 0.36 0.14 7.44 2.0 29.1%
Brandon Jenkins Florida State 3–4 -1.62 0.56 0.08 7.40 0.0 1.7%
Nathan Williams Ohio State 7–UDFA -0.27 0.26 0.14 6.99 0.0 7.2%
Red = received medical redshirt; e = estimated based on performance in other drills

Posted by: Guest on 15 Apr 2013

42 comments, Last at 02 May 2013, 10:49pm by Intropy

Comments

1
by Aaron Brooks Go... :: Mon, 04/15/2013 - 12:41pm

"Mingo has outperformed Jordan on all fronts and against stronger competition"

He was playing against SEC offenses, not SEC defenses.

4
by ccombs2 (not verified) :: Mon, 04/15/2013 - 1:14pm

Agreed, big difference. I think Jordan's career will hedge on how he's used. He is not Aldon Smith, more like Julian Peterson to me.

6
by Aaron Schatz :: Mon, 04/15/2013 - 1:39pm

From what I know, the SEC is better on offense than the Pac-12, if you go up and down the entire conference. The Pac-12 is stronger on offense if you only count the three or four best teams (the best of which Jordan did not face, except in practice).

13
by Mehllageman56 (not verified) :: Mon, 04/15/2013 - 2:38pm

At first I doubted you, but when I mapped out the respective FEI ratings, it makes an even stronger case for the SEC. Texas A and M and Alabama are both ranked above Oregon, and then things even out until you compare Florida and LSU vs ASU and Washington. The only issue with this is that you're comparing 14 teams with 12, but overall they're pretty even.

2
by slipknottin :: Mon, 04/15/2013 - 12:54pm

The big issue with measuring sack production with a guy like Jordan is that he simply did not rush the passer every down. He dropped into coverage, lined up wide against slot receivers. It frustrated me, in fact, as I'm sure it did many people to see how little Jordan actually rushed the passer

Hard to compare production when you aren't comparing snaps doing the same thing

3
by IrishBarrister :: Mon, 04/15/2013 - 1:10pm

I can see your point to a certain extent, but I would be very wary of a prospect who's alleged reason for lack of sacks was dropping into coverage ... and had a PD Rate of .05(!). This results in the strange argument: "My lack of sack production came from my frequency in dropping into coverage, where I had even worse pass defense production, and that's why I'm a great prospect." Not terribly compelling stuff there.

8
by Aaron Schatz :: Mon, 04/15/2013 - 1:42pm

SackSEER is not measuring which player will be better overall. SackSEER is measuring which player will be more productive as a pass rusher. If Jordan is more flexible and better in coverage, that will not be shown in SackSEER.

14
by IrishBarrister :: Mon, 04/15/2013 - 2:57pm

Perhaps I should rephrase.

Assume arguendo that sack and PD Rate are important elements to projecting pass rushers, as SackSEER does. We also assume that it is perfectly reasonable to say "his sack production is lowered because he dropped into coverage a lot", which of course Jordan did. At the same time, one pass defended every 20 games indicates something very concerning.

The point I was trying to make was that the "he drops into coverage a lot, therefore its okay that his PD Rate is low" is a very strange argument to make. One would think that his PD Rate would be higher and would somewhat offset his deflated sack production (that, or he's terrible in coverage). So either SackSEER is indicating something an analyst should be concerned about, or SackSEER's metrics really don't apply to OLBs who drop into coverage a lot.

29
by Mr Shush :: Tue, 04/16/2013 - 4:57am

I don't know that it is a strange argument to make. Players who drop into coverage can have a positive impact without defensing a lot of passes, and pass-rushers who defense a lot of passes may well be doing so on pass-rushes not when dropping into coverage. Indeed, batting passes when rushing would seem to me more likely to be indicative of good pass-rushing talent than doing so in coverage.

To take a really extreme example, JJ Watt defenses a lot of passes. He does not spend a lot of time in coverage.

16
by SackSEER :: Mon, 04/15/2013 - 3:25pm

The weird dynamic to Jordan's usage is that it's not like Oregon was coached by Ron Zook: they were coached by Chip Kelly, who is widely considered to be a "genius" (albeit on the opposite side of the ball). You could wonder if there might be a reason that Jordan was used the way that he was (of course, that all goes out the window if the Eagles take Dion Jordan at number four).

Also, for what it's worth, SackSEER doesn't hate Jordan, it just really likes Mingo and thinks that Jordan is just kind of average.

20
by Mehllageman56 (not verified) :: Mon, 04/15/2013 - 5:31pm

Dion Jordan's best skill is disrupting the short passing game by knocking recievers off their patterns at the line scrimmage and then quickly dropping into coverage. He looks like Gronkowski's anti-matter to me, especially since they both have injury issues. What Alotti (Oregon's defensive coordinator) likes to do is mix everything up, so the opposing team doesn't know who to block. So he would use Jordan all over the place, rushing him some and having him cover a lot.
The main thing to me is that Mingo never dropped into coverage, so I'm not sure how good he would be at it. A bunch of other 3-4 linebacker prospects have that issue, but none of them are going in the first round.

5
by slipknottin :: Mon, 04/15/2013 - 1:16pm

I disagree. I believe it's easier to get high pass deflection numbers up on the line, and often at DT than it is at linebacker.

Briggs and fletcher led the league in deflections from the LB spot with 11. Watt had 16...

I just don't think it's comparable.

7
by Aaron Schatz :: Mon, 04/15/2013 - 1:41pm

Watt is a very special, unique case. You don't get a defensive lineman with 16 PD's every year. Or most years, in fact.

For example, in 2010, 11 linebackers had 8+ PD's. Only two defensive linemen did: Kevin Williams with 10 and Julius Peppers with 9.

22
by slipknottin :: Mon, 04/15/2013 - 7:35pm

Right, but its still a pretty unfair argument to say that because Jordan was used in coverage so often, his pass deflection numbers should be higher. I don't think it works that way.

28
by Aaron Brooks Go... :: Tue, 04/16/2013 - 1:14am

Unless he was really good. How many PDs did Bailey or Revis get in a typical season?

10
by Rich A (not verified) :: Mon, 04/15/2013 - 1:50pm

Who was next after Watt though?

I think using Watt as evidence for something universally true about the DT position is like saying because Victoria Secret models are pretty then all girls are pretty.

12
by slipknottin :: Mon, 04/15/2013 - 2:00pm

Corey Liuget with 9. Abraham, Kevin Williams, Brian Robinson with 7.

For the most part none of those guys were in coverage either. That's passes batted at the line.

I just think you can't look at pass deflection numbers as a way to determine how good someone is in coverage.

Maybe JJ Watt is terrible in coverage, can't determine that off his numbers.

9
by Mehllageman56 (not verified) :: Mon, 04/15/2013 - 1:50pm

I think the problem with SackSeer is that it is trying to do too much. It measures the sack potential from usually two, but often 3 or 4 different positions: 4-3 defensive end, 3-4 linebacker, and then the position that really throws it off- 4-3 defensive tackle or 3-4 defensive end. All of these positions emphasize different skills, and a player lacking in one skill may not get on the field in a 3-4 but be perfect in a 4-3. A 3-4 linebacker better have good shuttle times, i.e., be able to cover. A 4-3 defensive end, not so much, a defensive tackle, not at all.
Then guys like Jason Pierre Paul throw a wrench into it by lining up as a defensive tackle on passing downs, at a position where strength and length might mean more than the ability to change direction swiftly.
That might explain why the Lewin forecast has been able to predict success a lot better; it only has one position and skill set to take into account.

All this aside, thanks again for SackSeer, I love this stuff.

15
by SackSEER :: Mon, 04/15/2013 - 3:04pm

Glad you like the projections Mehllageman; I like doing them.

I agree with you that the line drawing is very difficult, and it has been something that I have struggled with. The issue with the position is that these guys move all over the place. There's no way to look at these guys like you look at quarterbacks.

I'm hoping, though, that SackSEER might ultimately shake out a little bit like the LCF: a few rough bumps in the beginning, stabilizing through an update, and then hitting on a guy like LCF hit on Russell Wilson.

18
by cisforcookie (not verified) :: Mon, 04/15/2013 - 3:47pm

is there enough data to support rating players based on how they fill different roles?

It wouldn't amaze me if certain athletic attributes were more important for 4-3 DE than 3-4 OLB, etc.

and would it be possible to add a subjective grade to the calculus based on technique? sortof like how sackseer automatically excludes people projected to go lower than the top rounds. because otherwise you have a lot of noise from players who get by on skills not easily measured by sackseer except by their college sack numbers.

21
by Mehllageman56 (not verified) :: Mon, 04/15/2013 - 5:43pm

The funny thing is that SackSeer 2012 recomended Quinton Coples be moved inside, because his explosion index and shuttle times were poor for a 4-3 DE. Then the Jets put him at the 5 tech in their defense. He didn't light it up, but 5.5 sacks isn't bad for a guy projected for 18 or so in 5 years. So you guys are already on the track to figuring this stuff out.
I don't know if there's a much of a difference between 4-3 DE and a 3-4 LB, other than the linebacker will probably need to play pass coverage. I think the DT guys who sneak in (Jason Pierre Paul the big wrench here) are the ones breaking the system down.

11
by Mehllageman56 (not verified) :: Mon, 04/15/2013 - 1:57pm

I would argue Moch's problems in the NFL are almost all injury related, and have almost nothing to do with Collins. The only implication is that 240 speed guys get injured in the NFL, and Collins would have company with the UConn guys in that pile. Moch also had better production in college than Collins while playing for a better team; Nevada knocked off Boise State that year, while Southern Miss was winless. Moch had close to thirty sacks in his Nevada career.

17
by dgreene (not verified) :: Mon, 04/15/2013 - 3:29pm

Without combine/workout numbers for Tank Carradine, how did you incorporate him into the model?

Also, where does Datone Jones fall? He's another player I'm curious about as some feel that he's a 3-4 end, while others feel he's a 4-3 DE, while still others feel he could play DT or OLB.

35
by SackSEER :: Thu, 04/18/2013 - 8:23am

For Carradine, an estimated forty based on pre-combine estimates is used, in turn, to estimate the remaining drills. As reflected by the close to zero explosion index, this results in Combine factors that are very close to the mean, which results in the Combine factors having little pull in either direction on his projections.

19
by Anonumos (not verified) :: Mon, 04/15/2013 - 4:08pm

Jordan's coverage responsibilities were almost always to bump a slot receiver, backpedal five to eight yards and then tackle any player who caught the ball at or behind the line of scrimmage. He would have had almost no opportunities to defend passes like a cornerback when this was his assignment.

23
by Anonymous49 (not verified) :: Mon, 04/15/2013 - 8:53pm

Does Arms length have any predictive value?

24
by SackSEER :: Mon, 04/15/2013 - 9:08pm

Nope. I only have arm lengths for about 70% of players, but the relationship between arm length and success is negligible.

Here is everybody with 35"+ arms. It's basically Aldon Smith and some guys.

Aaron Maybin 35.75
Antwan Odom 35.75
Matt Shaughnessy 35.625
Jeremy Thompson 35.625
Carlos Hall 35.5
Aldon Smith 35.375
Lawrence Sidbury 35.375
Cameron Jordan 35
Greg Romeus 35
Mathias Kiwanuka 35
Manny Lawson 35
Khari Long 35
Alonzo Jackson 35

-----------
Sorry JPP!

25
by Anonymous49 (not verified) :: Mon, 04/15/2013 - 9:13pm

Is there a cutoff value (my guess would be around 33") below which you are not getting much production?

The way I can found arm length values is to google "Player Name Combine Results".

26
by SackSEER :: Mon, 04/15/2013 - 9:35pm

Well, Jared Allen, Elvis Dumervil, Osi Umenyiora, and Terrell Suggs all had 32" arms. There were only eleven guys with shorter arms than that. Admittedly, none of those guys did much in the NFL, but you're getting into small sample size theater at that point. I don't think Jared Allen would go from dynamo to chump if you somehow managed to shrink his arms half an inch.

Arm length is like the 10 yard split, which one would think would correlate to NFL sacks, but it turns out in this instance that the history just isn't there.

-----------
Sorry JPP!

27
by Anonymous49 (not verified) :: Mon, 04/15/2013 - 10:21pm

Surprising, indeed. The low number of shorter-armed (<32") pass rushers, I think, is from a selection bias. If a guy has short arms, he probably won't play DE/OLB in college.
Not that it would be relevant to this year's draft. The only two pass rushers (1st to 6th round) with <33" arms are Chase Thomas and DeVonte Holloman.

30
by DE_Broncofan (not verified) :: Tue, 04/16/2013 - 1:27pm

For arm length you forgot about Datone Jones (32.75"), Tourek Williams (32.5") and Craig Roh (32.375"), although you might be looking at projections that have both Williams and Roh going in the 7th or UDFA.

32
by Anonymous49 (not verified) :: Wed, 04/17/2013 - 1:01am

I have Jones as a 34DE, not as an outside pass rusher (nor is he listed in the article). I had Williams with 33.5", so I caught a typo in my database, thanks.

31
by Dr. Mooch :: Tue, 04/16/2013 - 11:31pm

Hey, remind me, what was Maybin's SackSEER projection?

34
by Aaron Schatz :: Wed, 04/17/2013 - 5:34pm

Maybin's projection was 21.2/71.8%. It overvalued Maybin compared to reality, but was smarter about Maybin than the Bills were. (That projection translates into a guy who goes in the second round, not a guy who should go 11th overall.)

36
by SackSEER :: Thu, 04/18/2013 - 8:30am

I have watched quite a bit of Maybin and it is pretty evident to me why he has underperformed his projection. May in is indeed fast--he almost always gets a great jump off the line. As a result, he can completely abuse lower echelon tackles--he invariably racks up lots of sacks every preseason. The problem is that pretty much even the lower tier starting tackles in the league are quick footed enough to stop an outside pass rush move if they don't have to worry about anything else. At that point, Maybin is done because he is stiff hipped (very bad 3 cone) and he has no functional strength (which SackSEER has no way to measure because the bench press is a useless metric for these players).

33
by Animynous (not verified) :: Wed, 04/17/2013 - 12:30pm

Does it seem to anyone else like there is an awful lot of curve-fitting going on in here?

37
by justanothersteve :: Thu, 04/18/2013 - 3:41pm

Some measurement tests explained (from various sources)

Arm length: A measurement is made from the point of the shoulder to the tip of the little finger.

20-yard shuttle: Also called the 5-10-5 Shuttle, this is a test of agility including speed, explosion and changing of directions. Technique is also important. Each player will be timed how fast they can go 5 yards laterally, then 10 back in the opposite direction, and finishing 5 yards back to the start line.

Three-cone drill: An agility test where the players runs around three cones placed in the shape of an "L". There are 5 yards between each cone.

60-yard shuttle: From a starting line, a player runs 5 yards and back, then 10 yards and back, then 15 yards and back, touching the line each time.

38
by Westley (not verified) :: Sun, 04/28/2013 - 5:54pm

What was Sio Moore's sack SEER?????

39
by Freemanator (not verified) :: Sun, 04/28/2013 - 8:01pm

Can I ask who were the seven guys that Sackseer rates higher than Jamie Collins?
Thanks

41
by SackSEER :: Thu, 05/02/2013 - 9:44pm

In order:

Jared Allen
Connor Barwin
Robert Mathis
Julius Peppers
Von Miller
Jevon Kearse
Julian Peterson

-----------
Sorry JPP!

42
by Intropy :: Thu, 05/02/2013 - 10:49pm

Were these guys in the training set?

40
by cram9030 :: Wed, 05/01/2013 - 10:51pm

I was just curious why you chose sacks instead of a less volatile statistic like pressures? It would seem to me that the use of a statistic that has slightly less variation would help the regressions. Or is the thought here that the 5 year sack totals are not that volatile because of the time scale being used?

I was also wondering if you have any opinion on a combined "Pass Stopper" statistic that would combine sacks, pressures, and pass defended based off the change in expected points for those plays? I think that we would probably find that sacks would be the most valuable because of loss of yards followed by pass defended, then pressures but that by regressing to those combined stats a more complete picture could be created.