Thursday, May 9, 2013

Explosive pass rushers!

If you were drafting a 3-4 outside linebacker, or a 4-3 defensive end, how much would you really care about their ability to stop the run, if they were a hellacious pass rusher?  You probably wouldn't care very much.  Guys who can stop the run are a dime a dozen.  You could pick one up in the fifth round without much difficulty.  Teams struggle much more with finding guys who can turn the opposing quarterback into a paraplegic.

If you have read through the other nonsense I have posted here, what I'm about to say will seem very predictable.  Teams should draft freak athletes!  Look for guys who can jump really high!  Blah, blah, blah.  I'm even bored with myself.  Once again, we're going to pull out the old Kangaroo Score.  The Kangaroo Score, in this case, relates to a player's ability to effectively bull rush and their explosiveness when the ball is snapped.  There will also be a visit from the Agility Score, and a special guest appearance by the often neglected bench press.  The bench press results from the combine don't really matter too much, and will only represent 10% of a player's total score.  At some point I might get rid of the bench press altogether.  It really only serves to eliminate the freakishly weak.

First of all, we'll look at the 3-4 outside linebackers.  Generally speaking they will be a bit more agile than 4-3 defensive ends.  It should make sense that a 3-4 outside linebacker would benefit from better agility, since they will drop back into coverage with greater frequency than defensive ends.  Other than that, the differences are insignificant, regardless of what stupid things people say about "Oh, he won't fit this scheme".  It's all nonsense.  For the most part they are pretty interchangeable.  Basically, the lighter the player is, the more they will require/benefit from increased agility, so as to avoid offensive tackles who could maul them.  I will also be including the player's average number of tackles for a loss in their last 2 years in college.  I'll explain why a bit later.

3-4 Outside Linebackers
Player                     Weight      Bench        Kang. Score  Agility Score       Total    Avg. TFL
Shawne Merriman272251.9600.5061.35113.25
Connor Barwin256211.6421.1861.28216
Brian Orakpo263311.979-0.3121.24014
Justin Houston270301.5810.5061.22616.75
Cameron Wake236201.3710.8430.9968
Demarcus Ware251270.7591.5290.97917.25
Adalius Thomas270241.573-0.3060.855            18
Von Miller246210.5321.8460.81419.5
LaMarr Woodley266291.195-0.0750.80015.25
Kamerion Wimbley248241.0080.1430.6518.25
Ryan Kerrigan267310.8050.0070.63222.25

Hmm, that sure is a lot of exceptional athletes there.  I wonder if that is a coincidence?  Some difficult individuals might be tempted to say,"Hey, why does Connor Barwin have a higher total score than Demarcus Ware?  That can't be right."  These people are assholes, who want to make my life difficult.  Yes, you.  The scores themselves don't really matter that much, as odd as that may sound.  All we care about is whether a player crosses certain thresholds when it comes to being more explosive and agile than his average peer.  These guys are.. so shut up.

4-3 Defensive Ends
Player                     Weight     Bench      Kang. Score    Agility Score        Total    Avg. TFL
Mario Williams295353.1230.0052.10419.5
JJ Watt290342.3081.0841.91818.25
Kyle Vanden Bosch270261.2111.5591.23814
Brian Robison259271.3830.9261.17212
Chris Long272241.2380.8140.99014.25
Will Smith275301.629-0.4540.96716.25
Aaron Schobel263211.0981.0430.912         N/A
Jevon Kearse262241.0830.7690.88310
Andre Carter249201.2300.5770.83119.5
Jason Babin260280.4920.9880.67729.5
Justin Tuck265241.060-0.0040.63716.5
Robert Quinn265220.9680.0760.56512.75
Greg Hardy281211.370-0.8460.50914.25
Rob Ninkovich260230.2671.0130.44613.25

What's that you say?  These guys look pretty athletic too?  Hmm, I'll have to look into that.  Surely it is just a coincidence.  In some ways, Brian Orakpo and Adalius Thomas should fit in better here, since their agility score were a bit lower.  Some of these players like Brian Robison, Kyle Vanden Bosch, Chris Long, etc. would likewise probably prove capable of playing quite well in a 3-4.  People would shriek and squeal,"No, Chris Long wouldn't fit in that scheme."  Why not?  He's about as agile as Cameron Wake, and no heavier than Shawne Merriman or Justin HoustonJason Babin really should be in a 3-4, since he doesn't have great explosive power, but has excellent agility.  This might be why he briefly thrived in the nonsensical "wide 9" defense in Philly.  It let him line up very wide outside the offensive tackle (similar to how he would line up in a 3-4) and avoid some physical confrontations.  The same thought probably applies to Rob Ninkovich.

Yes, J.J. Watt is really a 3-4 defensive end, which is more like a defensive tackle (sort of like Mario Williams who I also included), but I can't resist posting his numbers everywhere I possibly can.  They are amazing!

Yes, this is leaving out quite a few guys who scored well, but who became busts (it also leaves out some who were just "good").  This is why I include the Avg. TFL here.  If you additionally filter players who had a low number of TFL it eliminates most of the athletic guys who became failures.  In general, I sort guys like this.  If they have a total score over 0.500 and averaged 15 TFL in their last two college years, they get a first round grade.  Then the TFL requirement goes down to 14 for the second round, 13 for the third, etc.  If you stick to that your likelihood of success should fall somewhere between 65-82%, regardless of the round in which a player is taken.  Of course, this does require pursuing players that the computer suggests are the most valuable targets, but these players are also frequently available later than you might suspect.  The league-wide average success rate is somewhere between 20-30%.  The reason I give a range of percentages is because precisely defining success is tricky here, and this covers the strictest and loosest possible definitions of the term that I can imagine.  This is based on examining 702 linebackers and defensive ends, so sample size isn't really an issue.  Once you get to the third round, I think it is okay to start taking guys with total scores between 0.300 and 0.500, but the TFL meter gets reset.  So, 15 TFL for the third round, 14 TFL for the fourth, etc. 

So far, using this system has only resulted in one potential first round pick who turned out to be a bust.  Unfortunately that player was Vernon Gholston.  What can you do?  On the flip side, the system does alert you to an undrafted player like Cameron Wake, as a very interesting prospect.  I have no idea how teams weren't interested in someone as athletically gifted as him.  Sure, he wasn't very productive in college, but that sort of athleticism seems worth a late round pick.

Now let's look at some of the stupidest moments in recent NFL history.  Who were some highly drafted players that haven't turned out to be worth much?

Player                     Weight    Bench        Kang. Score    Agility Score     Total       Avg. TFL
Courtney Upshaw27222-0.754-0.843-0.74315.75
Brandon Graham26831-0.1200.7220.29123
Derrick Morgan266210.2250.0120.07914
Tyson Jackson29620-0.022-2.013-0.696         N/A
Aaron Curry254250.811-0.2660.43014.75
Aaron Maybin249220.705-0.6190.19912.25
Larry English25524-0.272-0.118-0.19616.5
Robert Ayers27217-0.811-1.056-0.94413.75
Everette Brown25626-0.459-1.092-0.55916.5
Derrick Harvey27131-0.320-0.082-0.07015
Quentin Groves259300.531-0.3250.34713
Jarvis Moss25016-0.517-0.105-0.50310

The guys in this group have two things in common. They were all taken in the first two rounds, and they all suck.  Aaron Curry almost has a passable score, but almost passable isn't what you want to see in a first round pick.  The computer would have viewed him as a 3rd rounder.  While many of them had impressive statistical production in college (high TFL numbers, in this case), their physical measurables suggested that there was a reasonable likelihood that this dominance wouldn't continue at the NFL level.  I also realize that it is awfully early to call Courtney Upshaw a bust, but I'm willing to go out on limb with that hunch.

There is an additional category of pass rushers, that relies very heavily on high agility scores, which I've now written about here.  I'll also try to get into some of the weird pass rushing outliers too.


  1. Wow upshaw's a bust...? I'm a ravens fan myself and i hardly beleive he'll bust..... I mean sure he didn't exactly do much his 1st year but c'mon it was his rookie year! Rookie year's mean nothing (unless ur a QB) he was getting used to being a 3-4 OLB he'll be a situational pass rusher until suggs is allowed to leave once his contract is up. Then he'll start next to dumerville. (or john simon if he becomes the next jarret johnson like i think he will)

  2. Just to be clear, I am not jumping to conclusions about Coutney Upshaw based on his rookie year. No, I wrote him off before he was even drafted. We'll just have to wait and see what happens. While I have my (STRONG) doubts about Upshaw, I do think John Simon could become an interesting player. Probably not a star, but perhaps solid and serviceable.

  3. Love the fact quentin groves,Aaronn curry,aaron maybin we're supposed to be hall of famers according to your score.... You love TFL upshaw has the 3rd most on your list that show's he an aggressive run stuffing linebacker who can get behind the line of scrimmage. For a high second round pick who rotates in the starting lineup in a 3-4 Defense who contributes say 5-7 sacks a year (I don't think he'll ever be an elite pass rusher but he'll be a solid pass rusher) Look upshaw proved to be a solid starting linebacker last year (as a rookie he was solid considering the defense was sh** for most of the season) i think he'll be solid with a chance to grow (Much like joe flacco)

    1. Actually, the numbers suggest that Quentin Groves, Aaron Curry, and Aaron Maybin were unlikely to be anything special. Maybe I just misunderstood your comment. While some of them may have done well in one of the categories I measure, overall all three of them had scores that were average at best. Aaron Curry had the best combination of physical traits (0.430 Total Score), and college productivity (14.75 Avg. TFL), but still didn't score highly enough to merit a first round pick. According to this system, Curry should have been, approximately, a 3rd round pick. Still, he has probably become the most productive out of the "busts" I listed him with.

      With Upshaw, you are right about his college productivity. It was fairly impressive. Still, that is only part of the puzzle for me, and athletically he was beyond just bad, according to what we can measure. I wouldn't be disappointed at all if he turns out to be good, I just don't think that is very likely to happen. As far as his ability to stop the run, that is obviously a good skill to have, but I don't think it is the primary trait a team should look for in their 3-4 OLBs or 4-3 DEs. It is kind of like saying that a wide receiver is good at blocking for his running back. Yeah, that's cool, but a WR's main job is still about catching the ball. I just think that pass rushing should be the primary skill for a 3-4 OLB or 4-3 DE. Like I said, I wish Upshaw the best, but I think if the team really had confidence in him they wouldn't have pushed him into a backup role by signing Dumerville. That Upshaw didn't noticeably outperform Albert McClellan (an undrafted OLB) also concerns me. It will take time for this all to sort itself out, so, we'll see what happens.

      You're right though, last year's defense was pretty horrible for the Ravens, so judging Upshaw becomes more tricky. If it makes any difference, I wouldn't be the least bit surprised if they are significantly better this year. The offensive line, on the other hand, makes me a bit nervous.

  4. The offensive line should be fine (guard wise.....) Center is an issue.... but gino did do very well against geno atkins in the final game. A.Q shipley graded out pretty good as the colts center when he started (he was traded for because our only backup center was a rookie and to provide real competition) LT should be okay-ish...... Mckinnie can't be any worse than oher.... Ever since his rookie campaign oher has sucked.... idk whether it's the o-line coach,scheme,weight or something..... but he looks like a mediocre-sucky RT. He has brilliant stretches.... (didn't give up a sack to demarcus ware bobbie williams did) Then...... ka-boom i remember when he had like 3 straight penalties on a drive..... my god he sucks...and 3-4 OLB's main quality isn't rushing the passer you have to be good in coverage and stop the run which upshaw can do. Everetee brown busted because of scheme fit... (3-4 OLB not a 4-3 DE) Upshaw isn't the most gifted athletically...... SAM linebacker (the position which upshaw played most of last year) Is basically a stopping the run,being good in coverage type of role. Look at jarret johnson.... was he a great pass rusher? Was he overly athletic? was he oversized? no he wasn't big he wasn't fast and he was iffy at rushing the passer.

    1. The interior of the line could be interesting. I'm excited to see what Shipley will do, since the computer is a fan of his, and he seems to have done well when given an opportunity. McKinnie is always a wild card, who seems to vary according to how interested he is in giving an effort. The computer suggested that Oher was likely to be mediocre, so I'm not surprised by what he has become. Hopefully he can step up this year.

      I guess we'll have to agree to disagree about 3-4 OLBs, to some degree. I'm not disagreeing that dropping into coverage and stopping the run matter, I just place less importance on it. I know we probably won't agree on this, but I never had a high opinion of Jarret Johnson. He was likable, hard working, and tough, but I never felt his impact on the team was that significant. Still, I can see why he was a fan favorite, and probably a good teammate. I also haven't seen much to suggest that he or Upshaw were ever particularly good in coverage, though this doesn't concern me too much.

      As for Everette Brown, it's fine to suggest that the scheme was the problem, but without seeing him succeed in another scheme, that is just speculation. At this point, all we know is that he failed to produce much, and that fits with what the computer expected.