Tuesday, July 16, 2013

High Agility Pass Rushers!

Well, my computer decided to attempt suicide this week.  The official reason would seem to be due to bad wiring in my house, causing an electrical surge.  The more obvious explanation is that the computer is being overwhelmed by the awesomeness of my amateur stats, foresees a world in which moxie is no longer valued in relation to objective data, and committed hara-kiri to save mankind from my idiocy.  Either way, we are limping along.

I've been meaning to put this here as a follow up to the Explosive Pass Rushers post, but kept getting distracted.  While I place a greater deal of faith in pass rushers who demonstrate explosiveness and power though their Kangaroo Score, there is a place in the world for other types of players.  This second category of pass rushers, who succeed because of exceptional agility, is a bit more difficult to pin down, and the overall rate at which they become successes is probably a bit lower in general.  So, there is a bit more risk with drafting these players, but still a decent chance that things can work out.

One of the reasons why I prefer the explosive variety of pass rushers, and why I believe they succeed more often, is due to the simplicity of how their traits are employed.  Beating the opposing tackle at the snap (an explosive first step), or simply overpowering them, really isn't that complicated.  When all else fails, being able to fall back on raw power, is a good thing.  The simplest approach to a problem (killing the QB, in this case), tends to be the most reliable.  So, I suspect that the highly explosive pass rusher is more likely to use the simple bull rush a good deal of the time.  It might not be pretty, but it works.

On the other hand, a highly agile pass rusher is a bit more likely to execute spin moves, or juking their way around their opponent, or even the dreaded squirrel move.  Since added bulk tends to diminish your agility (though possibly add to your raw power), many of the high agility pass rushers tend to be smaller guys.  Staying out of the clutches of the tackle, is probably more of a necessity, as they generally won't have the power to break free from a locked on tackle.  Of course, these are all generalizations, and all players have some combination of explosiveness and agility, some just rely on one trait much more heavily than the other.

One of the best current examples of a successful and highly agile pass rusher, is the Packers' Clay Matthews.  A lot has been made of the exceptional 10 yard split of 1.49 seconds, that he produced in his forty yard dash, but I'm still not sure that this was the measurement that mattered.  As predictors go, I'm much more interested in how he did in the short shuttle and 3-cone drill.  His short shuttle time of 4.18 seconds is 1.014 standard deviations above the average for his pass rushing peers.  Similarly, his 3-cone time of 6.90 seconds, is 1.242 standard deviations above average.  This gives him a combined Agility Score of 1.128.  These are rather exceptional results.  On the flip side, his Kangaroo Score of -0.238, is a little below average.

So, the numbers would suggest that Matthews had some intriguing athletic traits, though he probably wouldn't thrive as a player who bull rushes a lot, which is probably the most basic and straight forward pass rushing move there is.  I think this idea becomes particularly interesting if we look at a recent article that Pro Football Focus wrote about him.  In examining his strengths and weaknesses, they suggested that his primary shortcoming might be as a bull rusher.  They claim that in the last two years, he has only produced 8 pressures while bull rushing, and only 3 pressures in the last year (one in every 137 pass rush attempts).  At the same time, they suggest that he is producing a pressure once in every 15.8 attempts, when rushing outside, and going around the tackle.  This is in no way meant to be a condemnation of Matthews, but I merely mention it to illustrate how he seems to be basing his style of play on his physical strengths, and abandoning completely any attempt to be something which he isn't physically designed to be (a mauling bull rusher).

The real problem with examining the high agility pass rushers, is that there really aren't that many of them.  Often times, a player might be highly agile, but also have a good Kangaroo Score, making it difficult to assign them strictly to the high agility group.  Among the players who seem to solely rely on agility, you get a handful of players who have occasional successes, but generally don't consistently produce sacks at the same level as their more explosive counterparts (Matthews is a rather large exception to that statement).  Still, I thought I'd post up a list of some players with highly exceptional agility scores, but who might also be somewhat lacking in terms of explosive power.  As always, all scores are given in terms of the number of standard deviations that a player's results are away from the average for his position group.

Player                                 Avg. TFL      Kangaroo     Sh. Shuttle      3-Cone      Tot. Agility
Bruce Irvin 14.5 -0.162 1.845 2.012 1.929
Von Miller 19.5 0.532 1.679 2.012 1.846
Channing Crowder 6.75 -0.057 0.793 2.398 1.595
Kyle Vanden Bosch 14 1.211 1.568 1.550 1.559
Demarcus Ware 14.25 0.759 1.624 1.435 1.529
Pisa Tinoisamoa              N/A -1.322 1.624 0.972 1.298
Melvin Ingram 13 0.028 1.014 1.512 1.263
Miles Burris 19.5 0.319 0.904 1.589 1.246
Trevor Scott 14.25 0.001 0.959 1.473 1.216
Sam Acho 15.5 -0.026 0.239 2.051 1.145
Brian Cushing 6.5 -0.219 0.793 1.473 1.133
Clay Matthews 6 -0.238 1.014 1.242 1.128
Jerry Hughes 18 0.155 1.181 0.895 1.038
Rob Ninkovich 13.25 0.267 1.014 1.011 1.013
Jason Babin 29.5 0.492 1.236 0.741 0.988
Kabeer Gbaja-Biamila              N/A -0.354 0.461 1.242 0.851
Chad Greenway 9 -0.634 0.793 0.818 0.806
Rosevelt Colvin              N/A 0.545 0.959 0.548 0.754
Chike Okeafor              N/A 0.392 0.904 0.510 0.707
Brooks Reed 7.25 -0.360 0.461 0.433 0.447

For the most part, this list was assembled by just sorting my pass rusher database, looking at their Total Agility score, and pulling out the players who have achieved some measure of acclaim.  Generally, I avoided including players who also had high Kangaroo Scores, since that isn't the purpose of this examination.  Still, I included a few such as Von Miller, Demarcus Ware, Kyle Vanden Bosch, Jason Babin, and Rosevelt Colvin, who had at least moderately above average Kangaroo Scores, as well as exceptional agility scores.  Compared to the Explosive Pass Rushers, this list is a bit more hit-and-miss in terms of producing consistent pass rushers.  I would say that, even here, as the player's Kangaroo Score moves towards the average result or above, their success rate does seem to improve.  Similarly, as their Avg. TFL stat goes up (their average number of tackles for a loss in their last two years in college), things likewise seem to improve.  In Clay Matthews' case, his Avg. TFL number is actually unreliable, since he wasn't a regular starter in his last two year, so that can be ignored for the most part.  Including Brian Cushing, who is really a MLB, might also seem a bit odd, but I just threw it in on a whim.

While there are some good football players in this list, if you truly eliminated all of the players with Kangaroo Scores that were below average (less than zero), it quickly becomes less intriguing.  Suddenly, it becomes a list of perhaps good, solid football players, who might possess some versatility, but could be a bit inconsistent at getting to the quarterback.  Their exceptional agility may still demonstrate itself in their ability to drop back into coverage, but they are somewhat less likely to end up producing double digit sacks.

Calculating the success rate of these sorts of players is a bit tricky.  In forming this sort of list, I passed over numerous players who have amounted to nothing in their careers.  Finding an objective measure for what we can call a success, is also a bit problematic.  Since many of the high agility players also tend to be lighter in weight, they also find themselves being moved to other positions quite frequently, such as 4-3 LBs or 3-4 ILBs, which adds another complication.  All of these issues make calculating the success rate for pure high agility pass rushers somewhat impossible.

Still, if I look at the complete list of pass rushing prospects (about 591 of them) from the last 13 years, and examine just the players with total agility scores over 1.000 (the exceptionally agile), and Kangaroo Scores that are no better than 0.500 (reasonably explosive, but not shockingly so), we can make some rough estimates based on the 44 players that remain (an unfortunately small sample size, but that's life).  While some of these 44 ended up moving to positions that weren't exclusively oriented towards pass rushing, I'm only concerned with whether they emerged as regular, and productive,  muti-year starters for their team.  So, with all that blathering out of the way, I would say that pure high agility, pass rushers become reasonable successes approximately 43.18% of the time.  Obviously, this is still really just spitballing.  Compared to the highly explosive pass rushers, who I think can probably be predicted to succeed around 65-82% of the time (though these players are actually not purely high explosive types, and may possess good agility also), this lower proposed success rate may not seem too impressive.  With that said, it is still comparable or better than the typical league wide success rate of about 20-30%, when selecting pass rushers and linebackers.  With some careful examination of the individual prospect, a success rate of over 50% is probably still quite achievable, though this is just wild speculation on my part.  After all, the idea of a 43.18% success rate is just based on how many have succeeded based on sorting them according to their agility scores alone, and merely including the data relating to their college success (Avg. TFL), should improve things a fair bit on its own.  Again, this is mindless speculation, but it seems reasonable to me.

I should also mention something that I found rather interesting, if not surprising, about this small sample group.  When I divided the 25 "failures", and the 19 "successes", into their separate groups, the differences between them became quite apparent.  Among the successes, their average Kangaroo Score was -0.026, which is merely average, but still significantly better than the -0.628 that was seen by the "failure" group.  Also, the Avg. TFL number amongst the successes was 12.52, compared to a 9.25 for the failures.  So, the more productive college players did indeed turn out better, and even though I tried to de-emphasize the explosiveness of the Kangaroo Score, the more explosive players still came out on top.

With all of this being said, I still prefer the idea of a player who exhibits a high degree of explosiveness, in combination with good agility, as this seems to be the best of both worlds, and has the highest likelihood of working out.  In combination with a high Avg. TFL number, to demonstrate some history of living up to their physical potential, things tend to turn out well with these sorts of draft picks.  With the more purely nimbly-toed player, I might be intrigued, but the idea of selecting them before the 3rd round, would make me a bit nervous.  As compelling a success as Clay Matthews has been, he is still a bit of an oddity.  I think Ted Thompson (GM for the Packers) is clearly one of the top 3 GMs when it comes to the draft, but the Clay Matthews pick could have turned out very differently.  His more recent selection of Nick Perry, another peculiar USC pass rusher, who was selected with the 28th pick in the 2012 draft, might end up serving as an interesting counter-argument to Matthews' success.

Similarly to Matthews, Perry played at USC, and accumulated comparable stats, as far as Avg. TFL are concerned.  Over his last two years, he averaged 10.25 TFL, compared to Matthews 6 TFL average   Unfortunately, Clay accumulated his stats in just 12 games started, while Perry had 21 starts.  So, Matthews had 0.5 TFL per game started, while Perry had 0.488 per game, very similar, and generally a fairly poor result (though in Matthews case it clearly turned out to be a poor predictor of future success).  Normally I am looking for first round pass rushers to have closer to an average of 15 TFL, or somewhere around 1.25 per game, but that's just me.  Either way, while we can see Matthews athletic measurables above, Perry was a similarly unbalanced athlete, though his strengths leaned much more heavily towards the explosive side, with an exceptional 1.870 Kangaroo Score, and a rather poor -0.874 Agility Score.  Whether Thompson's gamble on Perry will turn out quite as well will be interesting.  For the most part, I think Perry is unlikely to amount to much, though his numbers suggest he could perhaps do better if he moved to a 4-3 defensive end position, where his abysmal agility might matter a bit less.  Either way, you have a team making very similar gambles, on vaguely similar types of risks.  It would somewhat defy the odds for this to work twice, but we'll have to wait and see.  All of this is just meant to say, that while I get excited about explosiveness, a balanced approach is always the most appealing to me.

No comments:

Post a Comment