Nick Driskill is a player I wanted to mention, even though there is
an extremely good chance that nobody will ever hear from him again. He
played strong safety at Mt. Union. He didn't get drafted, but was
picked up as an undrafted free agent by the Colts. I really have no
idea what to make of him, but some of the numbers that make up his story
are just fascinating.
Here are his stats from college:
TKL SACK TFL INT Passes Broken Up FF
2012 101 3 14.5 2 11 3
2011 113 6.5 17 4 9 1
2010 80 3 13.5 5 4 2
2009 43 0.5 6.5 1 3 1
Those
are the kinds of stats that should only exist in a video game. They
are absolutely silly. Granted, he played at a rather goofy school,
though Cecil Shorts and Pierre Garcon also went there. Even if you
wanted to downplay his production, how could you? He dominated every
defensive statistical category there is, at least by the standards of a
defensive back. How many defensive backs finish school with 13 sacks,
12 interceptions, 7 forced fumbles, and two seasons of triple digit
tackles? That doesn't even mention his 51.5 tackles for a
loss. He must have been constantly killing people behind the line of
scrimmage, because a lot of defensive ends who get taken in the first
round don't have those kinds of stats. His opponents at Heidelberg
University, Muskingum University, and Baldwin Wallace, must have been
trembling in their boots (wherever the hell these schools are)!
I
don't know how he accumulated these stats. Watching Mt. Union play
isn't something I get much opportunity to do (and is probably something I
would avoid). Were they running some peculiar defense, with the coach
saying, "Funnel all the plays to Driskill! I've got to make that kid a
star to win a bet with my buddy at Marietta!". I have no idea. This is
the only clip I can find of him playing, and it doesn't tell you very
much.
There is a catch though.
He is 5' 8.5" tall (that half inch matters here!), and 197 pounds. He
is a bit on the short side. You definitely aren't going to want him
covering tightends. His pro day results (since guys like him don't get
invited to the combine) were merely average, except for his vertical
jump, broad jump and bench press. Strangely, the Mayor of Munchkinland's bench press results were better than what most offensive linemen achieve, though this is of questionable value. The numbers sort of tell the story of
a guy who is explosive, quick and powerful, but not necessarily fast or
exceptionally agile. Here are his results, along with some other
comparable midgets.
Player Ht Wt 40-yd 10-yd BP Vert BJ Shrt-Shut 3-cone
Nick Driskill 5' 8.5" 197 4.53 1.53 29 43 10'8" 4.16 7.00
Cortland Finnegan 5' 9.5" 188 4.34 ? 14 44 10'8" 4.34 6.95
Brent Grimes 5' 9.2" 177 4.57 ? 4 33.5 10' 1" 4.25 7.13
Dre Bly 5' 9.7" 186 4.51 1.58 ? 34.5 9'6" 4.09 7.11
Bob Sanders 5' 8.3" 204 4.35 1.50 15 41.5 10' 8" 3.95 6.70
Some
of these players are clearly faster. Some are more nimble. Almost all
of them are oddballs with significant physical shortcomings, but also
some areas of strength. The player I would really like to compare him to is Jim Leonard, who was also a very dynamic toadstool, but I don't have combine data for him. Honestly, I probably wouldn't have bet on any of theses guys
achieving much success. In Nick Driskill's case, I still probably
wouldn't bet on him. In three years, I would place the odds strongly in
favor of him becoming a special teams player at best, or a Winnipeg
Blue Bomber at worst. Still, even though it runs contrary to my normal
views on players like him, I don't want to bet against him. Those ridiculous stats of his, as well as my love of a good vertical jump, just intrigue me too much.
I'm going to take the cowardly approach. I'm mentioning him in hopes that he becomes something special, in which case I will look like a clever boy. Yet, I am kind of betting against him, because I am feeling a bit gutless and know the odds are against him. If I do this a hundred times, eventually something prophetic will come out of my mouth, and then I'll just need to delete every post where I said something stupid.
Oh, and he was a biology major. So, perhaps with a little tweaking we could nudge a chromosome here and there, to make him a bit taller and faster.
Sunday, May 5, 2013
Saturday, May 4, 2013
Athleticism and the offensive line
People often say that offensive linemen are the safest picks in the first round. This may be true. I was looking at a list of linemen who had been taken in the first two rounds, from 1999 to 2012, and the average player selected in this range ended up starting in about 70% of the games during their career. Player's taken in just the first round started 75% of their games. Of course, this doesn't mean that they played well. I really don't want to attempt to judge that, but they at least managed to get on the field quite a lot.
The thing that strikes me as odd about this, is that offensive linemen should be a somewhat trickier position to evaluate. Like every player the team's have their combine data to consider. Unlike other players, offensive linemen don't provide any statistical production. If teams fail at a rather high rate when drafting quarterbacks and wide receivers with high draft picks, despite having more measurable information on them, how are they apparently doing much better at drafting offensive linemen, with less information? Maybe they're not.
Isn't it possible that they are screwing up on offensive linemen, just as much as they do at other positions, but it is just harder to notice? When a receiver repeatedly drops the ball, or a quarterback throws a harebrained interception, even the most casual fan will notice. With an offensive linemen, however, it will probably take repeated and glorious failures to make the headlines. Sure, you would think that teams would notice, even if we aren't paying close attention. Unfortunately, I think we are all familiar with some of the human turnstiles certain teams continue to employ. Maybe team's just develop blind-spots for certain guys, who they clearly believed in when they got drafted, and for whom they continued to have high hopes for later success. All I'm suggesting is that maybe some of these "safe" draft picks, are safe because they are so hard to criticize. After all, how successful is a newspaper article going to be if it spends its time analyzing a lineman's technique?
I'm not claiming to have any answers, but I enjoy examining combine data, and think there might be some trends worth noting. Initially I would have thought that successful interior linemen would have shown exceptional Kangaroo Scores, as I often envision them as the more brutish and less nimble in-line blockers. I expected offensive tackles to be heavily reliant on a good agility score (from the 3-cone drill and the short shuttle). On both counts I appear to have been somewhat wrong, and the complete opposite appears to be the case.
While the Kangaroo Score probably does tell you quite a bit about a player's explosive power, it also seems to relate to their quickness. Picture an offensive tackle dropping back at the snap of the ball. Though he is moving backwards, the degree to which he does this explosively is probably going to give him an edge against high quality speed rushers. Moving forward, this same explosiveness turns them into a run blocking force. So, the Kangaroo Score comes into play. Still, with high end left tackles you do see that their agility scores come to matter significantly more than for right tackles. The average results for a tackle who makes it to a Pro Bowl or All Pro roster are a 0.721 Kangaroo Score, and a 0.233 Agility Score. For just the left tackles, it was a 0.634 Kangaroo Score, and a 0.476 Agility Score.
Interior linemen, it seems, probably wind up playing inside due their lack of explosion in comparison to tackles. Not they can't be good in this department, but it tends to be less common. The best interior linemen do seem to compensate though, with excellent agility. Perhaps this agility gives them leverage, and that becomes the source of the power that they might otherwise lack. I don't know if that is the case. It's just an idea I am kicking around. Then we come to the high end centers that tend to have shockingly good short shuttle times, somewhere in the area of one standard deviation above average. The average results for interior linemen that make it to a Pro Bowl or All Pro roster are a 0.075 Kangaroo Score, and a 0.288 Agility Score. So, yes, the interior guys tend be somewhat less dynamically athletic, but a lot of that is because they are being graded in the same group as the tackles, who just throw off the curve for everyone else.
Since I am proposing that some highly drafted players might be sticking around as starters, when perhaps they shouldn't, let's take a look at some of the generally acknowledged busts from the first two rounds of the draft.
Player Position Pick # Kangaroo Score Agility Score
Marcus Johnson Tackle 49, in 2005 0.208 0.074
Toniu Fonoti Guard 39, in 2002 0.863 -0.762
Adam Terry Tackle 64, in 2005 0.009 -0.324
Jason Smith Tackle 2, in 2009 -1.351 0.575
Eben Britton Tackle 39, in 2009 -0.384 -0.565
Chris Williams Tackle 14, in 2008 -0.475 -0.559
Chilo Rachal Guard 39, in 2008 0.354 -1.401
Levi Brown Tackle 5, in 2007 -0.560 -0.516
Vladimir Ducasse Tackle 61, in 2010 -0.630 -1.384 (too soon?)
For the most part, these players do seem to be showing scores that are average at best, and horrific in most cases. When you compare them to the average results for Pro Bowl players at their listed positions, they all fall rather short of the mark. There are more guys who I suspect should be considered busts, but teams keep starting them. I will throw out this idea though. If you had a quality left tackle, would you ever let him go, or trade him? Yet two teams in 2013 either did this, or attempted to do this, with these players:
Player Position Pick # Kangaroo Score Agility Score
Jake Long Tackle 1, in 2008 -0.240 0.636
Branden Albert Tackle 15, in 2008 0.037 -0.423
I'm not saying that they are bad, merely that their numbers are somewhat mediocre. This might have something to do with their up and down careers. The fact that the Dolphins were supposedly in trade talks for Branden Albert, as a replacement for Jake Long, would appear to be a lateral move at best.
Once you get past the players who were taken in the first couple of rounds, where teams are heavily invested in proving they made the right pick, things become a bit more interesting. If you look at late round or undrafted players, who went on to have success, you see that the majority of them do demonstrate high levels of athletic ability. When compared to the average Pro Bowl player's results, they do quite well. This shouldn't be surprising since many of these players are the Pro Bowlers who set the standard in the first place. So, despite playing for a team that probably had relatively little faith in them, they made themselves impossible to ignore.
Player Position Pick # Kangaroo Score Agility Score
Jason Peters Tackle Undrafted 2.278 -0.005
Chris Myers Guard/Center 200, in 2005 0.213 1.727
Scott Wells Guard/Center 251, in 2004 0.278 1.233
Chris Kuper Guard 161, in 2006 0.172 0.640
John Sullivan Center 187, in 2008 -0.251 0.759
Carl Nicks Guard 164, in 2008 1.032 -0.072
Kyle Kosier Guard/Tackle 249, in 2002 -0.289 1.402
Matt Slauson Guard 193, in 2009 1.337 -0.425
Brian Waters Guard Undrafted 1999 -0.027 0.740
Alex Boone Tackle Undrafted 2009 0.396 0.044
Eric Heitmann Center 239, in 2002 -0.393 0.743
Yes, there are some highly drafted players with excellent numbers who have failed to live up to their potential. Alex Barron (2.317 Kangaroo Score and 0.338 Agility Score) and Winston Justice (2.368 Kangaroo Score and 1.584 Agility Score) are perfect examples of this. I'm not sure what you can do to avoid or explain that sort of situation. Some guys just don't live up to their potential. There are also undoubtedly numerous players with poor measurables who have done quite well. Still, if we take a look at two different ends of the offensive line spectrum (based on 2012 depth charts), this is what we see:
Player Position Pick # Kangaroo Score Agility Score
Patriots
Nate Solder Left Tackle 17, in 2011 1.281 1.592
Logan Mankins Left Guard 32, in 2005 -0.370 1.146
Ryan Wendell Center Undrafted N/A N/A
Dan Connolly Right Guard Undrafted 0.274 0.695
Sebastian Vollmer Right Tackle 58, in 2009 1.748 1.076
Maybe this has a little something to do with how the Patriots get away with using mediocre receivers and running backs. That is a shockingly athletic line, even if we don't know what Wendell's scores would be.
Lions
Gosder Cherilus Left Tackle 17, in 2008 -0.738 0.133
Stephen Peterman Left Guard 83, in 2004 -1.245 -0.079
Dominic Raiola Center 50, in 2001 0.789 1.373
Rob Sims Right Guard 128, in 2006 0.197 0.383
Jeff Backus Right Tackle 18, in 2001 -0.646 -0.914
So, which of these offensive lines would you expect to do better?
As for the 2013 draft, I will leave the top three tackles alone for now, but I want to point out one player. People kept claiming this guy was exceptionally athletic, and I just couldn't figure out why. So it will be interesting to see how it turns out in the long run, since even if he fails it could be years before anyone admits it.
Player Position Pick # Kangaroo Score Agility Score
Menelik Watson Tackle 42, in 2013 -0.732 -1.589
If you're still interested in all of this nonsense, you can jump over to Athleticism and the Offensive Line part 2. In that post I compare a player's athletic measurements to their success/productivity base on their CarAV score .
The thing that strikes me as odd about this, is that offensive linemen should be a somewhat trickier position to evaluate. Like every player the team's have their combine data to consider. Unlike other players, offensive linemen don't provide any statistical production. If teams fail at a rather high rate when drafting quarterbacks and wide receivers with high draft picks, despite having more measurable information on them, how are they apparently doing much better at drafting offensive linemen, with less information? Maybe they're not.
Isn't it possible that they are screwing up on offensive linemen, just as much as they do at other positions, but it is just harder to notice? When a receiver repeatedly drops the ball, or a quarterback throws a harebrained interception, even the most casual fan will notice. With an offensive linemen, however, it will probably take repeated and glorious failures to make the headlines. Sure, you would think that teams would notice, even if we aren't paying close attention. Unfortunately, I think we are all familiar with some of the human turnstiles certain teams continue to employ. Maybe team's just develop blind-spots for certain guys, who they clearly believed in when they got drafted, and for whom they continued to have high hopes for later success. All I'm suggesting is that maybe some of these "safe" draft picks, are safe because they are so hard to criticize. After all, how successful is a newspaper article going to be if it spends its time analyzing a lineman's technique?
I'm not claiming to have any answers, but I enjoy examining combine data, and think there might be some trends worth noting. Initially I would have thought that successful interior linemen would have shown exceptional Kangaroo Scores, as I often envision them as the more brutish and less nimble in-line blockers. I expected offensive tackles to be heavily reliant on a good agility score (from the 3-cone drill and the short shuttle). On both counts I appear to have been somewhat wrong, and the complete opposite appears to be the case.
While the Kangaroo Score probably does tell you quite a bit about a player's explosive power, it also seems to relate to their quickness. Picture an offensive tackle dropping back at the snap of the ball. Though he is moving backwards, the degree to which he does this explosively is probably going to give him an edge against high quality speed rushers. Moving forward, this same explosiveness turns them into a run blocking force. So, the Kangaroo Score comes into play. Still, with high end left tackles you do see that their agility scores come to matter significantly more than for right tackles. The average results for a tackle who makes it to a Pro Bowl or All Pro roster are a 0.721 Kangaroo Score, and a 0.233 Agility Score. For just the left tackles, it was a 0.634 Kangaroo Score, and a 0.476 Agility Score.
Interior linemen, it seems, probably wind up playing inside due their lack of explosion in comparison to tackles. Not they can't be good in this department, but it tends to be less common. The best interior linemen do seem to compensate though, with excellent agility. Perhaps this agility gives them leverage, and that becomes the source of the power that they might otherwise lack. I don't know if that is the case. It's just an idea I am kicking around. Then we come to the high end centers that tend to have shockingly good short shuttle times, somewhere in the area of one standard deviation above average. The average results for interior linemen that make it to a Pro Bowl or All Pro roster are a 0.075 Kangaroo Score, and a 0.288 Agility Score. So, yes, the interior guys tend be somewhat less dynamically athletic, but a lot of that is because they are being graded in the same group as the tackles, who just throw off the curve for everyone else.
Since I am proposing that some highly drafted players might be sticking around as starters, when perhaps they shouldn't, let's take a look at some of the generally acknowledged busts from the first two rounds of the draft.
Player Position Pick # Kangaroo Score Agility Score
Marcus Johnson Tackle 49, in 2005 0.208 0.074
Toniu Fonoti Guard 39, in 2002 0.863 -0.762
Adam Terry Tackle 64, in 2005 0.009 -0.324
Jason Smith Tackle 2, in 2009 -1.351 0.575
Eben Britton Tackle 39, in 2009 -0.384 -0.565
Chris Williams Tackle 14, in 2008 -0.475 -0.559
Chilo Rachal Guard 39, in 2008 0.354 -1.401
Levi Brown Tackle 5, in 2007 -0.560 -0.516
Vladimir Ducasse Tackle 61, in 2010 -0.630 -1.384 (too soon?)
For the most part, these players do seem to be showing scores that are average at best, and horrific in most cases. When you compare them to the average results for Pro Bowl players at their listed positions, they all fall rather short of the mark. There are more guys who I suspect should be considered busts, but teams keep starting them. I will throw out this idea though. If you had a quality left tackle, would you ever let him go, or trade him? Yet two teams in 2013 either did this, or attempted to do this, with these players:
Player Position Pick # Kangaroo Score Agility Score
Jake Long Tackle 1, in 2008 -0.240 0.636
Branden Albert Tackle 15, in 2008 0.037 -0.423
I'm not saying that they are bad, merely that their numbers are somewhat mediocre. This might have something to do with their up and down careers. The fact that the Dolphins were supposedly in trade talks for Branden Albert, as a replacement for Jake Long, would appear to be a lateral move at best.
Once you get past the players who were taken in the first couple of rounds, where teams are heavily invested in proving they made the right pick, things become a bit more interesting. If you look at late round or undrafted players, who went on to have success, you see that the majority of them do demonstrate high levels of athletic ability. When compared to the average Pro Bowl player's results, they do quite well. This shouldn't be surprising since many of these players are the Pro Bowlers who set the standard in the first place. So, despite playing for a team that probably had relatively little faith in them, they made themselves impossible to ignore.
Player Position Pick # Kangaroo Score Agility Score
Jason Peters Tackle Undrafted 2.278 -0.005
Chris Myers Guard/Center 200, in 2005 0.213 1.727
Scott Wells Guard/Center 251, in 2004 0.278 1.233
Chris Kuper Guard 161, in 2006 0.172 0.640
John Sullivan Center 187, in 2008 -0.251 0.759
Carl Nicks Guard 164, in 2008 1.032 -0.072
Kyle Kosier Guard/Tackle 249, in 2002 -0.289 1.402
Matt Slauson Guard 193, in 2009 1.337 -0.425
Brian Waters Guard Undrafted 1999 -0.027 0.740
Alex Boone Tackle Undrafted 2009 0.396 0.044
Eric Heitmann Center 239, in 2002 -0.393 0.743
Yes, there are some highly drafted players with excellent numbers who have failed to live up to their potential. Alex Barron (2.317 Kangaroo Score and 0.338 Agility Score) and Winston Justice (2.368 Kangaroo Score and 1.584 Agility Score) are perfect examples of this. I'm not sure what you can do to avoid or explain that sort of situation. Some guys just don't live up to their potential. There are also undoubtedly numerous players with poor measurables who have done quite well. Still, if we take a look at two different ends of the offensive line spectrum (based on 2012 depth charts), this is what we see:
Player Position Pick # Kangaroo Score Agility Score
Patriots
Nate Solder Left Tackle 17, in 2011 1.281 1.592
Logan Mankins Left Guard 32, in 2005 -0.370 1.146
Ryan Wendell Center Undrafted N/A N/A
Dan Connolly Right Guard Undrafted 0.274 0.695
Sebastian Vollmer Right Tackle 58, in 2009 1.748 1.076
Maybe this has a little something to do with how the Patriots get away with using mediocre receivers and running backs. That is a shockingly athletic line, even if we don't know what Wendell's scores would be.
Lions
Gosder Cherilus Left Tackle 17, in 2008 -0.738 0.133
Stephen Peterman Left Guard 83, in 2004 -1.245 -0.079
Dominic Raiola Center 50, in 2001 0.789 1.373
Rob Sims Right Guard 128, in 2006 0.197 0.383
Jeff Backus Right Tackle 18, in 2001 -0.646 -0.914
So, which of these offensive lines would you expect to do better?
As for the 2013 draft, I will leave the top three tackles alone for now, but I want to point out one player. People kept claiming this guy was exceptionally athletic, and I just couldn't figure out why. So it will be interesting to see how it turns out in the long run, since even if he fails it could be years before anyone admits it.
Player Position Pick # Kangaroo Score Agility Score
Menelik Watson Tackle 42, in 2013 -0.732 -1.589
If you're still interested in all of this nonsense, you can jump over to Athleticism and the Offensive Line part 2. In that post I compare a player's athletic measurements to their success/productivity base on their CarAV score .
Friday, May 3, 2013
Why people love statistical outliers
Football is a sport. Sports are played by athletes (unless it is bowling). The NFL Combine puts people through different tests of athletic ability. Then at the end of this testing, we gather together and say "Hey, you know that short slow guy? He's the one I want. The guy has moxie." Then a year or two later, when our man of determination and heart has been run off to play in the CFL, we try again. This time, we'll get it right. That other guy? Hell, he didn't have half the heart this new guy has. This new guy really brings his lunch pail to work (really, why the fuck do we say this?). Yup, it's going to be different this year.
Sure, being a hard worker or a guy with great character is a good thing. Do you think you can identify this trait? Was Randy Moss a hard worker? Or what about Terrell Owen's character? "Bah", you say,"They put up great stats, but I wouldn't want those cancerous guys on my team." Sure, sure. I get it. It's a team sport. It takes all 53 guys to win not just a couple guys who put up pretty stats. Okay then, so how are we going to identify these guys with character, that everyone seems to like so much? Do you have a Moxie-Meter hidden in your basement? A Lunch-Pail-O-Graph device that we can attach to their skull with electrodes, to measure this essential gritty determination? Even a Balboa-gauge would do in a pinch, if you have one on hand.
I suspect that no matter what team you root for, you have been in this position at least once. Your team has selected a wide receiver with a fairly high draft pick, only to find out that the guy just really can't catch the ball very well (Troy Williamson, Travis Taylor, etc.). So, somehow we are supposed to believe that NFL GMs can get an accurate read on a player's psyche, level of motivation, commitment, or whatever, when they can't even tell if a guy can catch. (which would seem like a fairly obvious trait to pay attention to, but perhaps gets overlooked while they are busy reading their DSM to analyze the player's psychological profile). Sure, I'm certain that most of these team's scouts, who were generally just failed former football players themselves, have keen analytical minds, highly trained and attuned to detect the faintest whiff of moxie in the air. Hell, that is probably why they stopped playing football in the first place. You don't want to damage a mind like that, with blows to the head.
Now, lets say we were working at NASA, and were looking to hire a new guy. We've got applications from all the best and brightest, from the finest universities in the land. If I turned to you and said, "You know what Bob? This guy here is the one we want. Sure, on paper he doesn't look so great. In fact he appears to be brain damaged. But you see, Bob, he just doesn't test well. I've talked to this guy and he's a real crackerjack. Sharp as a tack, I tell ya'. He's the guy we want to plan our mission to Mars." Well, best case scenario you would be laughed at. Worst case? Well, I suppose that would involve having the space shuttle go off course, and instead of heading to Mars, crashing into the White House (maybe I can pitch this idea to Michael Bay).
Yes, test results should matter to you. I know, we all hate to feel as if we are confined or trapped by some measurement, whether it is I.Q., 40 yard dash time, or whatever. We like to look in the mirror and see past that receding hairline, growing paunch, and strange rash in our armpit that just won't go away, to see the vigorous and brilliant man we know that we really are. Women just can't resist us. Oh, they play coy, but they notice us. As for other men, they tremble before our intimidating masculinity. No stupid number is going to tell us what we are, and it shouldn't do that to any man. This is America for god's sake, and no commie numbers are going to change that. USA! USA! USA!
So, basically, what I am trying to say is we tend to be morons. We're not all beautiful little snowflakes. Potatoes are all unique too, though nobody ever seems to compare themselves to one. We're not all just a step away from having our greatness discovered. We're not all going to be that outlier, who is going to defy expectations, but we still root for it to happen to someone. Somewhere inside us, we know that some guys are just better, more talented, but we don't have to like it.
Then we see a receiver run a 4.7 second forty yard dash. We pause. We think about all the great things that his coaches said about him. How his teammates would follow him through the gates of hell. He was the star of the Rose Bowl for Christ's sake! It's those damned commie numbers, coming to get us again! Anquan, he'll save us! Anquan Boldin ran a 4.72 forty yard dash, and things turned out great for him. He was gritty and tough wasn't he? That's all you need, man. Just give me some gritty tough guy, and you can keep those commie numbers of yours. Hell, look at Wes Welker. That sumbitch was 5'8" and running a 4.65 forty. Takes a lot of grit to be small and slow, doesn't it?
Sure those guys are great. They are anomalies. Outliers. Basically, they are like a cancerous tumor spreading through an otherwise healthy statistical analysis. They show up in your spreadsheet. They scare you a bit. You can't figure out what to do with them, or how to make them go away. So you learn to just accept them for what they are, a pain in your statistical ass (yes, statistical ass cancer). Though it is reasonable to find them oddly fascinating, it doesn't mean you should spend your life searching for them. That would inevitably prove rather depressing, sort of like a daily colonoscopy.
People might say that Wes Welker was underrated, or that he should have been drafted higher (actually he wasn't drafted at all). How highly should you draft a small slow wide receiver, with good but not exactly shocking college production (especially when they come from Texas Tech where everybody accumulates decent stats)? Maybe good old Wes was just hungover at the combine? Maybe Anquan didn't have his bran muffin that morning, allowing him to drop the ballast needed to increase his vertical jump? Okay, fine.
If you want to go looking for the next Wes Welker or Anquan Boldin, that's cool. I admire the quest that you are setting for yourself. I do think there are probably some guys out there who have a good eye for drafting talent, and maybe they will spot the next big thing. Still, at least with receivers (though this applies everywhere else too), most teams haven't shown that they are employing these savants in their scouting department. So, you can try to find the next guy who is going to defy all the odds, and over the next ten years you might find one. If you do manage to turn one up, I'd be the first to congratulate you. But can you do it again? If some scout out there can do that, say 60-70% of the time, then I would be very intrigued. If a team can't do this, then that underdog draft pick starts to look like a fluke, just blind luck. Or perhaps, they did have a brief moment of insight, where their internal Moxie-Meter went off, only to have it again go on the fritz for the following decade. I certainly can't rule that out.
Or, you can embrace what seems to be the most sensible solution, that those other guys, the bigger faster stronger ones who statistically pummeled their opponents all through college, might just be a better bet. It's just putting the odds in your favor. Even if they seem like assholes. Even if they seem unmotivated. Even if they lack moxie. Victor Cruz, Miles Austin, Marques Colston, now those are guys who you can rightfully justify searching for. While people might treat them as if they were the same sort of underdogs as Wes Welker, they really aren't. The numbers were there. The measurables and statistics existed. Teams just chose not to pay attention. Maybe they lacked grit?
Sure, being a hard worker or a guy with great character is a good thing. Do you think you can identify this trait? Was Randy Moss a hard worker? Or what about Terrell Owen's character? "Bah", you say,"They put up great stats, but I wouldn't want those cancerous guys on my team." Sure, sure. I get it. It's a team sport. It takes all 53 guys to win not just a couple guys who put up pretty stats. Okay then, so how are we going to identify these guys with character, that everyone seems to like so much? Do you have a Moxie-Meter hidden in your basement? A Lunch-Pail-O-Graph device that we can attach to their skull with electrodes, to measure this essential gritty determination? Even a Balboa-gauge would do in a pinch, if you have one on hand.
I suspect that no matter what team you root for, you have been in this position at least once. Your team has selected a wide receiver with a fairly high draft pick, only to find out that the guy just really can't catch the ball very well (Troy Williamson, Travis Taylor, etc.). So, somehow we are supposed to believe that NFL GMs can get an accurate read on a player's psyche, level of motivation, commitment, or whatever, when they can't even tell if a guy can catch. (which would seem like a fairly obvious trait to pay attention to, but perhaps gets overlooked while they are busy reading their DSM to analyze the player's psychological profile). Sure, I'm certain that most of these team's scouts, who were generally just failed former football players themselves, have keen analytical minds, highly trained and attuned to detect the faintest whiff of moxie in the air. Hell, that is probably why they stopped playing football in the first place. You don't want to damage a mind like that, with blows to the head.
Now, lets say we were working at NASA, and were looking to hire a new guy. We've got applications from all the best and brightest, from the finest universities in the land. If I turned to you and said, "You know what Bob? This guy here is the one we want. Sure, on paper he doesn't look so great. In fact he appears to be brain damaged. But you see, Bob, he just doesn't test well. I've talked to this guy and he's a real crackerjack. Sharp as a tack, I tell ya'. He's the guy we want to plan our mission to Mars." Well, best case scenario you would be laughed at. Worst case? Well, I suppose that would involve having the space shuttle go off course, and instead of heading to Mars, crashing into the White House (maybe I can pitch this idea to Michael Bay).
Yes, test results should matter to you. I know, we all hate to feel as if we are confined or trapped by some measurement, whether it is I.Q., 40 yard dash time, or whatever. We like to look in the mirror and see past that receding hairline, growing paunch, and strange rash in our armpit that just won't go away, to see the vigorous and brilliant man we know that we really are. Women just can't resist us. Oh, they play coy, but they notice us. As for other men, they tremble before our intimidating masculinity. No stupid number is going to tell us what we are, and it shouldn't do that to any man. This is America for god's sake, and no commie numbers are going to change that. USA! USA! USA!
So, basically, what I am trying to say is we tend to be morons. We're not all beautiful little snowflakes. Potatoes are all unique too, though nobody ever seems to compare themselves to one. We're not all just a step away from having our greatness discovered. We're not all going to be that outlier, who is going to defy expectations, but we still root for it to happen to someone. Somewhere inside us, we know that some guys are just better, more talented, but we don't have to like it.
Then we see a receiver run a 4.7 second forty yard dash. We pause. We think about all the great things that his coaches said about him. How his teammates would follow him through the gates of hell. He was the star of the Rose Bowl for Christ's sake! It's those damned commie numbers, coming to get us again! Anquan, he'll save us! Anquan Boldin ran a 4.72 forty yard dash, and things turned out great for him. He was gritty and tough wasn't he? That's all you need, man. Just give me some gritty tough guy, and you can keep those commie numbers of yours. Hell, look at Wes Welker. That sumbitch was 5'8" and running a 4.65 forty. Takes a lot of grit to be small and slow, doesn't it?
Sure those guys are great. They are anomalies. Outliers. Basically, they are like a cancerous tumor spreading through an otherwise healthy statistical analysis. They show up in your spreadsheet. They scare you a bit. You can't figure out what to do with them, or how to make them go away. So you learn to just accept them for what they are, a pain in your statistical ass (yes, statistical ass cancer). Though it is reasonable to find them oddly fascinating, it doesn't mean you should spend your life searching for them. That would inevitably prove rather depressing, sort of like a daily colonoscopy.
People might say that Wes Welker was underrated, or that he should have been drafted higher (actually he wasn't drafted at all). How highly should you draft a small slow wide receiver, with good but not exactly shocking college production (especially when they come from Texas Tech where everybody accumulates decent stats)? Maybe good old Wes was just hungover at the combine? Maybe Anquan didn't have his bran muffin that morning, allowing him to drop the ballast needed to increase his vertical jump? Okay, fine.
If you want to go looking for the next Wes Welker or Anquan Boldin, that's cool. I admire the quest that you are setting for yourself. I do think there are probably some guys out there who have a good eye for drafting talent, and maybe they will spot the next big thing. Still, at least with receivers (though this applies everywhere else too), most teams haven't shown that they are employing these savants in their scouting department. So, you can try to find the next guy who is going to defy all the odds, and over the next ten years you might find one. If you do manage to turn one up, I'd be the first to congratulate you. But can you do it again? If some scout out there can do that, say 60-70% of the time, then I would be very intrigued. If a team can't do this, then that underdog draft pick starts to look like a fluke, just blind luck. Or perhaps, they did have a brief moment of insight, where their internal Moxie-Meter went off, only to have it again go on the fritz for the following decade. I certainly can't rule that out.
Or, you can embrace what seems to be the most sensible solution, that those other guys, the bigger faster stronger ones who statistically pummeled their opponents all through college, might just be a better bet. It's just putting the odds in your favor. Even if they seem like assholes. Even if they seem unmotivated. Even if they lack moxie. Victor Cruz, Miles Austin, Marques Colston, now those are guys who you can rightfully justify searching for. While people might treat them as if they were the same sort of underdogs as Wes Welker, they really aren't. The numbers were there. The measurables and statistics existed. Teams just chose not to pay attention. Maybe they lacked grit?
Thursday, May 2, 2013
A statistical approach to drafting wide receivers
I thought I would post up a rather simplified example of what would happen if you asked a spreadsheet to analyze wide receiver prospects in the NFL Draft. This isn't exactly how I go about doing things, but is close enough to give some idea as to how it can improve a team's odds of making a decent pick. For this brief look, I will show what the computer thought of the draft classes from 2004 to 2012.
Player's will be given a score based in equal parts on their combine data, and their production in college. Their Stat Score makes a simple adjustment to their raw college stats to normalize things for the purpose of comparing guys who played in different offenses. The Stat Score is fairly similar to Shawn Siegele's Dominator Rating, though I use their college team's total offense rather than just their passing offense. I'm not claiming that this is an improvement over the Dominator Rating, it probably isn't, it is just the way I have historically done things. The other half of their score will be the Athletic Score, which will come from their combine results.
Since there is obviously going to be a difference between how a guy like DeSean Jackson succeeds, compared to a player like Calvin Johnson, players will be divided into two groups. The "Big" group will consist of player's over 200 pounds, and their Athletic Score will put more of an emphasis on their Kangaroo Score, though the other combine drills will still be a factor. These player's are expected to succeed by physically overpowering their opponent. The "Small" group will consist of player's under 210 pounds, and their Athletic Score will put more of an emphasis their agility score (based on the short shuttle and 3-cone drill), as well as raw speed. Basically, if you can't overpower your opponent, you want to be able to evade them. Player's who are between 200 and 210 pounds will be graded on both scales to see where they fit best.
One issue that occurs, is you sometimes have a player who so thoroughly dominates either his Athletic Score or his Stat Score, that it can bury a failure somewhere else. Take Troy Edwards, the 13th pick of the 1999 draft as an example. His Stat Score was 2.160 standard deviations above average, which is shockingly good. Unfortunately his Athletic Score was -0.630 standard deviations below average. Since I am looking for people who have a physical advantage over their opponent, and a history of meeting this potential, the computer will cut any prospect from consideration who wasn't at least average in both areas.
There is an additional issue. I'm still not sure whether to adjust the Stat Score for players who competed at the Division II or III level. Making such an adjustment would be easy, but I just haven't decided yet how much of a deduction to make.
So, here are the top 5 results from each year, including their final score, where they were selected, and whether they were graded as a Big or Small receiver:
2012
Derek Carrier Undrafted 1.606 Big
Michael Floyd pick #13 0.717 Big
Marvin McNutt pick #194 0.660 Big
Justin Blackmon pick #5 0.557 Big
Player's will be given a score based in equal parts on their combine data, and their production in college. Their Stat Score makes a simple adjustment to their raw college stats to normalize things for the purpose of comparing guys who played in different offenses. The Stat Score is fairly similar to Shawn Siegele's Dominator Rating, though I use their college team's total offense rather than just their passing offense. I'm not claiming that this is an improvement over the Dominator Rating, it probably isn't, it is just the way I have historically done things. The other half of their score will be the Athletic Score, which will come from their combine results.
Since there is obviously going to be a difference between how a guy like DeSean Jackson succeeds, compared to a player like Calvin Johnson, players will be divided into two groups. The "Big" group will consist of player's over 200 pounds, and their Athletic Score will put more of an emphasis on their Kangaroo Score, though the other combine drills will still be a factor. These player's are expected to succeed by physically overpowering their opponent. The "Small" group will consist of player's under 210 pounds, and their Athletic Score will put more of an emphasis their agility score (based on the short shuttle and 3-cone drill), as well as raw speed. Basically, if you can't overpower your opponent, you want to be able to evade them. Player's who are between 200 and 210 pounds will be graded on both scales to see where they fit best.
One issue that occurs, is you sometimes have a player who so thoroughly dominates either his Athletic Score or his Stat Score, that it can bury a failure somewhere else. Take Troy Edwards, the 13th pick of the 1999 draft as an example. His Stat Score was 2.160 standard deviations above average, which is shockingly good. Unfortunately his Athletic Score was -0.630 standard deviations below average. Since I am looking for people who have a physical advantage over their opponent, and a history of meeting this potential, the computer will cut any prospect from consideration who wasn't at least average in both areas.
There is an additional issue. I'm still not sure whether to adjust the Stat Score for players who competed at the Division II or III level. Making such an adjustment would be easy, but I just haven't decided yet how much of a deduction to make.
So, here are the top 5 results from each year, including their final score, where they were selected, and whether they were graded as a Big or Small receiver:
2012
Derek Carrier Undrafted 1.606 Big
Michael Floyd pick #13 0.717 Big
Marvin McNutt pick #194 0.660 Big
Justin Blackmon pick #5 0.557 Big
Alshon Jeffrey pick #45
0.483 Big
Derek Carrier is a bit of an oddball, coming from tiny Beloit College. It will be interesting to see if he becomes anything , or if I will have to start creating a penalty for players from such low levels of competition.
2011
Jonathan Baldwin
pick #26 1.098 Big Next stop, Bustville!
Julio Jones pick #6 0.814 Big
Torrey Smith pick #58 0.643 Small
Stephen Burton
pick #236 0.513 Big
Cecil Shorts pick #114
0.484 Small
2010
Dez Bryant pick #24 0.998
Big
Mike Williams pick #101
0.690 Big
Victor Cruz
Undrafted 0.685 Big or 0.590 Small(listing both because it’s
interesting to me)
Andre Roberts pick #88
0.640 Small
Emmanuel Sanders pick #82
0.588 Small
There was no combine data for Eric Decker, Demaryius Thomas, or
Danario Alexander. If there had been I suspect they would have made the list.
2009
Kenny Britt pick #30 0.807
Big
Hakeem Nicks pick #29
0.720 Big
Ramses Barden pick #85 0.705 Big
Mike Thomas pick #107
0.503 Small
Mike Wallace pick #84 0.315 Small
No combine for Michael Crabtree
2008
James Hardy pick #41
0.413 Big Ooops! That didn't turn out well.
Jordy Nelson pick #36 0.411
Big
Pierre Garcon pick #205 0.388
Big
Donnie Avery pick #33
0.363 Small
Earl Bennett pick #70
0.169 Small
This was just a terrible year for receivers.
2007
Calvin Johnson
pick #2 1.881 Big
Robert Meachem
pick #27 0.661 Big One of the more disappointing players.
Mike Sims-Walker
pick #79 0.490 Big
Dwayne Bowe pick #23 0.435 Big
Dwayne Bowe pick #23 0.435 Big
Laurent Robinson
pick #75 0.381 Big
2006
Miles Austin Undrafted
0.851 Big
Greg Jennings pick #52
0.750 Small
Brandon Marshall
pick #119 0.640 Big
Marques Colston
pick #252 0.608 Big
Derek Hagan pick #82 0.599
Big
A banner year for the computer. Fortunately the computer is a humble guy, and
doesn’t make much fuss about it.
2005
Vincent Jackson pick
#61 2.019 Big
Dante Ridgeway pick #192
0.990 Big Not one of the computer’s finer moments
Mike Williams pick #10
0.988 (now known as “the fat Mike Williams”) Big
Roddy White pick #27 0.815
Small
Braylon Edwards pick #3 0.658
Big
That’s not a typo. Vincent Jackson is indeed 2 standard
deviations above average.
2004
Larry Fitzgerald pick #3
1.014 Big
Lee Evans pick #13 0.855
Small
Reggie Williams pick #9 0.797
Big
Rashaun Woods pick #30 0.605
Big Honestly, I'm not sure what happened with this guy.
Jerricho Cotchery pick #108 0.588
Big
As I said before, this is just a very basic way of doing things, and not what I would really recommend. Trying to boil things down to one overall score just doesn't work as well as looking over a broader set of smaller scores. Using a broader set of data lets you get more of a sense as to how balanced a player is in a wide range of areas.
Still, while the computer does make some mistakes here (Jon Baldwin, Mike "The fat one" Williams, James Hardy, etc.) many of these mistakes are no worse than what actual NFL teams did. The spreadsheet also scored major coups in selecting players like Victor Cruz, Marques Colston, Brandon Marshall, Miles Austin, and Mike "Not the fat one" Williams, as well as others, all in the later rounds or undrafted. Overall, out of the 40 listed prospects (not counting the 2012 draft class, since it is too early for that), 67.5% arguably became successes according to my odd definition of the term. That is in comparison to a overall average success rate of 24.39% for all drafted receivers, or a 22.5% median league-wide success rate for individual NFL teams. Some players like Emmanuel Sanders also appear poised to enter the "success" list as their playing time increases, but I'll leave that alone for now.
I'm not saying that players should be graded in such a simple manner. I'm just saying that even a method this ridiculously simple should outperform most NFL GMs, and that a more sophisticated version, along with some limited film study, should produce excellent results. I'll get into exploring some of the ways to refine things even further here.
Beating Ozzie Newsome: 2013 Edition
For the past few years I have been playing a little game to make watching the NFL Draft more entertaining. When my home team, the Ravens, are on the clock, I make my own pick as to who we should draft. If the Ravens trade up or down in the draft, I can choose to go along with this decision, or stay put. Obviously, I can't initiate trades on my own. It's just fun to see how my picks turn out in comparison to the ones made by the pros.
The Ravens, unfortunately, always have a couple advantages over me. They can choose to give playing time and opportunities to people who just aren't very good, and neglect other guys who have potential. No matter what team you root for, I am sure there are some players that make you wonder why the team keeps giving them chances (Terrence Cody!). A lot of the people I pick in the later rounds don't even end up being drafted, though many of them eventually do make a team's roster. Those are the risks, and there isn't much I can do about it. There's also the possibility that a player I choose will be forced to play outside of what I consider to be their ideal position, say perhaps a 3-4 outside linebacker forced to play defensive end. Again, there's not much I can do about this. In the end, I am just betting on superior athletes, with a solid history of production in college, overcoming these sorts of obstacles.
So, now that I have made my excuses in advance, for what has a rather high probability of being a future source of embarrassment, I'll get to the point and unveil who I would have selected this year.
Team Kangaroo Team Ozzie
Round 1
pick #32 Jamar Taylor, CB Matt Elam, SS
Round 2
pick #56 --(I don't trade up)--- Arthur Brown, ILB
pick #62 Terron Armstead, OT --(lost in trade)--------
Round 3
pick #94 Da'Rick Rogers, WR Brandon Williams, DT
this could cause me trouble
Round 4
pick #129 Dave Quessenberry, OT John Simon, DE/OLB
pick #130 Ryan Swope, WR Kyle Juszczyk, FB
Round 5
pick #165 Eric Kush, C --(lost in trade)------
pick #168 Chris Jones, DT Ricky Wagner, OT/OG
Round 6
pick #199 Charles Johnson, WR --(lost in trade)-------
pick #200 Eric Martin, LB Kapron Lewis Moore, DT
pick#203 Glenn Foster, DT Ryan Jensen, OG/C
Round 7
pick #238 Paul Worrilow, LB Aaron Mellette, WR
pick #247 Wade Jacobson, OT Marc Anthony, CB
Before the mockery begins, I'll point out some of the weird issues with my draft class. Yes, five of my draftees never got drafted at all, including my third round pick Da'Rick Rogers. This is actually fairly typical, except for Da'Rick, and I expect a number of these oddballs to still make teams. The difference between a 6th round pick and an undrafted free agent is rather insignificant, and I believe in my guys (don't fuck up guys!).
A bigger question might be why I drafted 3 offensive tackles and 3 wide receivers. This probably seems like overkill, and somewhat impractical. Well, the Ravens tackles kind of suck, but fortunately this lack of talent is accompanied by no depth. So, they are bad, but fortunately there are only a few crappy guys to yell at rather than a lot of them. The situation will be getting even worse after this year when the admirably mediocre Michael Oher will be a free agent. So, I figured that as silly as it may seem the team could actually handle drafting three, and basically thought the value was there at the point they were picked at. Quessenberry could also be viewed as a guard, if that makes things seem a bit more sane.
The situation is pretty much the same at the wide receiver position, where after Torrey Smith things get very sketchy very fast. Even Torrey, who I like, kind of worries me. I'm just not convinced he is suited to be a number one receiver. I felt the value was there in the draft (except for Da'Rick Rogers evidently), so I attempted to rebuild the entire wide receiver group all at once. Yes, this would mean cutting pretty much all of the other receivers that are currently on the team. I'd be curious to hear why that would be a bad thing. Oh, fine. We can keep Jacoby Jones around to be our kick returner/dance instructor, but everyone else GONE!
Honestly, I feel that Ozzie had a better draft than he has had in years. This isn't as powerful a compliment as it may initially appear to be. I'm not sure if there will be any stars in it, but I suspect a few of his picks are pretty decent. My favorite pick is Brandon Williams, whom I considered taking myself. I see no way that he doesn't immediately take the nose tackle job away from Terrence Cody. Matt Elam might be a solid guy, though I don't like throwing out hunches on defensive backs without short shuttle and 3-cone times. I think Arthur Brown is an overrated mediocrity, but he will be surrounded by players that will probably make him look better than he is. People keep claiming he is some sort of speedy sideline to sideline monster, though I'm not sure where they are getting this crazy idea.
Time will tell how it all works out. Maybe I'll be right, and you'll become more skeptical of the wisdom of NFL GMs. Maybe Ozzie will be right, and you will become more skeptical of me (if that's possible). Either way, you manage to identify one idiot that you shouldn't be listening to, so it's kind of a win-win for you.
*Update 7/25/13- Now my 4th round pick, Ryan Swope, has retired before even starting a game, due to issues relating to concussions. This could be a rough year for Team Kangaroo.
The Ravens, unfortunately, always have a couple advantages over me. They can choose to give playing time and opportunities to people who just aren't very good, and neglect other guys who have potential. No matter what team you root for, I am sure there are some players that make you wonder why the team keeps giving them chances (Terrence Cody!). A lot of the people I pick in the later rounds don't even end up being drafted, though many of them eventually do make a team's roster. Those are the risks, and there isn't much I can do about it. There's also the possibility that a player I choose will be forced to play outside of what I consider to be their ideal position, say perhaps a 3-4 outside linebacker forced to play defensive end. Again, there's not much I can do about this. In the end, I am just betting on superior athletes, with a solid history of production in college, overcoming these sorts of obstacles.
So, now that I have made my excuses in advance, for what has a rather high probability of being a future source of embarrassment, I'll get to the point and unveil who I would have selected this year.
Team Kangaroo Team Ozzie
Round 1
pick #32 Jamar Taylor, CB Matt Elam, SS
Round 2
pick #56 --(I don't trade up)--- Arthur Brown, ILB
pick #62 Terron Armstead, OT --(lost in trade)--------
Round 3
pick #94 Da'Rick Rogers, WR Brandon Williams, DT
this could cause me trouble
Round 4
pick #129 Dave Quessenberry, OT John Simon, DE/OLB
pick #130 Ryan Swope, WR Kyle Juszczyk, FB
Round 5
pick #165 Eric Kush, C --(lost in trade)------
pick #168 Chris Jones, DT Ricky Wagner, OT/OG
Round 6
pick #199 Charles Johnson, WR --(lost in trade)-------
pick #200 Eric Martin, LB Kapron Lewis Moore, DT
pick#203 Glenn Foster, DT Ryan Jensen, OG/C
Round 7
pick #238 Paul Worrilow, LB Aaron Mellette, WR
pick #247 Wade Jacobson, OT Marc Anthony, CB
Before the mockery begins, I'll point out some of the weird issues with my draft class. Yes, five of my draftees never got drafted at all, including my third round pick Da'Rick Rogers. This is actually fairly typical, except for Da'Rick, and I expect a number of these oddballs to still make teams. The difference between a 6th round pick and an undrafted free agent is rather insignificant, and I believe in my guys (don't fuck up guys!).
A bigger question might be why I drafted 3 offensive tackles and 3 wide receivers. This probably seems like overkill, and somewhat impractical. Well, the Ravens tackles kind of suck, but fortunately this lack of talent is accompanied by no depth. So, they are bad, but fortunately there are only a few crappy guys to yell at rather than a lot of them. The situation will be getting even worse after this year when the admirably mediocre Michael Oher will be a free agent. So, I figured that as silly as it may seem the team could actually handle drafting three, and basically thought the value was there at the point they were picked at. Quessenberry could also be viewed as a guard, if that makes things seem a bit more sane.
The situation is pretty much the same at the wide receiver position, where after Torrey Smith things get very sketchy very fast. Even Torrey, who I like, kind of worries me. I'm just not convinced he is suited to be a number one receiver. I felt the value was there in the draft (except for Da'Rick Rogers evidently), so I attempted to rebuild the entire wide receiver group all at once. Yes, this would mean cutting pretty much all of the other receivers that are currently on the team. I'd be curious to hear why that would be a bad thing. Oh, fine. We can keep Jacoby Jones around to be our kick returner/dance instructor, but everyone else GONE!
Honestly, I feel that Ozzie had a better draft than he has had in years. This isn't as powerful a compliment as it may initially appear to be. I'm not sure if there will be any stars in it, but I suspect a few of his picks are pretty decent. My favorite pick is Brandon Williams, whom I considered taking myself. I see no way that he doesn't immediately take the nose tackle job away from Terrence Cody. Matt Elam might be a solid guy, though I don't like throwing out hunches on defensive backs without short shuttle and 3-cone times. I think Arthur Brown is an overrated mediocrity, but he will be surrounded by players that will probably make him look better than he is. People keep claiming he is some sort of speedy sideline to sideline monster, though I'm not sure where they are getting this crazy idea.
Time will tell how it all works out. Maybe I'll be right, and you'll become more skeptical of the wisdom of NFL GMs. Maybe Ozzie will be right, and you will become more skeptical of me (if that's possible). Either way, you manage to identify one idiot that you shouldn't be listening to, so it's kind of a win-win for you.
*Update 7/25/13- Now my 4th round pick, Ryan Swope, has retired before even starting a game, due to issues relating to concussions. This could be a rough year for Team Kangaroo.
Wednesday, May 1, 2013
Wide Receiver Success Rate
I was fiddling around with my wide receiver spreadsheets, trying to tweak my formulas, and I thought it would be interesting to see how NFL teams compare to one another when it comes to making selections at this position. This is intended to just get some vague idea as to how often teams actually manage to correctly identify quality wide receiver prospects, though it has some limitations.
First of all, this will include players selected between 2004 and 2012. Players taken in the last few years are probably still developing, and could possibly improve the historical grade I will assign to a team. For my purposes though, it really doesn't matter. I'm not trying to paint any team as being incompetent, I'm just trying to get a glimpse of the overall picture.
The next issue is coming up with a way of defining what we will call a "success". I generally set the bar at 35 yards per game played (not per game started). This would work out to a player that averages 560 yards per year. This isn't an exceptional amount of production, but is significantly more than most draft prospects will ever produce. At this level, a player is at the very least rosterable, and contributing to a reasonable degree. The reason I chose 35 yards as the cutoff point, was that it seemed to be the average result for wide receivers who were at least given significant enough playing time to prove themselves, one way or the other. I'm sure somebody will want to argue about this, and suggest a better method of defining what "average" is, so feel free to suggest something better. I'm just shooting for simplicity.
I'll explore wide receivers in more depth later, but for now here is how things worked out. Beside each team's name I will give the number of wideouts they have drafted since 2004, as well as what percentage of them managed to at least reach my magical line of mediocrity. Receivers who were undrafted, such as Victor Cruz or Miles Austin, aren't included in this examination. Teams only get credit for players they actually used a draft pick on.
Team # of Wideouts Taken Overall Success Rate
49ers 13 15.38%
Bears 9 33.33%
Bengals 15 13.33%
Bills 7 28.57%
Broncos 8 50%
Browns 10 30%
Buccaneers 10 10%
Cardinals 7 57.14%
Chargers 5 20%
Chiefs 12 8.33%
Colts 6 50%
Cowboys 6 33%
Dolphins 8 12.5%
Eagles 7 57.14%
Falcons 8 50%
Giants 8 38.57%
Jaguars 11 36.36%
Jets 8 28.57%
Lions 8 37.5%
Packers 9 44.44%
Panthers 8 12.5%
Patriots 7 0%
Raiders 13 15.38%
Rams 11 18.18%
Ravens 12 16.66%
Redskins 7 14.28%
Saints 7 28.57%
Seahawks 7 0%
Steelers 9 33.33%
Texans 8 0%
Titans 13 15.38%
Vikings 10 20%
The league-wide median success rate would be 22.5%. From 2004 to 2012, 287 wide receivers have been drafted, and 70 (24.39%) of them managed to become at least average by my definition of the word. For teams like the Broncos, Cardinals, Colts, Eagles, Falcons, and Packers, congratulations, you are probably doing something right (and having a good quarterback might be biggest factor). For everybody else, well, you might as well be blindly pulling names out of a hat.
The average NFL team has spent about 9 draft picks on the wide receiver position during this time frame. If your team has used more picks than this, while having a success rate below 24.39%, it would seem that there are two possibilities that could explain what is going wrong. They are either mind bogglingly unlucky, or they are are actively pursuing bozos. A purge of the scouting department would seem to be in order, as a computer spreadsheet of combine data and college stats would probably do a better job.
Update: 8/1/13 - You can also go to the post Wide Receiver Success Rate part 2, to see what happens when we raise the bar to 45 yards per game. In this post, I list who these receivers are, and get into what traits they all might all have in common.
First of all, this will include players selected between 2004 and 2012. Players taken in the last few years are probably still developing, and could possibly improve the historical grade I will assign to a team. For my purposes though, it really doesn't matter. I'm not trying to paint any team as being incompetent, I'm just trying to get a glimpse of the overall picture.
The next issue is coming up with a way of defining what we will call a "success". I generally set the bar at 35 yards per game played (not per game started). This would work out to a player that averages 560 yards per year. This isn't an exceptional amount of production, but is significantly more than most draft prospects will ever produce. At this level, a player is at the very least rosterable, and contributing to a reasonable degree. The reason I chose 35 yards as the cutoff point, was that it seemed to be the average result for wide receivers who were at least given significant enough playing time to prove themselves, one way or the other. I'm sure somebody will want to argue about this, and suggest a better method of defining what "average" is, so feel free to suggest something better. I'm just shooting for simplicity.
I'll explore wide receivers in more depth later, but for now here is how things worked out. Beside each team's name I will give the number of wideouts they have drafted since 2004, as well as what percentage of them managed to at least reach my magical line of mediocrity. Receivers who were undrafted, such as Victor Cruz or Miles Austin, aren't included in this examination. Teams only get credit for players they actually used a draft pick on.
Team # of Wideouts Taken Overall Success Rate
49ers 13 15.38%
Bears 9 33.33%
Bengals 15 13.33%
Bills 7 28.57%
Broncos 8 50%
Browns 10 30%
Buccaneers 10 10%
Cardinals 7 57.14%
Chargers 5 20%
Chiefs 12 8.33%
Colts 6 50%
Cowboys 6 33%
Dolphins 8 12.5%
Eagles 7 57.14%
Falcons 8 50%
Giants 8 38.57%
Jaguars 11 36.36%
Jets 8 28.57%
Lions 8 37.5%
Packers 9 44.44%
Panthers 8 12.5%
Patriots 7 0%
Raiders 13 15.38%
Rams 11 18.18%
Ravens 12 16.66%
Redskins 7 14.28%
Saints 7 28.57%
Seahawks 7 0%
Steelers 9 33.33%
Texans 8 0%
Titans 13 15.38%
Vikings 10 20%
The league-wide median success rate would be 22.5%. From 2004 to 2012, 287 wide receivers have been drafted, and 70 (24.39%) of them managed to become at least average by my definition of the word. For teams like the Broncos, Cardinals, Colts, Eagles, Falcons, and Packers, congratulations, you are probably doing something right (and having a good quarterback might be biggest factor). For everybody else, well, you might as well be blindly pulling names out of a hat.
The average NFL team has spent about 9 draft picks on the wide receiver position during this time frame. If your team has used more picks than this, while having a success rate below 24.39%, it would seem that there are two possibilities that could explain what is going wrong. They are either mind bogglingly unlucky, or they are are actively pursuing bozos. A purge of the scouting department would seem to be in order, as a computer spreadsheet of combine data and college stats would probably do a better job.
Update: 8/1/13 - You can also go to the post Wide Receiver Success Rate part 2, to see what happens when we raise the bar to 45 yards per game. In this post, I list who these receivers are, and get into what traits they all might all have in common.
Eric Kush
Eric Kush was a prospect in the 2013 Draft that I found very interesting. Unfortunately, this selection seems to have not been a big hit among some of the fans of the Chiefs, the team that drafted him. From what I can gather, people seem to be wondering why the Chiefs drafted this center from the University of California (PA), when many draft websites weren't even listing him as a draftable prospect. So, let me see if I can make people feel more upbeat about why this could have been an excellent selection, despite what ESPN may be telling you..
Identifying offensive line prospects is a little peculiar for me, as I have to rely almost entirely on data from the NFL Combine to come up with my deranged conclusions. I do prefer drafting positions that accumulate some sort of stats, to sort of see whether a prospect was performing to a level that matches their physical ability. As a final step, it is also nice to watch the player in action to see if there is an "Oh shit, did you see that?" moment to visually confirm what all the numbers are suggesting. Unfortunately this last step is mind numbingly dull when scouting centers, though I suppose there are people who are into that sort of thing.
Despite these issues, a computer can do a rather good job of identifying offensive line talent based on just the player's combine/pro day data. The prospect of making decisions based on something like this seems to revolt most people, but I think they fail to recognize just how horrendously most teams do in the draft, so beating NFL GMs really isn't that hard. Based on this limited information you should still pick a successful player about 50-70% of the time, regardless of what round they are taken in. People who swear by the idea of watching game tape are undoubtedly going to laugh at this, but that same film study approach generally only produces about a 21.5% success rate for NFL teams. By betting on a player's combine numbers and college stats to predict their success, we're basically counting cards. As you see with counting cards, you have to stick to the system, avoiding gut reactions or subjective opinions, and just play the odds.
One of the interesting things about offensive linemen is the odd story their combine numbers tell. To a rather large degree you can tell what position a player is destined for, or would thrive in, just by looking at his combine data. Centers in particular have a weird anomaly that really makes them stand out, and that is their short shuttle times. While I would expect positive times in the agility drills to be beneficial for offensive linemen, I have no real theory at this time as to what possible explanation there could be for this short shuttle phenomenon. There does seem to be some sort of relationship between the short shuttle and lower body flexibility, and change of direction skills. What this could mean for centers, I'm still not sure of though. So, for now, I have just decided to go with it, since it seems to work.
Listed below are pretty much all of the centers from the last ten years who have made the Pro Bowl, and the All Pro Team. I am including their Kangaroo score as well as their Short Shuttle Score, and 3-Cone Drill score. As I said, it is the short shuttle score which matters the most here, the other scores can be seen as bonuses if a player does well in them.. All scores are given in relation to what the average results were for offensive linemen. Basically, any score that is one standard deviation above average is going to put you at somewhere above the 70th percentile for your position.
Player Short Shuttle Score Kangaroo Score 3-Cone Score
Max Unger 1.208 -1.493 1.396 1 PB
Chris Myers 1.943 0.213 1.511 2 PB
Nick Mangold 1.698 -0.584 1.051 4 PB
Scott Wells 1.698 0.278 0.767 1 PB
Ryan Kalil 1.992 -0.840 0.980 3 PB
Maurkice Pouncey -0.850 -1.149 0.130 3 PB
Alex Mack -0.017 -0.026 1.652 1 PB Alt.
Dan Koppen 0.914 -0.762 -1.709 1 PB
Nick Hardwick 0.473 0.851 -0.222 1 PB
LeCharles Bentley 1.012 -0.983 -0.258 2 PB
John Sullivan 0.963 0.251 0.555 1 All Pro
Now, I realize that correlation doesn't equal causation, but I just can't help but find this to be interesting. Out of these 11 individuals who have achieved some fame/success at the center position, 8 had ridiculously high scores on the short shuttle. Scores above one standard deviation are rare enough that there may only be one or two people in a given draft that can hit this benchmark, and a lot of these guys are significantly exceeding it. Of the remaining three players, Nick Hardwick was still significantly above average at 0.473 but had the highest Kangaroo Score at 0.851, which suggests he has better lower body explosiveness than his peers. Alex Mack basically came out average across the board except for his 3-cone score, but he was also the only one on the list who only made the Pro Bowl as an alternate. The real mystery is Maurkice Pouncey, who scored terribly on nearly everything. I have no explanation for him, except to say that outliers do exist, although it is also possible that he is simply overrated and perhaps not quite as good as his reputation leads people to believe. I really can't say. However you want to look at it, the statistical chances of this many guys making the Pro Bowl, with these sorts of exceptionally rare short shuttle times, would be incredibly slim.
Now, obviously, using Pro Bowl selections as a benchmark for greatness is a questionable proposition. The problem is that there really is no truly objective method for saying whether a player is particularly good. Some people prefer to use Career Approximate Value as a measurement, but this has its own flaws. It tends to work better at making generalizations about groups of players, and then falls apart when looking at individual players. I prefer some combination of CarAV, plus stats, plus reading criticisms of a player to get a sense of him, but this is fairly time consuming. This approach is also difficult to apply to a large group of players. At some point I think it is probably reasonable to allow some subjective judgments to get in there, and the Pro Bowl, as stupid as it is, at least gives some general indication that a guy was probably doing something right.
So, how does all of this relate to Eric Kush? Well, here are Eric's numbers:
Player Short Shuttle Score Kangaroo Score 3-Cone Score
Eric Kush 1.943 0.221 1.334
Among the three primary things I look for in centers, he did exceptionally well in all of them. His Kangaroo Score might seem like a weak result, but is still above average for his position. The bell curve for the Kangaroo Score is somewhat distorted by the fact that offensive tackles are graded with centers and guards, and the best tackles tend to dominate the Kangaroo Score, making it somewhat unusual for centers and guards to even reach the average level. Beyond this issue, Eric performed admirably with 25 reps on the bench press, and had a respectable 40 yard dash of 5.04 (and a 1.73 ten yard split), both of which have some added but lesser correlations with success. All in all, he is pretty much the prototype of what teams are looking for athletically from a center, and easily the most appealing center prospect in the 2013 draft. Though opportunity, injury, and off the field issues can all affect a player's chances of success, I would probably say the odds are in his favor of making a Pro Bowl at some point in the future, considering the rarity of players with his traits, and the frequency with which they end up becoming exceptional. That would be quite a nice payoff for a team that only invested a sixth round pick. I rather strongly doubt that any of the four centers taken ahead of him will turn out nearly as well.
None of this should be taken to mean that a player can't succeed with a lower short shuttle time, or that a player with a good short shuttle time is guaranteed success. I'm merely suggesting that there is a rather strong history of things working out well for centers who do well in this drill. We'll see what happens. In the meantime, it's certainly more entertaining to place your bets on a 6th round guy, that some people might be overlooking.
***Update- In the brief time since I wrote this, I wrote something else that I think better describes the subject of NFL centers. If you are interested you can go to the post on Short Shuttle Times and NFL Centers.
Identifying offensive line prospects is a little peculiar for me, as I have to rely almost entirely on data from the NFL Combine to come up with my deranged conclusions. I do prefer drafting positions that accumulate some sort of stats, to sort of see whether a prospect was performing to a level that matches their physical ability. As a final step, it is also nice to watch the player in action to see if there is an "Oh shit, did you see that?" moment to visually confirm what all the numbers are suggesting. Unfortunately this last step is mind numbingly dull when scouting centers, though I suppose there are people who are into that sort of thing.
Despite these issues, a computer can do a rather good job of identifying offensive line talent based on just the player's combine/pro day data. The prospect of making decisions based on something like this seems to revolt most people, but I think they fail to recognize just how horrendously most teams do in the draft, so beating NFL GMs really isn't that hard. Based on this limited information you should still pick a successful player about 50-70% of the time, regardless of what round they are taken in. People who swear by the idea of watching game tape are undoubtedly going to laugh at this, but that same film study approach generally only produces about a 21.5% success rate for NFL teams. By betting on a player's combine numbers and college stats to predict their success, we're basically counting cards. As you see with counting cards, you have to stick to the system, avoiding gut reactions or subjective opinions, and just play the odds.
One of the interesting things about offensive linemen is the odd story their combine numbers tell. To a rather large degree you can tell what position a player is destined for, or would thrive in, just by looking at his combine data. Centers in particular have a weird anomaly that really makes them stand out, and that is their short shuttle times. While I would expect positive times in the agility drills to be beneficial for offensive linemen, I have no real theory at this time as to what possible explanation there could be for this short shuttle phenomenon. There does seem to be some sort of relationship between the short shuttle and lower body flexibility, and change of direction skills. What this could mean for centers, I'm still not sure of though. So, for now, I have just decided to go with it, since it seems to work.
Listed below are pretty much all of the centers from the last ten years who have made the Pro Bowl, and the All Pro Team. I am including their Kangaroo score as well as their Short Shuttle Score, and 3-Cone Drill score. As I said, it is the short shuttle score which matters the most here, the other scores can be seen as bonuses if a player does well in them.. All scores are given in relation to what the average results were for offensive linemen. Basically, any score that is one standard deviation above average is going to put you at somewhere above the 70th percentile for your position.
Player Short Shuttle Score Kangaroo Score 3-Cone Score
Max Unger 1.208 -1.493 1.396 1 PB
Chris Myers 1.943 0.213 1.511 2 PB
Nick Mangold 1.698 -0.584 1.051 4 PB
Scott Wells 1.698 0.278 0.767 1 PB
Ryan Kalil 1.992 -0.840 0.980 3 PB
Maurkice Pouncey -0.850 -1.149 0.130 3 PB
Alex Mack -0.017 -0.026 1.652 1 PB Alt.
Dan Koppen 0.914 -0.762 -1.709 1 PB
Nick Hardwick 0.473 0.851 -0.222 1 PB
LeCharles Bentley 1.012 -0.983 -0.258 2 PB
John Sullivan 0.963 0.251 0.555 1 All Pro
Now, I realize that correlation doesn't equal causation, but I just can't help but find this to be interesting. Out of these 11 individuals who have achieved some fame/success at the center position, 8 had ridiculously high scores on the short shuttle. Scores above one standard deviation are rare enough that there may only be one or two people in a given draft that can hit this benchmark, and a lot of these guys are significantly exceeding it. Of the remaining three players, Nick Hardwick was still significantly above average at 0.473 but had the highest Kangaroo Score at 0.851, which suggests he has better lower body explosiveness than his peers. Alex Mack basically came out average across the board except for his 3-cone score, but he was also the only one on the list who only made the Pro Bowl as an alternate. The real mystery is Maurkice Pouncey, who scored terribly on nearly everything. I have no explanation for him, except to say that outliers do exist, although it is also possible that he is simply overrated and perhaps not quite as good as his reputation leads people to believe. I really can't say. However you want to look at it, the statistical chances of this many guys making the Pro Bowl, with these sorts of exceptionally rare short shuttle times, would be incredibly slim.
Now, obviously, using Pro Bowl selections as a benchmark for greatness is a questionable proposition. The problem is that there really is no truly objective method for saying whether a player is particularly good. Some people prefer to use Career Approximate Value as a measurement, but this has its own flaws. It tends to work better at making generalizations about groups of players, and then falls apart when looking at individual players. I prefer some combination of CarAV, plus stats, plus reading criticisms of a player to get a sense of him, but this is fairly time consuming. This approach is also difficult to apply to a large group of players. At some point I think it is probably reasonable to allow some subjective judgments to get in there, and the Pro Bowl, as stupid as it is, at least gives some general indication that a guy was probably doing something right.
So, how does all of this relate to Eric Kush? Well, here are Eric's numbers:
Player Short Shuttle Score Kangaroo Score 3-Cone Score
Eric Kush 1.943 0.221 1.334
Among the three primary things I look for in centers, he did exceptionally well in all of them. His Kangaroo Score might seem like a weak result, but is still above average for his position. The bell curve for the Kangaroo Score is somewhat distorted by the fact that offensive tackles are graded with centers and guards, and the best tackles tend to dominate the Kangaroo Score, making it somewhat unusual for centers and guards to even reach the average level. Beyond this issue, Eric performed admirably with 25 reps on the bench press, and had a respectable 40 yard dash of 5.04 (and a 1.73 ten yard split), both of which have some added but lesser correlations with success. All in all, he is pretty much the prototype of what teams are looking for athletically from a center, and easily the most appealing center prospect in the 2013 draft. Though opportunity, injury, and off the field issues can all affect a player's chances of success, I would probably say the odds are in his favor of making a Pro Bowl at some point in the future, considering the rarity of players with his traits, and the frequency with which they end up becoming exceptional. That would be quite a nice payoff for a team that only invested a sixth round pick. I rather strongly doubt that any of the four centers taken ahead of him will turn out nearly as well.
None of this should be taken to mean that a player can't succeed with a lower short shuttle time, or that a player with a good short shuttle time is guaranteed success. I'm merely suggesting that there is a rather strong history of things working out well for centers who do well in this drill. We'll see what happens. In the meantime, it's certainly more entertaining to place your bets on a 6th round guy, that some people might be overlooking.
***Update- In the brief time since I wrote this, I wrote something else that I think better describes the subject of NFL centers. If you are interested you can go to the post on Short Shuttle Times and NFL Centers.
Subscribe to:
Posts (Atom)