Monday, September 2, 2013

Short Shuttle Times and NFL Centers

As the name of this blog would suggest, I believe that the simplest answers to a problem tend to be the best ones.  Among these simple (and potentially stupid) answers that I believe in, is the possibility that if you are looking for a center for your team's offensive line, you might not need to go any further than examining the player's short shuttle time.  Strange and magical things seem to happen when a player's short shuttle time approaches the 4.50 second mark. 

Yes, I realize that this is probably going to sound incredibly foolish.  Yes, I realize that this doesn't involve film study.  Yes, character, hard work, motivation, proper blocking technique, blah, blah, blah...that stuff matters I guess (Does it?  I'm really not sure.).  I will admit that this seems a bit crazy, and yes, I do consider other factors.  I just suspect that, in the end, the short shuttle time probably is the main deciding factor in a player's success at the center position.  So, let's proceed with my attempt at demonstrating this lunacy, and you can laugh at me afterwards.

What I'm going to do, is to list every single player who was the intended starting center at the beginning of the 2012 season.  While some of these players ended up getting hurt, and not finishing the season, I think it is best to just include the players that the team wanted to start, not the ones they were forced to play.  I will make one exception to this rule though, and that is for the Green Bay Packers.  I'll explain why I'm making that exception in a little bit.  For the most part, I believe this list is accurate, though in a couple cases it was a bit unclear who the intended center was.  Overall, this should be pretty reliable though.

The 32 starting centers will be divided into 3 tiers.  Along with their short shuttle time, I will also show their Kangaroo Score, and their Agility Score (half of which also comes from the short shuttle drill), which are shown in the form of how many standard deviations they are away from the average results for a player in their position group.  I'll also show how many games they have started in their career (GS), their number of Pro Bowl selections (PB), and their number of All Pro selections.  I should also mention that the average short shuttle time for all offensive linemen is approximately 4.74 seconds.

Group 1
The Average and the Unknown- Starting centers with short shuttle times of 4.60 seconds or worse.
Player     Year   Pick#   Sh. Sht.  Kangaroo     Agility        GS    PB All Pro
Rodney Hudson 2011 55 4.96 -1.545 -0.971 4

Maurkice Pouncey 2010 18 4.92 -1.149 -0.359 45 3 3
Brian De La Puente 2008          U 4.77 -1.915 0.220 28

Alex Mack 2009 21 4.75 -0.026 0.817 64 1
Ted Larsen 2010 205 4.66 -1.226 0.153 27

Phil Costa 2010          U 4.65 1.420 0.443 20

Nick Hardwick 2004 66 4.65 0.851 0.125 119 1
Mike Pouncey 2011 15 4.64 -1.409 0.468 32

Stefen Wisniewski 2011 48 4.64 -0.235 0.733 31

Ryan Wendell 2008          U         N/A            N/A          N/A 21

Fernando Velasco 2008          U         N/A            N/A          N/A 19

Peter Konz 2012 55         N/A            N/A          N/A 15

Jonathan Goodwin 2002 154         N/A            N/A          N/A 92 1
Lyle Sendlein 2007          U         N/A            N/A          N/A 77

-0.581 0.181

Players in Group 1 were drafted, on average, around the 104th pick, which would be in the 3rd round.  The five players at the bottom of this group, for whom I have no data, aren't factored into any of this.  I'll explain why I dumped these unknowns into this group later.  Group 1 is sort of a hodgepodge of highly drafted players, and people that nobody has probably ever heard of before.  While numerous players in this group have short shuttle times that are better than the average result of 4.74 seconds, their results still aren't exactly shocking.  I'm not saying that such players will do poorly.  I am only suggesting that I generally wouldn't be inclined to bet on them, if I had other options.

This is only going to cause me trouble.  You see that guy in the list above?  The one with the most Pro Bowls, 3?  Yes, Maurkice Pouncey!  Well, he might be overrated (please don't hurt me).  Really, I have no axe to grind here, but if you take the fellows at Pro Football Focus seriously (a debatable proposition), he's just kind of average (and, they mention it again here), and they suggest his brother Mike Pouncey is probably better.  Now, I can neither confirm nor deny the validity of their analysis.  That's not what I do.  All I can say is that this view of the two Pounceys is closer to what their measurables would suggest was a likely outcome.  Is it hear me out here... is it possible that a player taken in the 1st round, with all the hype and expectations that come with this, and is also selected by a celebrated and highly successful team, might be getting more credit than he deserves?  While Pro Bowls are a questionable way to measure success (especially for offensive linemen who people probably don't watch as closely as they think they do), I think they become even more meaningless with high draft picks, who sometimes just need to avoid screwing up, in order to make people happy.  A 7th round Pro Bowler, on the other hand, well, they probably really earned it.  Maybe Pouncey is an outlier, or maybe I am just completely wrong about all of this.

Outside of the Pouncey brothers, the main name that probably stands out is that of Nick Hardwick.  While his short shuttle time is only a little bit above average, his Kangaroo Score suggests he has much more explosive lower body power than most of his peers.  While he might not have had an exceptional score when it came to the primary trait that I look for, he did come through on the secondary trait I consider for a center.

Group 2
The Pretty Damned Impressive- Starting centers with short shuttle times from 4.50-4.59 seconds.
Player     Year   Pick#   Sh. Sht.  Kangaroo     Agility        GS     PB All Pro
Dan Koppen 2003 164 4.56 -0.762 -0.397 132 1 1
John Sullivan 2008 187 4.55 -0.251 0.759 61
Roberto Garza 2001 99 4.53 0.977 0.649 148
Eric Wood 2009 28 4.51 -0.117 0.963 47

Kyle Cook 2007          U 4.51 -1.059 0.963 50

Max Unger 2009 49 4.50 -1.493 1.289 48 1 1
-0.451 0.704

When players begin sneaking into Group 2, I start to pay attention.  How rare is it for a center to score this highly in their short shuttle drill?  Well, that is a somewhat complicated question.  So many players are shuffled around from position to position on the offensive lines, that it becomes hard to measure players who are strictly centers.  Still, since pretty much all centers are under 6'5" tall, we can narrow things down a bit.  Out of the 521 offensive linemen that I have in my database, I have short shuttle data for 498 of them.  If we take these 498 players, and eliminate everyone over 6'5", then we are left with only 17.87% of the remaining players who ran a short shuttle time under 4.59 seconds.  Since there are, on average, 41 offensive linemen taken in the NFL Draft, then that would mean the average draft class has 7.32 players who should probably qualify for this group.

The tricky thing here is that many of these 7.32 players per draft class still end up playing other positions.  Many end up playing guard quite successfully, such as David Decastro (4.56), Marshal Yanda (4.58), Andy Levitre (4.52) Logan Mankins (4.45) or even tackle, such as Matt Light (4.49) or Jordan Gross (4.39).  It's impossible to anticipate how teams will utilize these players (it is, either way, still a positive trait for other positions to have too).  In the 2013 Draft, 6 players were selected who met this short shuttle time/height requirement (a below average year, unfortunately), with 1 currently being listed by their team as a tackle, 3 listed as guards, and only 2 currently listed as centers.  What position they end up playing is anyone's guess, but this might give us some idea as to how these things typically break down.

Still, considering that these players with a short shuttle time under 4.59 seconds only make up about 17.87% of the draft population, it seems odd that they ended up with 18 out of 32 (56.25%, this includes the upcoming Group 3 players) of the starting center jobs in 2012.  That means that there were 3.15 times as many starting centers last year, with times under 4.59 seconds, than there should have been considering their relative rarity.  Players in Group 2 were also available about 24 picks later in the draft than the players in Group 1.

Dan Koppen is sort of the poster child for this whole theory.  He pretty much bombed every event at the combine...except for one.  The nice thing about this is that, if we believe that measurable athletic traits matter, then this helps to narrow down which drills matter for which positions.  Koppen was a pioneer.  A man ahead of his time.  Setting forth on a path of adventure, he went on a quest to see how far an athletically inferior specimen could go, with only a great short shuttle time.  I salute you Dan Koppen.

Overall, Group 2 looks reasonably impressive, and seems to have fewer random bozos in it than Group 1 (LRBP- Low Random Bozo Population, almost always an important factor).  Half of the players here have managed to make it to a Pro Bowl or an All Pro team (with only one of these three players being selected before the 5th round, which I believe adds some validity to the honor).  Roberto Garza, while lacking in awards, has at the very least had a fairly impressive career, having started 148 games.

Now, we're going to take things up a notch, and everything is going to get much weirder.

Group 3
The Freaks!- Starting centers with a short shuttle time under 4.49 seconds 
Player     Year   Pick#   Sh. Sht.  Kangaroo     Agility        GS     PB All Pro
David Baas 2005 33 4.48 0.518 1.108 128

Brad Meester 2000 60 4.43 0.702 1.336 193

Will Montgomery 2006 234 4.43 0.172 1.442 47

Nick Mangold 2006 29 4.40 -0.584 1.374 110 4 3
Evan Dietrich-Smith 2009          U 4.40 -0.778 1.587 9

Scott Wells 2004 251 4.40 0.278 1.233 107 1
Dominic Raiola 2001 50 4.35 0.789 1.373 188

Chris Myers 2005 200 4.35 0.213 1.727 95 2
Ryan Kalil 2007 59 4.34 -0.840 1.486 68 3 1
Samson Satele 2007 60 4.29 0.007 1.662 85

Matt Birk 1998 173 4.18            N/A 2.374 187 6 2
Jason Kelce 2011 191 4.14 -0.537 2.472 18

-0.005 1.598

UPDATE- 9/10/2014 : In the time since I first posted this, I have come across another, more plausible source, for Matt Birk's short shuttle time.  This new source lists his time as 4.45 seconds.  In the end, that still places him in Group 3, but I just wanted to make a note of this.

Well damn, that is one good looking group of centers we have there, and on average, they will only cost you a pick at the very beginning of the 5th round.  While I have to wonder about the authenticity of Matt Birk and Jason Kelce's short shuttle times, even if I added 0.25 seconds to their results, they would still be in Group 3.  It probably shouldn't come as a shock at this point to see that Birk, one of the most accomplished centers of the past decade would have shocking results in this area, though I am still stunned at how well he did.

Out of these 12 players, 5 have have gone to Pro Bowl or All Pro teams.  Two more players, Brad Meester and Dominic Raiola, have had at least solid Roberto Garza type careers, with a remarkably high number of games started by them.  You almost have to wonder if a Reverse Pouncey Effect might be at play here, since the relatively poor history of these player's teams (Jaguars and Lions respectively) probably doesn't bring them as much attention.

So, how rare are the players in Group 3?  Well, using the same method I used for Group 2, the numbers would suggest that only 8.03% percent of drafted players are likely to qualify.  Out of the 41 offensive linemen taken in an average draft, you should only expect to find 3.29 players who qualify.  You should remember, these 3.29 players are also part of the same group of 7.32 players that would fit into Group 2, and not in addition to them.  Just like with Group 2, there is still the issue that some of them will also end up playing positions other than center.  In the 2013 Draft, 3 players were selected who could fall into this group, and only one is currently listed as a center, while the other 2 are currently listed as guards (David Quessenberry and Jeff Baca).  Are you curious who this one unnamed center is?  I'll get to that in a minute.

You should also consider this.  These 12 players in Group 3 make up 37.5% of the starting centers, while only being 8.03% of the draft population.  This means that there are 4.6 times as many of them starting as you would expect, considering their rarity.  As players' short shuttle times improve, they continue to take an increasingly disproportionate number of the starting jobs.  They are also available, on average, nearly a full round later than their more common Group 1 peers, though this also means that many of them have to overcome their team's lower expectations of them.

The one player, out of all three of these lists, that I had to sort of cheat to include, was Evan Dietrich-Smith.  He doesn't really belong in this list at all, since the Packers had no intention of him being the starter.  Legendary center Jeff Saturday was the actual opening day starter, a 37 year old free agent acquisition that the Packers had just signed.  In fact, Saturday did start the first 14 games, before being benched (in his defense, you have to expect a decline in performance from a 37 year old).  Unfortunately, I don't have the combine data for Jeff Saturday, but it strikes me as very interesting to see who they replaced him with, and who is now penciled in as the opening day starter for 2013.  That they just happened to replace him with a player who had a 4.40 short shuttle time would seem like a peculiar coincidence.  You have to wonder how this happened.  Did Mike McCarthy, frustrated with Saturday, say,"Put in that undrafted kid from Idaho State.  Yeah the one who has been lurking around here for 4 years, and only started 3 games.  I'm feeling lucky!"  It seems unlikely that Dietrich-Smith's excellent short shuttle time was a factor they consciously  considered when making this move, since teams haven't shown many signs that this factor registers with them as terribly important in their draft decisions.  Nonetheless, the undrafted oddball rose to the top of the heap, sort of like Jeff Saturday, who was also an undrafted player.

Among the players in this list, Will Montgomery's might not be a household name.  Still his story is kind of interesting.  Drafted at the very end of the 7th round, back in 2006, it is probably safe to say people expected very little of him.  He bounced around the league, going from the Panthers, to the Jets, and finally to the Redskins, starting only 15 games in his first 6 years in the league, mostly at guard.  Then in 2011, someone got the bright idea to give him a shot at the center position, and according to Pro Football Focus, he flourished with this switch, and was ranked as their 5th overall center for the 2012 season.  The possibility that people would discover this hidden ability to play center, shortly before Montgomery turned 30, kind of strikes me as a cruel joke.

There is also the peculiar situation with Samson Satele (4.29 Short Shuttle), who started the first 11 games for the Colts, before suffering a knee injury.  So, here's a quick trivia question.  Who finished the remaining 5 games for the Colts?  Why it was A.Q. Shipley, who just happens to have a short shuttle time of 4.40 seconds, which would also put him into Group 3 (DUN! DUN! DUN!).  So, how did he perform, this undrafted player from the 2009 draft, who had never started a game in 4 years, on any of the three teams he had played for?  I really have no idea, but the Ravens seemed to think that this formerly forgotten player was worth acquiring in a trade, even though they had supposedly drafted their center of the future, Gino Gradkowski, in the 2012 draft (who happens to have a short shuttle time of 4.78 seconds, uh, oh!).  The odd coincidences seem to keep piling up.

In all of this rambling, I still haven't accounted for the 5 players from Group 1, for whom we have no combine data.  When you consider the disproportionate degree to which players with better short shuttle times seem to be dominating the center position, it could lead you to the not so wild hunch that 2 or 3 of these unaccounted for players probably would fall into Group 2 or 3 as well.  This would just push the overall representation of athletically gifted players even higher.  If just 2 of these 5 players only made it as high as Group 2, then 20 out of 32 (62.5%) starting center jobs would be occupied by a group of players that only make up 17.87% of the population.  It could conceivably be even higher.

You may have also noticed that as the players short shuttle times improved from Group 1, to Group 2, to Group 3, that there were corresponding improvements to the average Kangaroo Score for each group.  In reality though, there is only a R^2 value of about 0.17 between the individual player's short shuttle times and their Kangaroo Scores, which is a rather low correlation.  Still, the Kangaroo Score would be the second factor that I would probably consider when looking for a center.  As I said earlier, I think this additional factor might partially explain how Nick Hardwick has succeeded, while being in Group 1.  So, using the same criteria as before, what percentage of the population has a Kangaroo Score that is at least average (average being a result of "0"), while also running the short shuttle drill in under 4.49 seconds (might as well be greedy here and shoot for Group 3)?  That would be 4.01% of the population, or 1.64 players per draft class.  Despite the rarity of this sort of physical ability, 7 seven such players (21.875% of the 32) were starters in 2012, which is 5.45 times as many as you would expect if this was all just dumb luck.  In the 2013 Draft, two players were selected who fit this mold.  One was David Quessenberry, who I mentioned is now listed as a guard for the Texans, and the other is that same Group 3 center prospect that I previously said I would reveal.  So, let's see who this player is, and match them up against the first center who was selected in the 2013 draft.

Player     Year   Pick#   Sh. Sht.  Kangaroo     Agility
Travis Frederick 2013 31 4.76 -0.492 -0.091
Eric Kush 2013 170 4.35 0.221 1.639

Who would you guess will end up doing better?  Travis Frederick?  Or, Eric Kush?  Plus, Eric Kush was available 139 picks later than Travis Frederick.  Whether Kush will actually end up doing well, is impossible for me to say.  Despite that, the numbers would suggest that there is a high probability of this happening.

To what degree do NFL teams seem to care about these measurable factors when drafting a player?  Well, let's divide the 2012 starting centers, from whom we have gathered data, into 2 groups, and compare their results.  The first group will be players who were taken in the first two rounds.  The second group will be players who were taken later than the 6th round, or who went undrafted.

   # of Players     Sh. Sht.
1st to 2nd round 13 4.55
6th round to undrafted 11 4.45

Despite the fact that they are a seriously disproportionate number of the players who end up becoming starters (relative to the rarity of their physical ability), teams seem to place less emphasis on this in the first 2 rounds of the draft.  This would seem to suggest that they don't actually trust these numbers very much.  Still, despite their apparent lack of interest, they end up having to give (begrudgingly?) a surprisingly high number of the starting jobs to these athletically superior late round picks.  If we count the 2012 starters that we don't have short shuttle times for, then we wind up with an equal number (14) of starters that came from the first 2 rounds, as we have from the 6th round through the undrafted.  Then, when you consider that 7 out of 32 teams (21.875%) were starting centers who went undrafted, it really starts to paint a picture that suggests that teams might not really know what they are doing, or what they are looking for.

With such a disconcerting possibility, who can we place our faith in?  Maybe, we can place our faith in the freakishly gifted athlete.  Even when obstacles (GMs), and low expectations (their draft status), are placed in their path, they still just seem to rise to the top.

You can start laughing at!


  1. Wow......Samson satele is a sucky center..... Alex mack is a stud.... They are exceptions to this rule.... alex mack is a stud who's pushed halo-ti Ngata around a lot when he's played against him as for satele.... he's definitley not a starting caliber center as for matt birk.... wow.... I wouldn't have thought he was that good of an athlete when he was younger.... Rominic raiola.... age has caught up to him as for maurkice pounce he is definitley not average.... he's a very good center he is overrated.... to a degree.. 3 All pros for a good center...?he's dfinitley a pro bowler (he won't be this season in that zone blocking scheme...) As for the rest of these centers.... it's all dead on.... Almost eveyrone in group 3 was or is an elite cenetr Evan smith,Baas,Satele,kelce are either average,sucky or haven't started yet tho... this system isn't flawless but it's pretty good.... Love this site man

    1. For the most part, I wouldn't disagree with any of what you said. Alex Mack is somewhat interesting since his measurable traits outside of his short shuttle time are fairly respectable. So, while I wouldn't necessarily have bet on him to become a good center, I wouldn't have bet against him either.

      In the end, I posted this just because I thought it might be amusing to take an extremely simple view of the center position, and it ended up producing rather peculiar and interesting results. We could always add more factors to refine the process further. Still, if I had to choose between drafting 2 centers, I probably would be much more comfortable with one that fell into Group 3, where they have the best short shuttle results.

    2. I hate to say this..... But I think matt birk was almost guarenteed success.... and I love this post..... it has its flaws.... but it's really good. Birk was almost guarenteed sucess dynamic atheltiscm.... he was probaly the smartest man in football over the last decade..... I wonder how much he helped to teach and develop flacco I also agree if I ever become a football scout/front office person (Which I plan on becoming after I finish my high school and possible college football career if I don't make it in the NFL) Anyway love your blog it definitley is one of the better football blogs around.