Athleticism and the Offensive Line part 2

I've mentioned some of my thoughts about how athleticism relates to the offensive line, before, but thought it would be fun to add on to this subject a bit.  There are plenty of ways to try to examine the historic patterns as to which draft picks succeed, versus which ones fail.  Since it can be hard to separate a players performance from their hype, I prefer to mainly look at prospects who were taken in the mid-to-late rounds, that managed to emerge as successes despite being somewhat less heralded.  If these successes have certain traits in common, that is probably worth paying attention to.  On the flip side of that, if the highly drafted players, who become busts/disappointments, have certain weaknesses that they share with each other, that is also interesting.

While I'm not a big fan of using CarAV (Career Approximate Value) as a stat, it can be useful sometimes.  For one, it has some established history, and people are already somewhat aware of it as a statistical measure.  Secondly, when applied to players at similar positions, from the same draft classes, it gives a reasonably acceptable measure of who is achieving some degree of success.  I still prefer not to use it for broader comparisons, across the span of different draft classes.

So, what I'm going to do, is to show the three offensive linemen from each draft class who achieved the highest CarAV score, while being selected in the 3rd round or later (not including undrafted players).  For each draft class, I will also include the player with the lowest CarAV, who was selected in the first two rounds, to represent the "bust" for that year.  To keep this from running too long, I will just show the results from the 2005 to 2010 draft classes.

Since I am comparing a player's success based on CarAV to their measurables from the combine, the comparison won't work if the player either didn't perform at the combine or their college pro day, or at least didn't perform enough of the drills to get an accurate picture.  In this case, I will just take the next highest (or lowest rated player, for the 'bust' group) based on CarAV, who has enough data available.  Fortunately this was rarely an issue, and I only had to make two such substitutions.

Beyond just showing their CarAV score, I will show their overall draft position, Kangaroo Score, Agility Score, and their average number of games started per year since they have been in the league (GS/Year).  As always, the Kangaroo Score and Agility Score are given as the number of standard deviations above or below average that a player is, relative to the average results at their position group. 

Player                     Year     Pos.      Pick#          Kangaroo         Agility     CarAV    GS/Year

Jared Veldheer	2010	T	69	1.070	1.246	18	14.33
J'Marcus Webb	2010	T	218	0.002	-0.207	15	14.66
J.D. Walton	2010	C	80	-0.869	0.451	14	12
Vladimir Ducasse	2010	G/T	61	0.630	-1.384	1	0.33
Louis Vasquez	2009	G	78	1.070	-0.267	25	13.5
T.J. Lang	2009	G/T	109	-0.571	1.600	21	8.5
Matt Slausen	2009	G	193	1.337	-0.425	15	16
Jason Smith	2009	T	2	-1.351	0.575	7	6.5
Carl Nicks	2008	G	164	1.032	-0.072	45	13.6
Josh Sitton	2008	G	135	1.049	1.005	34	12.5
John Sullivan	2008	C	187	-0.251	0.759	23	12.2
Chris Williams	2008	T	14	-0.475	-0.559	15	7.6
Marshall Yanda	2007	G	86	-0.611	1.146	36	12
Jermon Bushrod	2007	T	125	0.294	0.129	36	10.33
Doug Free	2007	T	122	1.248	1.233	26	9.16
Aaron Sears	2007	G	35	0.524	0.864	11	15.5
Jahri Evans	2006	G	108	-0.724	-0.686	78	16
Eric Winston	2006	T	66	0.612	1.879	42	14.71
Jeromey Clary	2006	T	187	-0.088	0.171	40	11.14
Ryan Cook	2006	C	51	-0.225	-0.083	19	7.28
David Stewart	2005	T	113	1.011	N/A	42	13
Chris Myers	2005	C	200	0.213	1.727	41	11.87
Nick Kaczur	2005	T	100	0.679	0.120	35	12.4
Marcus Johnson	2005	G/T	49	0.208	0.074	7	3.6
Avg for Successes			130	0.361	0.577
Avg. for Busts			35.33	-0.114	-0.085

Before I get any further, I should mention a few issues.  Aaron Sears, who is the low CarAV player representing the 2007 draft class, had his career cut short by injury.  By all accounts, he played quite well.  Unfortunately there wasn't a better player to choose from for 2007 to represent the the disappointing way that things can turn out.   If we removed Sears from the equation, the average Kangaroo Score results for the 'busts' would drop to -0.242 (from -0.114), and the average Agility Score for busts would drop to -0.275 (from -0.085).  I also could have used Jeff Otah, rather than Chris Williams, to represent the 'bust' pick for 2008, but the data for Williams was more complete, since he did the agility drills.  Either way, their results were similarly disappointing, and wouldn't have significantly changed the overall results.

In T.J. Lang's case I only had his short shuttle score, but not his 3-cone score.  So, I used this, since it was all that was available.  The same issue applies to Jermon Bushrod, who also didn't do the 3-cone drill.  Similarly, with Josh Sitton, I had to use just his broad jump score, without his vertical jump score, to get his Kangaroo Score. 

Lumping all offensive linemen together, without any regard to their actual position, isn't my preference.  As I have said before, I think the Kangaroo Score is more vital to tackles, than to guards.  Centers, I feel, are mostly dependent on their short shuttle score, which is part of their overall Agility Score.  Just because somebody is listed above, doesn't mean that I would vouch for them, or believe that they are particularly good.  This is just meant to show which players are getting on the field and achieving some success in the eyes of CarAV, despite being taken later in the draft.  Still, even with this less than ideal approach to the issue, I think certain things are fairly obvious.

Among the mid-to-late round players, who became successful to varying degrees, their athletic ability tended to be above average.  Out of 18 arguable successes, only Jahri Evans (and possibly J.D. Walton to a lesser degree), demonstrated what could be seen as significantly below average athletic results.  If we were to say that Evans was a statistical outlier, and remove his results, the average Kangaroo Score for the 'successes' would go up to 0.425 (from 0.361), and the Agility Score would go to 0.655 (from 0.577).  For the most part, the other players demonstrated at least average overall athletic ability, though most were fairly dominant in at least one of the two categories.  11 out of 18 went so far as to produce scores of at least one full standard deviation above the average for their position, in at least one category, which is quite exceptional.

As for the highly drafted players, who perhaps failed to meet the expectations of their draft position, their results tended to be predictably average to poor.  Ignoring the Aaron Sears issue, only Jason Smith and Vladimir Ducasse managed to produce somewhat above average results in either of the two athletic measurements.  Unfortunately for both of them, these acceptable scores in one category, were overwhelmed by their remarkably awful scores in the other.  While I feel that a player doesn't necessarily need to excel in both measurements, these sorts of stunningly bad scores are fairly obvious warning signs.

So, once again, it seems to me that the athletically gifted players are the sensible ones for a team to select.  The late round players, whom teams had lower expectations, seem to rise to the top when they have above average measurable athletic ability.  The highly drafted players, who are given every opportunity to succeed since their team had so much faith in them, appear to repeatedly fail when their athletic ability just doesn't measure up.  If we were to leave out the results of Jahri Evan and Aaron Sears, the average Kangaroo Score of the mid-to-late round successes would be 0.667 standard deviations higher than the early round 'busts'.  Similarly, the mid-to-late round successes would have an average Agility Score that was 0.930 standard deviations higher than the average result for their more highly drafted, and disappointing counterparts.  You would think that this would be obvious, but it doesn't seem to have much effect on people's decision making process.

Or, maybe I'm just nuts.