SITE NEWS: We are moving all of our site and company news into a single blog for Sports-Reference.com. We'll tag all PFR content, so you can quickly and easily find the content you want.
Also, our existing PFR blog rss feed will be redirected to the new site's feed.
Pro-Football-Reference.com ยป Sports Reference
For more from Chase and Jason, check out their work at Football Perspective and The Big Lead.
ShareThisPro success of college award winners
This post is a fuller description of a study I did for ESPN Magazine. They asked me to look at the NFL success of the winners of the various postseason college football awards, such as the Heisman, the Outland, the Lombardi, and so on. Do Heisman Trophy winners, in general, have more or less NFL success than winners of the Doak Walker Award, or the Jim Thorpe, or the Biletnikoff? Which college award has traditionally produced the best pros? Let's find out.
The first thing I did was to collect the winners of 17 awards:
- The Heisman Trophy
- The Butkus Award (top linebacker)
- The Outland Trophy (top interior lineman)
- The Bednarik Award (defensive player of the year)
- The O'Brien Award (top QB)
- The Unitas Golden Arm Award (top senior QB)
- The Doak Walker Award (top running back)
- The Fred Biletnikoff Award (top wide receiver)
- The Jim Thorpe Award (top defensive back)
- The John Mackey Award (top tight end)
- The Lou Groza Award (top place kicker)
- The Bronco Nagurski Award (defensive player of the year)
- The Ray Guy Award (top punter)
- The Vince Lombardi Award (top lineman)
- The Hendricks Award (top defensive end)
- The Rimington Trophy (top center)
- The Tatupu Award (top special teams player)
The study covers the award winners from 1993 to 2005. For the several awards that didn't exist in 1993, it covers all award winners through 2005.
So, with all the players in place, I need a way to measure their NFL success. As we go through it, keep in mind that it is not meant to be a precise metric, but rather an approximate measure of value. Comparing a linebacker who has been in the league for nine years to a running back who is in his second year is very tough to do, so all I'm hoping to do is group guys into broad categories that seem reasonable. I'm going to put a number between 0 and 18 on each player. Peyton Manning is an 18. So are Warren Sapp and Randy Moss. Julius Peppers is a 14, Garrison Hearst a 12, Terrell Suggs a 10, Dominic Raiola an 8, E.J. Henderson a 6, Rashaan Salaam was a 4, Byron Hanspard a 2, and Eric Crouch a zero.
Again remember that the goal here is not to forever put an end to the debate about whether Daniel Graham or Antonio Bryant has had the better career. That's too ambitious a goal. We simply want to classify them both as being a bit better than Michael Bishop or Travis Dorsch, but not as good as Terry Glenn or Carson Palmer.
OK, so here's how I arrived at these ratings:
First, I ignored the 2007 NFL season altogether. At the time I had to turn the study in, many of the metrics I would be using --- pro bowl status, for example --- weren't yet determined. One consequence of this is that the 2006 award winners are not included in the study at all. Another is that players who had great 2007 seasons will be slightly undervalued.
For players whose college careers ended in 2003 or earlier, their rating is
Score = YP + YS + 2*PB
where YP = years played (that is, number of seasons in which he appeared in an NFL game), YS = years starting (the number of years in which he was his team's main starter at his position), and PB = Pro Bowls.
Further, I put a cap on each category. If a player had a YP of more than 3, I simply counted it as 3. If he had a YS of more than 5, it counts as 5. If he had a PB of more than 5, it counts as 5. The reasoning here is that it doesn't really tell us much about Will Shields and Orlando Pace that Shields went to 12 pro bowls and Pace only went to 7. If you went to five or more pro bowls, you were pretty great. And, for this exercise, we're not interested in splitting hairs of greatness.
Now, for players who haven't had as much of a chance to pad their resume, we need to project just a little bit. We'll do so by giving them a bit of extra credit for being a starter and for making a pro bowl. For players whose college careers ended in 2004, their rating is:
Score = YP + 3*YS + 5*PB
For players whose college careers ended in 2005:
Score = YP + 5*YS + 8*PB
Now that we've got all the players rated, it's just a matter of adding up the average rating of the winners of each award. Note that multiple winners of the same award are only counted once. Jason White won the O'Brien Award twice, for instance, but his zero only counts once in the average for the O'Brien Award.
Here is the summary:
Players Pts AVG ------------------------------- lombardi 13 114 8.8 outland 14 122 8.7 nagurski 12 103 8.6 bednarik 9 73 8.1 walker 13 83 6.4 thorpe 14 89 6.4 butkus 13 81 6.2 biletnikoff 12 70 5.8 heisman 14 75 5.4 unitas 14 62 4.4 obrien 12 53 4.4 hendricks 3 13 4.3 rimington 6 26 4.3 mackey 6 25 4.2 tatupu 9 34 3.8 groza 12 38 3.2 guy 5 11 2.2
I'll attach the complete data set below, but the one-sentence summary is that, while the awards for the glory positions sometimes go to NFL non-entities (Danny Wuerffel, Eric Crouch, Luke Staley, Mike Hass), the trophies for defenders and linemen seem to go to players who are always at least adequate performers at the NFL level. Even the worst Lombardi winners of the period studied ended up being NFL starters for at least two seasons.
Here's the full data set:
Year Award Player NFL value Notes ============================================================================================== 2006 heisman Troy Smith still in college or rookie in 2007, not counted. 2005 heisman Reggie Bush 1 2004 heisman Matt Leinart 6 2003 heisman Jason White never played in NFL (before 2007) 2002 heisman Carson Palmer 10 2001 heisman Eric Crouch never played in NFL (before 2007) 2000 heisman Chris Weinke 4 1999 heisman Ron Dayne 3 1998 heisman Ricky Williams 10 1997 heisman Charles Woodson 16 1996 heisman Danny Wuerffel 3 1995 heisman Eddie George 16 1994 heisman Rashaan Salaam 4 1993 heisman Charlie Ward never played in NFL (before 2007) 1992 heisman Gino Torretta 2 2006 bednarik Paul Posluszny still in college or rookie in 2007, not counted. 2005 bednarik Paul Posluszny still in college or rookie in 2007, not counted. 2004 bednarik David Pollack 2 2003 bednarik Teddy Lehman 4 2002 bednarik E.J. Henderson 6 2001 bednarik Julius Peppers 14 2000 bednarik Dan Morgan 10 1999 bednarik LaVar Arrington 14 1998 bednarik Dat Nguyen 7 1997 bednarik Charles Woodson 16 1996 bednarik Pat Fitzgerald never played in NFL (before 2007) 1995 bednarik Pat Fitzgerald multiple winner, this row not counted 2006 biletnikoff Calvin Johnson still in college or rookie in 2007, not counted. 2005 biletnikoff Mike Hass never played in NFL (before 2007) 2004 biletnikoff Braylon Edwards 5 2003 biletnikoff Larry Fitzgerald 8 2002 biletnikoff Charles Rogers 3 2001 biletnikoff Josh Reed 4 2000 biletnikoff Antonio Bryant 7 1999 biletnikoff Troy Walters 3 1998 biletnikoff Troy Edwards 4 1997 biletnikoff Randy Moss 18 1996 biletnikoff Marcus Harris never played in NFL (before 2007) 1995 biletnikoff Terry Glenn 10 1994 biletnikoff Bobby Engram 8 2006 butkus Patrick Willis still in college or rookie in 2007, not counted. 2005 butkus Paul Posluszny still in college or rookie in 2007, not counted. 2004 butkus Derrick O. Johnson 8 2003 butkus Teddy Lehman 4 2002 butkus E.J. Henderson 6 2001 butkus Rocky Calmus 4 2000 butkus Dan Morgan 10 1999 butkus LaVar Arrington 14 1998 butkus Chris Claiborne 8 1997 butkus Andy Katzenmoyer 3 1996 butkus Matt Russell 1 1995 butkus Kevin Hardy 10 1994 butkus Dana Howard 2 1993 butkus Trev Alberts 3 1992 butkus Marvin Jones 8 2006 groza Art Carmody still in college or rookie in 2007, not counted. 2005 groza Alexis Serna still in college or rookie in 2007, not counted. 2004 groza Mike Nugent 8 2003 groza Jonathan Nichols never played in NFL (before 2007) 2002 groza Nate Kaeding 8 2001 groza Seth Marler 2 2000 groza Jonathan Ruffin never played in NFL (before 2007) 1999 groza Sebastian Janikowski 8 1998 groza Sebastian Janikowski multiple winner, this row not counted 1997 groza Martin Gramatica 10 1996 groza Marc Primanti never played in NFL (before 2007) 1995 groza Michael Reeder never played in NFL (before 2007) 1994 groza Steve McLaughlin 2 1993 groza Judd Davis never played in NFL (before 2007) 1992 groza Joe Allison never played in NFL (before 2007) 2006 lombardi Lamar Woodley still in college or rookie in 2007, not counted. 2005 lombardi A.J. Hawk 6 2004 lombardi David Pollack 2 2003 lombardi Tommie Harris 10 2002 lombardi Terrell Suggs 10 2001 lombardi Julius Peppers 14 2000 lombardi Jamal Reynolds 3 1999 lombardi Corey Moore 2 1998 lombardi Dat Nguyen 7 1997 lombardi Grant Wistrom 8 1996 lombardi Orlando Pace 18 1995 lombardi Orlando Pace multiple winner, this row not counted 1994 lombardi Warren Sapp 18 1993 lombardi Aaron Taylor 8 1992 lombardi Marvin Jones 8 2006 walker Darren McFadden still in college or rookie in 2007, not counted. 2005 walker Reggie Bush 1 2004 walker Cedric Benson 2 2003 walker Chris Perry 3 2002 walker Larry Johnson 9 2001 walker Luke Staley never played in NFL (before 2007) 2000 walker LaDainian Tomlinson 16 1999 walker Ron Dayne 3 1998 walker Ricky Williams 10 1997 walker Ricky Williams multiple winner, this row not counted 1996 walker Byron Hanspard 2 1995 walker Eddie George 16 1994 walker Rashaan Salaam 4 1993 walker Bam Morris 5 1992 walker Garrison Hearst 12 2006 nagurski James Laurinaitis still in college or rookie in 2007, not counted. 2005 nagurski Elvis Dumervil 1 2004 nagurski Derrick O. Johnson 8 2003 nagurski Derrick Strait 3 2002 nagurski Terrell Suggs 10 2001 nagurski Roy Williams 16 2000 nagurski Dan Morgan 10 1999 nagurski Corey Moore 2 1998 nagurski Champ Bailey 18 1997 nagurski Charles Woodson 16 1996 nagurski Pat Fitzgerald never played in NFL (before 2007) 1995 nagurski Pat Fitzgerald multiple winner, this row not counted 1994 nagurski Warren Sapp 18 1993 nagurski Rob Waldrop 1 2006 obrien Troy Smith still in college or rookie in 2007, not counted. 2005 obrien Vince Young 14 2004 obrien Jason White never played in NFL (before 2007) 2003 obrien Jason White multiple winner, this row not counted 2002 obrien Brad Banks never played in NFL (before 2007) 2001 obrien Eric Crouch never played in NFL (before 2007) 2000 obrien Chris Weinke 4 1999 obrien Joe Hamilton 1 1998 obrien Michael Bishop 1 1997 obrien Peyton Manning 18 1996 obrien Danny Wuerffel 3 1995 obrien Danny Wuerffel multiple winner, this row not counted 1994 obrien Kerry Collins 10 1993 obrien Charlie Ward never played in NFL (before 2007) 1992 obrien Gino Torretta 2 2006 outland Joe Thomas still in college or rookie in 2007, not counted. 2005 outland Greg Eslinger never played in NFL (before 2007) 2004 outland Jammal Brown 13 2003 outland Robert Gallery 6 2002 outland Rien Long 3 2001 outland Bryant McKinnie 8 2000 outland John Henderson 12 1999 outland Chris Samuels 16 1998 outland Kris Farris 1 1997 outland Aaron Taylor never played in NFL (before 2007) 1996 outland Orlando Pace 18 1995 outland Jonathan Ogden 18 1994 outland Zach Wiegert 8 1993 outland Rob Waldrop 1 1992 outland Will Shields 18 2006 rimington Dan Mozes still in college or rookie in 2007, not counted. 2005 rimington Greg Eslinger never played in NFL (before 2007) 2004 rimington Ben Wilkerson 1 2003 rimington Jake Grove 5 2002 rimington Brett Romberg 1 2001 rimington LeCharles Bentley 11 2000 rimington Dominic Raiola 8 2006 tatupu A.J. Trapasso still in college or rookie in 2007, not counted. 2005 tatupu Ryan Hoffman never played in NFL (before 2007) 2004 tatupu Chad Owens 2 2003 tatupu Wes Welker 3 2002 tatupu Glen Pakulak never played in NFL (before 2007) 2001 tatupu Kahlil Hill 1 2000 tatupu J.T. Thatcher never played in NFL (before 2007) 1999 tatupu Deltha O'Neal 11 1998 tatupu Chris McAlister 14 1997 tatupu Brock Olivo 3 2006 thorpe Aaron Ross still in college or rookie in 2007, not counted. 2005 thorpe Michael Huff 6 2004 thorpe Carlos Rogers 5 2003 thorpe Derrick Strait 3 2002 thorpe Terence Newman 7 2001 thorpe Roy Williams 16 2000 thorpe Jamar Fletcher 3 1999 thorpe Tyrone Carter 4 1998 thorpe Antoine Winfield 8 1997 thorpe Charles Woodson 16 1996 thorpe Lawrence Wright 2 1995 thorpe Greg Meyers never played in NFL (before 2007) 1994 thorpe Chris Hudson 7 1993 thorpe Antonio Langham 7 1992 thorpe Deon Figures 5 2006 unitas Brady Quinn still in college or rookie in 2007, not counted. 2005 unitas Matt Leinart 6 2004 unitas Jason White never played in NFL (before 2007) 2003 unitas Eli Manning 5 2002 unitas Carson Palmer 10 2001 unitas David Carr 8 2000 unitas Chris Weinke 4 1999 unitas Chris Redman 3 1998 unitas Cade McNown 3 1997 unitas Peyton Manning 18 1996 unitas Danny Wuerffel 3 1995 unitas Tommie Frazier never played in NFL (before 2007) 1994 unitas Jay Barker never played in NFL (before 2007) 1993 unitas Charlie Ward never played in NFL (before 2007) 1992 unitas Gino Torretta 2 2006 mackey Matt Spaeth still in college or rookie in 2007, not counted. 2005 mackey Marcedes Lewis 1 2004 mackey Heath Miller 8 2003 mackey Kellen Jr. Winslow 3 2002 mackey Dallas Clark 7 2001 mackey Daniel Graham 6 2000 mackey Tim Stratton never played in NFL (before 2007) 2006 guy Dan Sepulveda still in college or rookie in 2007, not counted. 2005 guy Ryan Plackemeier 6 2004 guy Dan Sepulveda still in college or rookie in 2007, not counted. 2003 guy B.J. Sander 2 2002 guy Mark Mariscal never played in NFL (before 2007) 2001 guy Travis Dorsch 1 2000 guy Kevin Stemke 2 2006 hendricks Lamar Woodley still in college or rookie in 2007, not counted. 2005 hendricks Elvis Dumervil 1 2004 hendricks David Pollack 2 2003 hendricks David Pollack multiple winner, this row not counted 2002 hendricks Terrell Suggs 10
This entry was posted on Thursday, January 3rd, 2008 at 11:02 am and is filed under Approximate Value, College, General, History. You can follow any responses to this entry through the RSS 2.0 feed. Both comments and pings are currently closed.
You could also say that defense is defense, whether in college or the pros. A lot of those Heisman and O'Brien awards went to guys who play in exotic offenses that don't exist in the NFL. But if you're fast and can hit, you'll probably do alright in any league.
I would be inclined to count multiple winners. If we are measuring which awards correlate to NFL success more, and an award chooses to make a player a two time winner and they never play in the NFL, that should count against the award.
How concerned are you that some of the separation with your rankings is infected with "information vs. reputation" problem. To get to 18, a player must have made 5 pro bowls. Who is most likely to make consecutive pro bowls for several years? Lineman.
I'd say the O'Brien award would be lower ranked, once the newer awards have more time. It went almost 10 years between Manning and Vince Young without having a QB who would start for more than a season.
Special Teams:
Hmm I'd say the only one where your method is dubious is special teams, as someone who is a great special teams player may be a great special teams player in the NFL but never be a starter.
Other than that pretty fascinating.
Does this analysis account for the difference int total number of starters at all? For example, there are only 32 starting quarterbacks but 64 starting offensive tackles yet 5 years played at being a starting quarterback counts the same in the rating system as 5 years starting at offensive tackle.
Maybe you could weight the quarterback/tailback/kicker/punter position differently or maybe you can change the rating system so a player only gets points in his "slotted" position (like a person who played tackle in college and started his career as a tackle would only get points as a tackle and not if he shifted to guard).
I think Malemute is on to something important here. There would probably have been more award winning QBs who were starters if the league needed 64 starting QBs.
I'm going to disagree with Malemute.
There are twice as many starting tackle slots in the NFL as there are starting QB slots, but there are also twice as many people vying for those slots. There are twice as many tackles in the league as there are QBs (roughly). There are twice as many tackles drafted as there are QBs (again, roughly). A starting NFL tackle is (roughly) in the top 64 of a group of 128. A starting NFL QB is in the top 32 of a group of 64ish. I don't see why it makes a difference.
Or, to look at a different way, what if I just changed the categories? Who says I was using "tackle" as a position anyway? There are only 32 starting LTs in the league, same as QBs. A player gets as much credit for starting at LT as he does for starting at QB. If you argue that LT and RT are really the same position, then why not take that logic a step further and say that "OL" is a position? There are 160 starting offensive linemen in the league, and only 32 starting Strong Safeties. Therefore much more credit should be given to starting Strong Safeties than to starting offensive linemen. Do you believe that? There are 352 starting "defensive players" in the NFL and only 32 starting TEs. Do you believe that starting tight ends should be given more credit (11 times more!) than starting defensive players?
Yes, I know I'm being ridiculous here. The point I'm trying to make is that the labels we give to the positions guys play shouldn't change the value of being a starter.
One more example: let's say that about half the league was currently using a one-back set as their base offense. So there are only 16 starting fullback slots in the NFL. Would you then argue that starting fullbacks should be given more credit than starting QBs?
If you want to argue that a starting QB should get more credit than a starting LT because QB is a more important position than LT, that's reasonable (although it opens up a much bigger can of worms than I know how to deal with). But if you want to argue that a starting QB should be given more credit than a starting LT, simply because there happens to be another position named RT, I don't understand the argument.
I'll try to jump in the middle of this. Doug, I don't think Malemute's argument is that a QB should be given more credit than a LT because there is also a position called RT. I think he is saying more that an award winner who won the award as a Tackle and is drafted at Tackle (Let's call him "Robert Gallery") should only get credit for starting at Tackle, and if he is not good enough to play tackle, but continues to find employment at a guard position, then that kind of player will be overvalued relative to a similarly disappointing QB, who will either be the starter at his drafted position, or sitting the bench, and not get to move to a lesser valued position, while still accumulating "starts".
I guess this depends on what question you are trying to answer, but I see the merit in both sides, as to the specific issue posed here. I will lean toward Doug on this particular topic, though. Basically, you are trying to see who even sniffed the NFL, who was a star, etc. At best, it would change about 1 point (Gallery, after all, did start 3 years at Tackle before moving to Guard this year).
If the question, on the other hand, was "Are highly drafted Quarterbacks more likely to be busts than highly drafted Left Tackles", then Malemute's point would be correct. Teams do not draft offensive linemen in the top 10 because they hope they can play right guard. Using GS to answer that question might lead one to conclude that drafting an offensive lineman was much safer (after all, Tony Mandarich ended up starting started 63 games in his career, 0 at LT, 53 at RT, and 10 at RG), even if he could never play the position that he was drafted, and paid big money, to play.
Also, in thinking about this, I think their may be some circular reasoning, related to a "reputation/information" issue in college. In college, the reputation is your perceived blue chip stock and future draft status. The information is your stats.
Let's take Glenn Dorsey, and Brian Brohm as examples. Coming into this season, I could have told you that these guys are top NFL prospects, based on reputation, and I have not seen enough of any of them play to offer a different opinion.
Brohm plays a position that accumulates lots of statistics, and he also gets credit/blame for Wins/Losses, even if they are not primarily his fault. Information trumps Reputation here, and when his team doesn't even make a bowl game (mostly due to a bad defense), he has no chance of winning any awards.
Dorsey plays a position, in interior defensive line, that is less defined by statistics. Dorsey plays hurt part of the year, and the team defense performance goes down during that period. If he were a quarterback, there is no way he wins an award if he is playing hurt part of the year and it affects his numbers. But, when he was/is healthy, he was dominant. Reputation, including the knowledge that he will be a top 5 pick, before the award is selected. He wins.
So, is it that Outland winners are more likely to be successful in the NFL, or is it that players who are more likely to be successful in the NFL (higher draft picks at their position) are more likely to win awards like the Outland where the decision is based more on reputation than information?
Let's look at the K position. Each roster has one K. There are 32 YP points available to kickers. There are 32 starting K. There are 32 YS points available to kickers. There are two Pro Bowl roster spots for kickers. There are 20 PB points available for kickers. Add them all up and there are 84 points available to be divied up by kickers.
Now let's take a look at the OLB position. Each roster has at least 3 OLB, There are at least 94 YP points available. There are two starting OLB. There are 64 YS points available. There are 6 Pro Bowl roster spots for OLB. There are 60 PB points for OLB. Total there is 218 points to be divied up by OLB.
Under the current point system, there are more points available to certain positions.
Let's divide the awards into Category A and Category B. Category A: awards that reward positions that are limited to 32 starting NFL positions (Mackey, Groza, Guy, O'Brien, Unitas, Walker, Rimington), and Category B: awards that reward positions that have more than 32 NFL starting positions available (the remainder of the awards). I will leave out the Tatupu award (special teams). The average of the 'AVG' point score column for Category A is 4.16. The average Category B is 6.92; a mean higher than the highest point value in Category A.
Another explanation for the spread of points may have to do with position turnover (ie how many new jobs at that position are available each year). Maybe the positions where players last longer (K, P, etc.) are harder to score points in because less of the NFL jobs become available each year.
I will try to clarify my earlier argument. I am not saying that a QB should be given more credit than a LT. My point was, is that for every P, K, TE, etc that wins an award there are only 32 NFL positions he can play. I theorized that a potential way to balance that would be to only let players get credit for the position they were "slotted" as. For example, if Gallery is drafted as a LT he only gets points for playing or pro bowls as a LT. If he starts as a RT, G, C whatever then he would not get any points. Since there are 32 starting LT spots and 32 starting QB spots it would be equal.
I would argue that there is roughly the same number of people competing for each position and that the few thousand NFL players are picked from a pool that starts in the hundreds of thousands that play some form of amateur football. Some positions have the benefit of having more points available and the benefit of more roster spots; allowing that player more chances to stick around on a roster and get more chances to score from a larger pool of points.
Today on a mailing list I peruse someone made the claim that the last 10 NFL QB's to have made their first playoff start are 0-10. Any way to check that?
Is anyone else completely shocked that the punting and kicking awards are dead last?
I would have assumed going in that they would have the highest correlation, what with the college and pro games being largely the same for them, and the relatively low number that get in per season (assuming the average punter/kicker lasts about 10 years, that's only 3 per draft), the award winners would be the front-runners to get those slots.
Also, Malemute has a point: 6 QBs make the pro bowl each year. 2 Punters and 2 Kickers do. That's a higher standard to reach.
And getting into the pro bowl on reputation as a punter is probably a lot harder than other positions: The stats generally give a solid representation of the best in the position for the year. This years are #1 and #2 in both gross and net yards per punt. Andy Lee also set the record for most punting yards in a season. The announcers somehow missed that when they called the last 49er game.
Interesting survey, and valuable info. Some comments:
First of all, perhaps this was explained and I missed it, but why on earth did this survey go with an 18 point scale? It's cumbersome and non-intuitive. If you had to use 18 for some reason, why not multiply all the results to standardize them to a 10 point scale familiar to those of us accustomed to math on the planet Earth?
It's pointless to quibble over the individual ratings when they're averaged out by multiple winners but... Vince Young 14? I hope that's a typo, though I'm not sure if the 1 or the 4 is the extra digit, in that case.
I also vote to count multiple winners twice. If that was the best guy in the nation that year at his position, then count his pro career.
I'm surprised how little use there is from the punter and kicker spots. College kickers are so uniformly awful, place kickers cost top teams crucial games every year shanking 27 yarders, and pro kickers are quite often awful too. Yet kickers are hardly ever drafted, and half the teams in the league are looking for a new kicker every year. I'd think the top college kickers would translate directly to the pros, since the skill set is identical and only the ball differs a bit, but obviously that's not the case.
As for QB value, the odds are stacked against QBs since the position is unique. Big fat guys at tackles/guards/centers can switch around, cornerbacks/safeties/nickel backs rotate, linebackers/DEs overlap, etc. Furthermore, defenses change constantly with different defensive packages, subs go in to keep the starters fresh, etc. A talented athlete on defense can (and will) therefore fit into a variety of positions and contribute in meaningful fashion even if he's not a starter at his college position. Special teams skews things too, since no one is drafted for special teams play, but it's definitely a plus if it exists, and players can stick around the league as non-starters if they're good on kick coverage/returns. Imagine if Devin Hester had won a college CB award, for instance?
None of this is true for QBs. The odd experiment as a WR or "slash" player aside, QBs have to be good as QBs or they're out of the league. Plus, there's just one QB playing at a time per team, no switches for different game situations, no subing in a backup to keep the starter fresh (resting for playoffs/blowouts aside), and I'd also think (without stats to back it up) that QB is a harder position to get playing time in, since established QBs stick around longer than any other position, other than maybe kicker. You don't see a lot of 35 y/o starting defenders. Even 33 or 34 most are too slow. RBs are over the hill at 30, when QBs are just hitting their stride. This means new Qbs have fewer starting opportunities, and need to wait longer for their chance, than any other position on the field. This might be balanced out somewhat by marginal qbs sticking around for a decade and gradually increasing their point value, though.