Are people looking at the wrong stats when assessing QBs - can we come up with something better?

[Darth Tater]

True, but here, the debate is about how to assess probability. My position is that all statistics labeled "QB stats" actually measure the QB combined with his support system -- and, as such, are useless numbers. What's your opinion on that?

I don't know if how good your measurement assess that probability. If it woks, you get paid. If not, you get fired.

Thing is, are you trying to predict the guy's future performance or are you just trying to show that the guy has not performed?

[Lombardi's_kid_brother]

I don't think OF is accurate about the idea that a team makes a QB so all the individual numbers are worthless.

Steve Young was mentioned earlier. For two years in Tampa, his completion percentage stunk. For two years in San Fran, his completion percentage stunk.

He did not become a better quarterback simply because his supporting cast and scheme got better. He became a better quarterback because he suddenly became a better quarterback. On the '88 49ers, he was a 53 percent passer who threw as many interceptions as TDs. (It's 100 passes, which is a pretty decent sample size for a backup). He was still an incaccurate mess despite the presence of Bill Walsh and Jerry Rice.

It really seems like for him, it took a long time for the light to go on.

A lot was made in The Blind Side of Walsh turning DeBerg into an accurate passer for one year. But DeBerg played forever and was always a pretty accurate passer regardless of the system he played in. He was awful as a rookie, but a ton of players have been awful as rookies - like John Elway.

[Oldfan]

I don't know if how good your measurement assess that probability. If it works, you get paid. If not, you get fired.

Thing is, are you trying to predict the guy's future performance or are you just trying to show that the guy has not performed?

Most in this thread are looking for a way to accurately grade the ability of a QB using statistics. If that was possible, they could accurately assess the probability of the QB making a successful transition from college to the NFL. I maintain that statistics are useless for that purpose because they are measuring the QB along with his support system and not the QB in isolation.

[Lombardi's_kid_brother]

Most in this thread are looking for a way to accurately grade the ability of a QB using statistics. If that was possible, they could accurately assess the probability of the QB making a successful transition from college to the NFL. I maintain that statistics are useless for that purpose because they are measuring the QB along with his support system and not the QB in isolation.

So, the only way to judge a QB is to stand him up and watch him throw the ball? At the end of the day, you are putting an awful lot of faith in the fat man with the tobacco staring at a 21-year-old's ****ing footwork.

Do you think ANY individual stat in football is worthwhile?

And it's not all raw numbers for me. I would never draft a QB shorter than 6'1. (And, yes, I would have missed on Drew Brees under that rule. So be it). I would only draft a QB from a D-1 program or from a D-1-AA program that went really really far in the D-1-AA playoffs. (And, yes, I would refuse to call it FBS or whatever). I would avoid QBs who ever ran the read option.

Like I said, this exercise is really about narrowing the field more than anything. Let the other guy take the risks. You would miss on Brees and Cutler under some of the rules I set. But you would avoid Joey Harrington (a 55 percent college passer).

I really think that completion percentage in college is supremely important. In the pros, you have to be able to hit the spot. And yes - maybe your receivers or system stunk. But it's not like the windows will get larger in the NFL.

And yes, I acknowledge that under my system, Heath Shuler would appear to be utterly can't miss. What the **** are you gonna do?

[Oldfan]

I don't think OF is accurate about the idea that a team makes a QB so all the individual numbers are worthless.

In 2006, Tom Brady's TD/INT ratio was 24/12. In 2007, it was 50/8.

The difference? in 2006, the Patriots had the worst WR corps in the NFL. In 2007, they had the best. All the rest of his support system remained the same.

In 2009, Brady played the Jets twice. In one, his QB rating was about 50. In another, it was over 100. The difference? Belichik was unprepared for Ryan's overload blitzes in their first meeting. Brady was sacked only once, but he was throwing off his back foot even when the pressure didn't come. How can you grade QBs when they experience different degrees of pressure?

In the Football Outsiders Almanac which you recommended to me, Brady's numbers in the shotgun for 2007 were 25% better than when he went under center. Mike Martz rarely puts Cutler in the shotgun. Brady's in the gun 55% of the time. How do you fairly compare the completion percentages?

[Lombardi's_kid_brother]

In 2006, Tom Brady's TD/INT ratio was 24/12. In 2007, it was 50/8.

The difference? in 2006, the Patriots had the worst WR corps in the NFL. In 2007, they had the best. All the rest of his support system remained the same.

In 2009, Brady played the Jets twice. In one, his QB rating was about 50. In another, it was over 100. The difference? Belichik was unprepared for Ryan's overload blitzes in their first meeting. Brady was sacked only once, but he was throwing off his back foot even when the pressure didn't come. How can you grade QBs when they experience different degrees of pressure?

In the Football Outsiders Almanac which you recommended to me, Brady's numbers in the shotgun for 2007 were 25% better than when he went under center. Mike Martz rarely puts Cutler in the shotgun. Brady's in the gun 55% of the time. How do you fairly compare the completion percentages?

In 1990, Tony Gwynn hit .309. In 1987, he hit .370. Statistics are not going to remain the same from year to year. Tom Brady's 2007 is the outlier of all outliers and probably does more to prove that Randy Moss is the greatest receiver in the history of the NFL than anything else. This year is way outside the mean as well.

But look at Brady from 2002 to 2009. He is always between 23 and 28 TDs.

I mean...yea...guys sometimes have years that are outside all that is predictable. Brady Anderson hit 50 home runs once. Albert Connel was a 1,000 year receiver.

I never claimed to be able to understand EVERYTHING by individual stats. The 2007 Patriots had 10 ungodly weeks. It's one of the few times that the entire NFL has been caught flat-footed by something and seemed (collectively) unable to adjust.

The other statistical year that makes no ****ing sense is Randall Cunningham's '98.

Of course, Randy Moss was a factor there as well. Those are also quite possibly the only two seasons in Randy Moss' career where he gave a ****. If Randy Moss had Jerry Rice's work ethic, it's possible that he would have been a Wayne Gretzky type figure. Someone who puts up numbers that are so ridiculous in historical context that you can't really even make sense of them. (Gretzky has more assists than any other player has points...which is just stupid. No great player should be THAT much better than his peers. Numerically, Gretzky is 50 percent better than the second greatest player ever. It would be as if Barry Bonds hit 1100 home runs).

You are obsessed with Brady by the way and use him to bash anyone with any notion about football. Nothing about Brady's career makes sense. In the 2000s, the greatest QB of all time should not slip to the sixth round. A QB should not suddenly double his TD production in one year. A QB should not be able to succeed with that hair cut. There is no model that anyone could build into which Tom Brady would fit. Likewise, if you were to create a model for the greatest physicist ever, Einstein would not fit the mold.

It's impossible to predict Tom Brady. He should not exist. But he does. And that's one of things that makes sports wonderful. You can predict 995 out of 1000 things...but then...Tom Brady. A skinny dude who shared time at Michigan with a third baseman.

You know who he should be compared to? John Daley. There is no way anyone could predict that an alcoholic fat guy who should be winning long-driving competitions would win two major titles, including the goddamn British. Seriously, Daley won a major playing links golf. But if you were trying to figure out how to become a golf pro, I would not recommend trying to model myself after him.

[Oldfan]

...It's impossible to predict Tom Brady. He should not exist. But he does. And that's one of things that makes sports wonderful. You can predict 995 out of 1000 things...but then...Tom Brady. A skinny dude who shared time at Michigan with a third baseman.

If I picked out a QB who played for Mike Leach at Texas Tech and produced outrageous stats, you'd brush it off. "He's a system quarterback, you'd say." But, you look at Brady's outrageous stats and search for mystical explanations.

The explanation is simple: He's a system quarterback!

ALL QBs are system quarterbacks, from the pee wee leagues to the NFL -- their performance is dependent upon their support as well as their own individual effort. Teamwork is a two way street. And, Football is the ultimate team sport.

[K.O. Johnny]

You aren't going to find a magic formula. When dealing with Russell' date=' the number that most mattered was his weight. There are some things that you can only find out on deep background - is the guy on drugs, does he have 14 kids, is he a fat ass?

Think of it this way. If you were interviewing someone for a job and they were top of their class at Harvard...had great references....were good looking and well dressed...you would hire them. If a year later, you found out the guy was embezzling to pay for a gambling problem, you wouldn't change the criteria you use to hire. "From now on I only hire ugly people from community college!"

[/quote']

This

I guess i started this thread because of all the folks creaming their pants over Luck, like a bunch of pre-adolescent school girls over Justin Bieber. It truly is becoming hysterical - people are really advocating doing a Ditka or some such thing.

I'd take the Harvard guy and the sixth rounder and the free agent with good stats and a good attitude - and this is the crucial part, GREAT NOT GOOD, BUT GREAT COACHING IS WHAT MAKES THE DIFFERENCE. Patience, experience and adjusting the scheme to fit the system to the QB. That's what made Gibbs so great, different QB's at different stages in their careers coming away with league or SB MVP honors.

The QB needs basics, like accuracy, athleticism and clutch but I think if you look back through the NFL you could perhaps fit a Michael Vick into Randall Cunningham's scheme or Marino into Jurgensen's system. You get my drift. So let's examine numbers, and do interviews and wonderlics and inkblot tests and all the due diligence we can, but let's not sell the farm.

[Lombardi's_kid_brother]

If I picked out a QB who played for Mike Leach at Texas Tech and produced outrageous stats, you'd brush it off. "He's a system quarterback, you'd say." But, you look at Brady's outrageous stats and search for mystical explanations.

The explanation is simple: He's a system quarterback!

ALL QBs are system quarterbacks, from the pee wee leagues to the NFL -- their performance is dependent upon their support as well as their own individual effort. Teamwork is a two way street. And, Football is the ultimate team sport.

I don't think I would say that about ANY Mike Leach QB. It was pretty obvious last year that Sam Bradford was different than the previous OU quarterbacks who put up ungodly numbers in that system but obviously weren't going to be pro QBs. The first thing to look for is an NFL body. (Brady had the height, but was so skinny as a college player). Then you start breaking down the numbers. The dude from Boise State probably won't make it in the NFL because he is built like Brandon Banks. If he was built like Rex Grossman he would be intriguing.

And, yes, football is a team sport. But there was a reason that a half-crippled Joe Montana was still connecting on 60 percent of his passes in Kansas City at age 63.

Your entire philosophy strikes me as a cul-de-sac. If a team doesn't have a good system, you shouldn't draft a QB. But you can't know if you have a good system until you have a QB that matches it. I can't determine if you think the Skins should NEVER draft a QB or if they should draft 10 a year until they find the right match.

---------- Post added January-5th-2011 at 09:01 PM ----------

This

I guess i started this thread because of all the folks creaming their pants over Luck, like a bunch of pre-adolescent school girls over Justin Bieber. It truly is becoming hysterical - people are really advocating doing a Ditka or some such thing.

I'd take the Harvard guy and the sixth rounder and the free agent with good stats and a good attitude - and this is the crucial part, GREAT NOT GOOD, BUT GREAT COACHING IS WHAT MAKES THE DIFFERENCE. Patience, experience and adjusting the scheme to fit the system to the QB. That's what made Gibbs so great, different QB's at different stages in their careers coming away with league or SB MVP honors.

The QB needs basics, like accuracy, athleticism and clutch but I think if you look back through the NFL you could perhaps fit a Michael Vick into Randall Cunningham's scheme or Marino into Jurgensen's system. You get my drift. So let's examine numbers, and do interviews and wonderlics and inkblot tests and all the due diligence we can, but let's not sell the farm.

Oh, I think Luck is as close to a sure thing as you are ever going to find. If you were going to set up a checklist of everything a college QB needs, he would hit every single one of them.

NFL body? He's 6'4, 235. That's about perfect.

College numbers? 64 percent completion percentage. Nearly 4:1 TD/INT ratio. Again, about perfect.

College experience? 25 games. You would probably like 5 or 6 more games. But, that's a B grade at worst.

Intelligence? He went to Stanford. His dad was a finalist for a Rhodes Scholar and a law school grad. I'm guessing he's pretty damned bright.

Skeletons in the closet? I actually know a family in Houston that is friends with his family. And I know people that know his dad because of WVU. The Lucks seem annoyingly perfect. It would actually make me happy if he turned out to be running a white slavery ring.

Clutchiness? I put no faith in this, but he's won at Notre Dame, in the Orange Bowl, and at USC. So...ok.

In all honesty, he's probably the best pro prospect on paper since Manning.

I mean, maybe he turns into Tim Couch, but it would really be remarkable if he wasn't a regular Pro Bowler.

[Oldfan]

...Your entire philosophy strikes me as a cul-de-sac. If a team doesn't have a good system' date=' you shouldn't draft a QB. But you can't know if you have a good system until you have a QB that matches it. I can't determine if you think the Skins should NEVER draft a QB or if they should draft 10 a year until they find the right match.[/quote']You start by designing a scheme. One of your choices is whether to design it for a non-athletic pocket passer like Tom Brady or Peyton Manning or whether you want to design it for an athletic QB like Jay Cutler or John Elway. Another option is to go WCO with a mobile QB who doesn't have a grade A arm (Montana).

If you go for the pocket passer, or the WCO-type you might find the guy you need from the third-round down. If you go for the QB athlete, you will probably be picking number one in the draft unless he played for Vanderbilt.

The advantage Shanahan had with Cutler is that he could move him around and make the O-line's job easier in protection. So, Belichik's scheme needs a better O-line than Shanaahan's, but Mike needs a more talented QB.

I'd go with Belichik's scheme because picking that QB #1 is high risk. Moreover, if Brady goes down, you can plug in a Matt Cassel and still get decent production, but if Shanahan's #1 athletic guy goes down, it's a long drop to #2.

[K.O. Johnny]

I don't think I would say that about ANY Mike Leach QB. It was pretty obvious last year that Sam Bradford was different than the previous OU quarterbacks who put up ungodly numbers in that system but obviously weren't going to be pro QBs. The first thing to look for is an NFL body. (Brady had the height' date=' but was so skinny as a college player). Then you start breaking down the numbers. The dude from Boise State probably won't make it in the NFL because he is built like Brandon Banks. If he was built like Rex Grossman he would be intriguing.

And, yes, football is a team sport. But there was a reason that a half-crippled Joe Montana was still connecting on 60 percent of his passes in Kansas City at age 63.

Your entire philosophy strikes me as a cul-de-sac. If a team doesn't have a good system, you shouldn't draft a QB. But you can't know if you have a good system until you have a QB that matches it. I can't determine if you think the Skins should NEVER draft a QB or if they should draft 10 a year until they find the right match.

---------- Post added January-5th-2011 at 09:01 PM ----------

Oh, I think Luck is as close to a sure thing as you are ever going to find. If you were going to set up a checklist of everything a college QB needs, he would hit every single one of them.

NFL body? He's 6'4, 235. That's about perfect.

College numbers? 64 percent completion percentage. Nearly 4:1 TD/INT ratio. Again, about perfect.

College experience? 25 games. You would probably like 5 or 6 more games. But, that's a B grade at worst.

Intelligence? He went to Stanford. His dad was a finalist for a Rhodes Scholar and a law school grad. I'm guessing he's pretty damned bright.

Skeletons in the closet? I actually know a family in Houston that is friends with his family. And I know people that know his dad because of WVU. The Lucks seem annoyingly perfect. It would actually make me happy if he turned out to be running a white slavery ring.

Clutchiness? I put no faith in this, but he's won at Notre Dame, in the Orange Bowl, and at USC. So...ok.

In all honesty, he's probably the best pro prospect on paper since Manning.

I mean, maybe he turns into Tim Couch, but it would really be remarkable if he wasn't a regular Pro Bowler.

Which Manning are you talking about? I'm not kidding either - their stats are remarkably similar

Peyton

Year

Attempts Comp. Comp % Yards INT TD's

1994 144 89 61.8 1141 6 11

1995 380 244 64.2 2954 4 22

1996 380 243 63.9 3287 12 20

1997 477 287 60.37 3819 11 37

Total 1354 851 62.85 11201 33 90*

Eli

Season Team GP Rating Att Comp Pct Yds TD Int Att Yds TD

2000 Ole Miss Rebels 6 117.4 53 28 52.8 337 3 2 7 4 0

2001 Ole Miss Rebels 11 144.8 408 259 63.5 2,948 31 9 31 9 0

2002 Ole Miss Rebels 13 125.6 481 279 58.0 3,401 21 15 39 -120 2

2003 Ole Miss Rebels 13 148.1 441 275 62.4 3,600 29 10 48 -28 3

Career Ole Miss Rebels 43 138.1 1383 841 60.8 10,286 84 36 125 -135 5

See what I mean, and the same family, genes, upbringing. Even college didn't seperate them. So the big question is, is it scheme/system and coaching that has made the difference in their pro careers. I really think it's a matter of coaching.

[Lombardi's_kid_brother]

Which Manning are you talking about? I'm not kidding either - their stats are remarkably similar

Peyton

Year

Attempts Comp. Comp % Yards INT TD's

1994 144 89 61.8 1141 6 11

1995 380 244 64.2 2954 4 22

1996 380 243 63.9 3287 12 20

1997 477 287 60.37 3819 11 37

Total 1354 851 62.85 11201 33 90*

Eli

Season Team GP Rating Att Comp Pct Yds TD Int Att Yds TD

2000 Ole Miss Rebels 6 117.4 53 28 52.8 337 3 2 7 4 0

2001 Ole Miss Rebels 11 144.8 408 259 63.5 2,948 31 9 31 9 0

2002 Ole Miss Rebels 13 125.6 481 279 58.0 3,401 21 15 39 -120 2

2003 Ole Miss Rebels 13 148.1 441 275 62.4 3,600 29 10 48 -28 3

Career Ole Miss Rebels 43 138.1 1383 841 60.8 10,286 84 36 125 -135 5

See what I mean, and the same family, genes, upbringing. Even college didn't seperate them. So the big question is, is it scheme/system and coaching that has made the difference in their pro careers. I really think it's a matter of coaching.

I think Peyton is smarter and likes football more. I think Peyton might end up like Favre and be trying to play when he is 42. I think Eli is done in his mid 30s.

---------- Post added January-5th-2011 at 10:26 PM ----------

You start by designing a scheme. One of your choices is whether to design it for a non-athletic pocket passer like Tom Brady or Peyton Manning or whether you want to design it for an athletic QB like Jay Cutler or John Elway. Another option is to go WCO with a mobile QB who doesn't have a grade A arm (Montana).

If you go for the pocket passer, or the WCO-type you might find the guy you need from the third-round down. If you go for the QB athlete, you will probably be picking number one in the draft unless he played for Vanderbilt.

The advantage Shanahan had with Cutler is that he could move him around and make the O-line's job easier in protection. So, Belichik's scheme needs a better O-line than Shanaahan's, but Mike needs a more talented QB.

I'd go with Belichik's scheme because picking that QB #1 is high risk. Moreover, if Brady goes down, you can plug in a Matt Cassel and still get decent production, but if Shanahan's #1 athletic guy goes down, it's a long drop to #2.

I don't know if the word I would use to describe Manning is "unathletic." Does anyone run a pure drop back/pocket passer scheme any longer? Gibbs second run was really the last gasp of that style of play, wasn't it?

[Mahons21]

See what I mean, and the same family, genes, upbringing. Even college didn't seperate them. So the big question is, is it scheme/system and coaching that has made the difference in their pro careers. I really think it's a matter of coaching.

I strongly disagree. Peyton is known for his desire to be more prepared than anyone else on the field, its what drives him to spend countless hours studying tape. What Peyton has cannot be coached, and we'll be lucky to see another QB as talented as him.

---------- Post added January-5th-2011 at 09:50 PM ----------

I don't know if the word I would use to describe Manning is "unathletic." Does anyone run a pure drop back/pocket passer scheme any longer? Gibbs second run was really the last gasp of that style of play' date=' wasn't it?[/quote']

The Giants run a similar scheme in my opinion. It seems to me that Indy uses the stretch to setup play-action, while the Giants use much more of a power running scheme to set up the PA.

[Lombardi's_kid_brother]

I strongly disagree. Peyton is known for his desire to be more prepared than anyone else on the field, its what drives him to spend countless hours studying tape. What Peyton has cannot be coached, and we'll be lucky to see another QB as talented as him.

---------- Post added January-5th-2011 at 09:50 PM ----------

The Giants run a similar scheme in my opinion. It seems to me that Indy uses the stretch to setup play-action, while the Giants use much more of a power running scheme to set up the PA.

The Giants run that Gilbride read and react/option route thing. I'm sure OF can explain how its part of the Coryell legacy since every offense is apparently part of the Coryell legacy.

My understanding of what Gilbride does is that every receiver has several choices for their route based on the D. The QB and the receiver have to make the same read. It's similar to the Colts except (I think) the Colts have several options when they come to the line and then Peyton makes the choice. I do think Gilbride is a more vertical oriented passing game where a lot of the Colts' attack seems more horizontal with the way they use backs and tight ends.

Here's the thing: I think OF's key mistake is that he locks these teams into boxes. The Patriots switched from a vertical to a horizontal attack in mid-season. It's not like the 70s or 80s where everyone is locked into a scheme. Every playbook now seems to have pretty much everything in it. The Steelers' would appear to be run/set up the play action/and throw deep to Wallace team. Except, I think they run more passes to receivers behind the line of scrimmage than anyone.

The big change in offenses over the last 10 years has been the formations and personnel groupings. I bet Kyle Shanahan uses more formations in a half than Gibbs used in a month in the 80s - and Gibbs was more innovative than anyone. The Packers broke out a T-Formation last month for God's sake. 15 years ago, you never saw things like TEs in the side car in the shotgun. Now, every team does it twice a half.

So, I don't get terribly caught up in this "scheme" thing. Al Saunders' 700 page playbook was not built around one concept or philosophy.

[zoony]

great thread, very good reading. best discussion on the web still takes place here

[Oldfan]

I don't know if the word I would use to describe Manning is "unathletic." Does anyone run a pure drop back/pocket passer scheme any longer? Gibbs second run was really the last gasp of that style of play' date=' wasn't it?[/quote']I use non-athlete and athlete here to describe QBs, not to compare them to the general population.

Peyton rarely leaves the pocket by design. Most of this throws are from the shotgun or five-step drops. It's the same with Brady. Defenses try to get him to move because he doesn't throw nearly as well on the move.

---------- Post added January-5th-2011 at 11:24 PM ----------

...Here's the thing: I think OF's key mistake is that he locks these teams into boxes...

Not sure where you got that idea. I seldom talk about scheme labels.

When we talk about QBs, scheme fit is a key issue. People would laugh if Mike Shanahan brought Peyton in and had him running his signature play, the play-action rollout off the zone stretch.

A couple of weeks ago, when talking about McNabb and mechanics, Mike pointed out that it's much more difficult to throw accurately on the move. He pointed out that Peyton doesn't have that problem because his feet are always under him in the pocket.

It's just common sense. The better athletes throw better on the move. Jay Cutler isn't a good fit for Mike Martz's scheme. He has a QB who moves extremely well and he's using him primarily as a pocket passer. Cutler was sacked only 11 times in Denver 2008 because Mike moved him away from the rush.

[Lombardi's_kid_brother]

Cutler seems to be having the same year he always has. He was a 60 percent passer who threw way too many interceptions in Denver. He is now a 60 percent passer who throws way too many interceptions. His numbers are pretty steady throughout his career.

And I'm not sure any QB is a good fit for Martz' offensive philosophy, mainly because they get the hell knocked out of them. He caught lightning in a bottle on a team where everything briefly lined up perfectly. An offense where you have to get the ball out in a blink met a QB with a lightning release who happened to have the greatest safety valve in NFL history - all while playing in laboratory conditions. That may be one of the rare times when "scheme" and players aligned perfectly. And, of course, it was completely by accident.

The problem with Martz has never been an inability to match up QB with scheme. It's the fact that he never cared about protecting his QB. I really think Martz got extremely lucky over a three year period or he would be considered Mouse Davis. It's telling that NO ONE has ever tried to copy what he does. This is a leauge where teams are still trying to get mileage out of the wildcat.

(I should also point out that Warner revived his career in Ken Wisenhunt's caveman offense).

Shanahan's signature play may also be a chicken and an egg thing. He spent half his career with supremely athletic QBs who were in place before he ever showed up - Young and Elway. Was he running a ton of bootlegs with Jay Schroeder back with the Raiders?

[Oldfan]

Cutler seems to be having the same year he always has. He was a 60 percent passer who threw way too many interceptions in Denver. He is now a 60 percent passer who throws way too many interceptions. His numbers are pretty steady throughout his career.

No they are not. His stats in 2008, taken as a whole, are much superior to anything he's put up in Chicago.

When you look at a completion percentage, do you adjust for any reason? For example, is Peyton's 66.3% completion pct., accomplished behind the #1 ranked line for protection still superior to Cutler's 60.4% accomplished behind the #32 ranked line?

[Darth Tater]

Shanahan's signature play may also be a chicken and an egg thing. He spent half his career with supremely athletic QBs who were in place before he ever showed up - Young and Elway. Was he running a ton of bootlegs with Jay Schroeder back with the Raiders?

Well, Shanahan's signature action play is also the signature action play of the veer offense which is what he ran as a QB in HS and as an offensive assistant in college. The fact is, the whole run game he uses is a modified veer attack.

[LD0506]

great thread, very good reading. best discussion on the web still takes place here

Have to second this, this is a phenomenal thread with some of the best back-n-forth seen here in a while.

What I seem to get though is that, no matter how you want to look at it or what stats you choose, it is essentially impossible to "know" anything about any QB until their career is over, which might be interesting but doesn't help us with any prognostications. If the exceptions pro and con seriously skew the results this badly than analysis is worthless, or at least this form of analysis is. I may be a naif when it comes to football but I do know a thing or two about analytical systems, and it just seems as though some essential element is being overlooked that once factored in would make a lot more sense of it all. I know that OF is correct in that trying to look at any QBs stats in a vacuum is ineffective, but I will not deign to even suggest an alternative. This is imp in the machine with any attempt to interpret a macro-process, you cannot accurately assess individual elements of a wider-synergistic result. There has to be a wider view of the entirety with enough precision to deconstrust the individual players' contributions. The closest example I can think of is climatologists trying to understand global weather patterns.

Great thread, I hope the players pick up where they left off today and give us even more.

[Lombardi's_kid_brother]

No they are not. His stats in 2008, taken as a whole, are much superior to anything he's put up in Chicago.

When you look at a completion percentage, do you adjust for any reason? For example, is Peyton's 66.3% completion pct., accomplished behind the #1 ranked line for protection still superior to Cutler's 60.4% accomplished behind the #32 ranked line?

No, I don't. Because I don't think pass rush impacts completion percentage that dramatically. I think it impacts ypa and maybe interceptions. But offensive coordinators adjust the plays they call based on the quality of everyone else on the team. Bradford is a perfect example. His best receiver is a slow white dude. His line stinks. Yet, he managed to hit on 60 percent of his passes as a rookie. The telling thing is that his ypa is a ridiculous 5.95.

Roethlisberger is another example. His completion percentage appears to go UP as his line gets worse. His best season ever was 2009, and that line was truly abysmal at pass protection.

The bottom line is that even behind awful lines, most QBs only get hit on maybe one in five of their dropbacks. There seemed to be this idea with Campbell among his most ardent supporters that every single time he dropped back he was in full retreat from 5 DeMarcus Wares.

I mean, if I adjusted the numbers as you suggest, I would have to up Ben Roethlisberger's 2009 to a comical degree. He hit on 66 percent of his passes while getting sacked 50 times. Manning hit on 68 percent while taking 10 sacks. Does that mean that Roethlisberger is actually 5 times the passer that Manning is? I wouldn't even know how to adjust that.

As I said, I think it's a three or four part balancing test, and as a lawyer, I'm really good at thinking in terms of balancing tests. Completion percentage tells you if the QB can hit what he is aiming at. I think it's a fairly independent stat. TD/Int ratio tells you if he makes good or bad decisions. That is somewhat independent as well, though I think it has a much higer "luck" factor. (For instance, if you throw 5 screen passes to Adrian Peterson that go for touchdowns, that makes this stat look good though you didn't necessarily have a ton to do with it). YPA tells you how impactful your throws are. I think it is the stat that is most dependent on other players. (In Roethlisberger's case, he has Mike Wallace. In Vick's case, he has DeSean Jackson. In Bradford's case, he has a slow white guy). Attempts would probably be the last thing I glanced at. To me it's probably 50 percent, 30 percent, and 20 percent in terms of their importance. Your mileage may vary.

One of the things to consider is that, yes, somewhat like Campbell might be on a bad team that gives up too many sacks and has no playmakers. That would drive down his relevant stats. At the same time, he's probably playing from way behind a lot. Which means he's facing a ton of deep zones which allow for easy dumpoffs. Larry Centers regularly put up 80-catch seasons by being the safety valve in garbage time for really bad teams. In his 100 catch season, he caught soemething like 25 passes in 4th Quarters where his team was down by ten or more.

Baseball certainly provides for more accurate stats, but football provides unique opportunities to pad stats that other sports like. In the 9th inning of a 6-run victory, Albert Pujols is still facing a major league quality pitcher who still needs to get him out. In 20 point NBA games, the benches empty. NFL teams don't really have enough backups to clear the bench the way the NBA or NHL can.

[K.O. Johnny]

have to second this, this is a phenomenal thread with some of the best back-n-forth seen here in a while.

What i seem to get though is that, no matter how you want to look at it or what stats you choose, it is essentially impossible to "know" anything about any qb until their career is over, which might be interesting but doesn't help us with any prognostications. If the exceptions pro and con seriously skew the results this badly than analysis is worthless, or at least this form of analysis is. I may be a naif when it comes to football but i do know a thing or two about analytical systems, and it just seems as though some essential element is being overlooked that once factored in would make a lot more sense of it all. I know that of is correct in that trying to look at any qbs stats in a vacuum is ineffective, but i will not deign to even suggest an alternative. This is imp in the machine with any attempt to interpret a macro-process, you cannot accurately assess individual elements of a wider-synergistic result. There has to be a wider view of the entirety with enough precision to deconstrust the individual players' contributions. The closest example i can think of is climatologists trying to understand global weather patterns.

Great thread, i hope the players pick up where they left off today and give us even more.

wtf

To hell with it, let's just run it down their throats - up the middle and a cloud of dust

No ld0506, just kidding, thanks too for your contribution - and i think perhaps the only conclusion we can definitely draw (if i'm understanding you correctly) is that QB play/passing/running/offense/winning and losing this game is more than the sum of it's parts and to try and deconstruct it will ultimately wreck it.

It's why I love this game more than any other team sport in the world - i just wish they could go back to beating the cr@p out of QBs a bit more :pfft:

[Lombardi's_kid_brother]

One quick thought to explain my thinking on completion percentage. It's not a perfect stat obviously. And neither is batting average, since it only tells one part of a baseball players' strengths. But as far as I can tell, no sabremetricians have tried to improve on batting average by factoring in the quality of the pitchers faced. (They do adjust for leagues and ballparks however). There is an assumption that - over time - all those factors even out. That's the point I'm making with completion percentage. It's the one football stat that is built on a healthy sample size. So, if Johnny Utah drops back 500 times, the fact that his line can't pick up a double A gap blitz might impact 10 percent of the stats, but the fact that he plays in a division with terrible corners might balance that out.

So, Bengals running backs might have skewed stats because they face the Steelers and Ravens four times a year in their career. But they generally play a last-placed schedule, which makes up for that.

[LD0506]

wtf

To hell with it, let's just run it down their throats - up the middle and a cloud of dust

No ld0506, just kidding, thanks too for your contribution - and i think perhaps the only conclusion we can definitely draw (if i'm understanding you correctly) is that QB play/passing/running/offense/winning and losing this game is more than the sum of it's parts and to try and deconstruct it will ultimately wreck it.

It's why I love this game more than any other team sport in the world - i just wish they could go back to beating the cr@p out of QBs a bit more :pfft:

My point (whether expressed well or not) is that trying to reduce this all by typical statistical analysis has come up short, woefully short in some cases. You really can't just throw out Tom Brady because he screws up the numbers for example. I hope that OF and LKB can chew their way through this all to some GUT of QB drafting that might make sense of it all.

And I agree, that is one of the things that makes football stand out above the rest, it is the complete team effort that defines the game.

[RiggosMohawk]

(I should also point out that Warner revived his career in Ken Wisenhunt's caveman offense).

yuck... unfortunately you brought up evidence that helps prove OF's tangibles argument, and I dislike agreeing with him on principle.

Warner has the top 3 tangibles in my opinion for a QB. They can be observed/measured independent of scheme and during live game action:

Decision making

Release speed

Accuracy

No matter what his scheme has called for, these 3 qualities have let him be successful (except the Giants who didnt want to block for him). I think they are the top 3 qualities you need in the NFL where the QBs margin of error is so much less than college, and where the QBs mistakes are much more damaging to winning %.

Also agree that this is a great thread