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

[Lombardi's_kid_brother]

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

Those probably are the three most important tangible qualities. The question is, how do you measure them short of having Gil Brandt say, "He's got a release like Joe Gilliam!"

Decision making? I think TD/Int does that to some degree.

Accuracy? I think Completion percentage does that.

Release? I'm not sure how you measure that as an isolated number. It's probably embedded in everything. (A slow release certainly hurts the other two since it allows DBs to react).

---------- Post added January-6th-2011 at 11:21 AM ----------

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.

I'm not suggesting throwing Tom Brady out. I'm saying you treat him as an outlier - one of those things that make life interesting.

If your investments made a 5 to 7 percent return every year for 30 years - except 1997 because you got into tech stocks at the right time and got out at the right time so you made 26 percent - you wouldn't try to build your entire investment model around 1997. Crazy **** happened that year.

The way I read OF is that he wants to start with Tom Brady's 2007 and Randall Cunningham's 1998 and try to account for that first. I want to build from every year but 2007 and then try to figure it in last. When I get there, I write down "Randy Moss" and move on with my life.

[Oldfan]

LKB: 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...

That's true, but the point you seem to be ignoring is that offensive coordinators affect the completion percentage for all QBs. So, the completion percentage reflects the QBs accuracy, both the quality and philosophy of the offensive coordinator, the quality of the receivers and other factors. The completion percentage is a useless number.

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.

Sam Bradford is a more accurate passer than Peyton Manning. I arrived at that opinion by watching them throw a football. Your useless completion percentage stats won't support or deny that conclusion because the two players operate in completely different environments. Your stat has no more value in comparing QBs than the final times of a race between two sprinters in which one has to carry a 30-pound weight on his back.

[K.O. Johnny]

Okay then, let's look at this from the wrong end of the telescope - we're trying to separate what makes great QBs and we're getting all confused with the misdirection of schemes, running game, o-line, opposition etc, so let's take these out of the equation.

Let's try and single out QBs who were better than they should have been and whose stats are not that outstanding but we would still take them in a heartbeat. I hope I'm not making matters worse!

Elway - career rating 79.9 (nothing that special) and that figure is 'massaged' because his 4 best years (stats wise) had Terrell Davis at RB.

BUT he took the pre Terrell Davis Broncos to 3 SBs with a weedy running game, an undersized o-line, barely above average receivers and a bend-not-break rather than hard-nosed defense to 3 super bowls and kept the Broncos competitive for over a decade. I really think Elway was/is up there. What did he have though, I can't work it out? Clutch and mobility, arm strength? Definitely, but there's something that we're missing.

Can people come up with other better than their stats QBs and see what they had in common?

[LD0506]

I'm not suggesting throwing Tom Brady out. I'm saying you treat him as an outlier - one of those things that make life interesting.

If your investments made a 5 to 7 percent return every year for 30 years - except 1997 because you got into tech stocks at the right time and got out at the right time so you made 26 percent - you wouldn't try to build your entire investment model around 1997. Crazy **** happened that year.

The way I read OF is that he wants to start with Tom Brady's 2007 and Randall Cunningham's 1998 and try to account for that first. I want to build from every year but 2007 and then try to figure it in last. When I get there' date=' I write down "Randy Moss" and move on with my life.[/quote']

I get that, don't mind me, I am just kibitzing on the sidelines. You mavens have at it.

[Oldfan]

The way I read OF is that he wants to start with Tom Brady's 2007 and Randall Cunningham's 1998 and try to account for that first. I want to build from every year but 2007 and then try to figure it in last. When I get there' date=' I write down "Randy Moss" and move on with my life.[/quote']I use Tom Brady's 2007 and 2008 numbers to show how one supporting factor, the quality of the WR corps can make a huge difference in a QB's performance. I use Brady's numbers in the shotgun versus Brady's numbers under center to show how just one element of a scheme can influence a QB's numbers. I use Brady's numbers in two games against the Jets to show the effect of pressure on a QB's numbers.

I can take Cutler's numbers in Denver (2008) and his numbers in Chicago (2009) and Orton's numbers in Chicago (2008) and his numbers in Denver (2009) and, by comparing them, show any reasonable mind that the numbers are useless in grading QBs because they don't actually measure the QB in isolation.

[darrelgreenie]

Okay then, let's look at this from the wrong end of the telescope - we're trying to separate what makes great QBs and we're getting all confused with the misdirection of schemes, running game, o-line, opposition etc, so let's take these out of the equation.

Let's try and single out QBs who were better than they should have been and whose stats are not that outstanding but we would still take them in a heartbeat. I hope I'm not making matters worse!

Elway - career rating 79.9 (nothing that special) and that figure is 'massaged' because his 4 best years (stats wise) had Terrell Davis at RB.

BUT he took the pre Terrell Davis Broncos to 3 SBs with a weedy running game, an undersized o-line, barely above average receivers and a bend-not-break rather than hard-nosed defense to 3 super bowls and kept the Broncos competitive for over a decade. I really think Elway was/is up there. What did he have though, I can't work it out? Clutch and mobility, arm strength? Definitely, but there's something that we're missing.

Can people come up with other better than their stats QBs and see what they had in common?

I consider using Jason Campbell's 2009 production numbers as a QB 'mendozza' line.

Jason Campbell is a average to good QB who had very little help from any direction FO/coaching/playcaller changes, OL, running game, average WRs

yet was still able to put up decent stats therefore any good QB should at least be able to match that level of production.

For example if you view Campbell's 2009 production against Carson Palmer they had nearly identical stats but was in much better situation or Matt Cassell who was in a similar situation but had far worse stats.

*The Randy Moss effect is an easy to identify indicator that the quality of the talent around a QB effects there production.

Tom Brady, Randall Cunningham, Brad Johnson, Dante Culpepper all had the one of their most productive seasons of their careers w/ Randy Moss as their WR.

[Oldfan]

...I really think Elway was/is up there. What did he have though, I can't work it out? Clutch and mobility, arm strength? Definitely, but there's something that we're missing.

John Elway, Jay Cutler and Archie Manning are the three best QB athletes I've ever seen.

Archie played for the hapless Saints and had the worst won/lost percentage in NFL history for any QB with 100 or more starts. He isn't in the HOF because his team stunk. He was much more of an individual threat to defenses than either of his sons.

Athletic QBs can do more with less supporting talent than other QBs. By using Cutler's ability to move and throw on the run, Mike Shanahan could make his O-line look better than they really were (fewest sacks in the NFL in 2008 with eleven). Using Cutler's arm strength, Mike stretched the field vertically.

In the shoes of an offensive coordinator, I ask: What tools does this QB have that I can use to win games? That's how I grade QBs.

[PF Chang]

LKB is clearly a brilliant guy and I agree with his posts 99% of the time. That being said, OF is killing it here. Completion percentage is one of the furthest things from being an independent variable in football. It's like saying that Jamaal Charles is the best running back in the league because he had the best YPA. Of course it's an indicator of QB quality but the benchmarking games seem pointless. Why even have the NFL combine if you'd never draft a QB with under 60% completion rate and under 6'1"? If it's all statistical and we can evaluate a QB purely through team-dependent statistics then there's really no need for the combine.

Where I disagree with OF is the relative importance of a QB's tools. I think he overrates arm strength and underrates intelligence/decision making, as we debated in the Manning thread (I think Manning is the best QB in the league). But that's a disagreement about the specifics of the evaluation and not the principles.

[Lombardi's_kid_brother]

I use Tom Brady's 2007 and 2008 numbers to show how one supporting factor, the quality of the WR corps can make a huge difference in a QB's performance. I use Brady's numbers in the shotgun versus Brady's numbers under center to show how just one element of a scheme can influence a QB's numbers. I use Brady's numbers in two games against the Jets to show the effect of pressure on a QB's numbers.

I can take Cutler's numbers in Denver (2008) and his numbers in Chicago (2009) and Orton's numbers in Chicago (2008) and his numbers in Denver (2009) and, by comparing them, show any reasonable mind that the numbers are useless because they don't actually measure the QB in isolation.

You get way too bogged down in small sample sizes.

Want to explain Cutler's 2009 numbers?

It's easy: 3 games.

Green Bay: 17-36; 4 picks

San Fran: 29-52; picks

Baltimore: 10-27 94 yards 3 picks

He had 3 utterly abysmal games in 2009.

It's not a great stat, but Cutler's QB rating over his career is:

88.5

88.1

86.0

76.8

86.3

If he was any more consistent, he would be a metronome. 2009 had three ridiculousy, embarrassingly awful games.

Cutler has a tendency to **** the bed. Martz' offense encourages bed-****ting. It's not an ideal matchup. I don't think Shanahan's scheme necessarily limits turnovers. I think Shanahan limits turnovers. If Cutler threw 4 picks under Shanahan, Mike would a) throw away the plays that were causing this, bench Cutler, and c) have him rolled into a rug and thrown off his yacht. Martz keeps plugging away while Lovie Smith - I dunno - listens to a smooth jazz station on his headset. (Ridiculous turnovers are a Lovie Smith trademark).

Generally speaking, QBs have tendencies towards turnovers and coaches can corral or exacerbate that. Cutler always threw more picks than the average bear. (That's a pun!) Martz loves turning the ball over. Put them together...and duck!

Question: Do you see any difference between "Scheme" and "Coaching?"

By that I mean, do you think Mike Martz-coached teams turn the ball over at an ungodly rate because of Mike Martz' scheme or because of Mike Martz?

It's not a great comparison because RB and QB are so different, but I think of Tiki Barber. He stopped fumbling in part because his coach made him stop fumbling. If he was playing for Martz, I assume that he would have put up more yards and kept fumbling. Because Martz doesn't seem to care about turnovers.

It's a difference between scheme and philosophy.

Back to your godhead, Belichick. I don't know what his offensive "scheme" is. One year, it's run Corey Dillon until his heart explodes. One year, it's throw 6 post patterns to moss a game. This year, it's you can never have too many pick plays to the tight end. He appears to be the only coach in football who has ever solved Dick Lebeau. Everytime he plays the Steelers, he runs 65 pass plays versus 8 runs and scores 35 points. When he started in New England, his scheme was "Let's pretend that Drew Bledsoe is Dan Marino and hope for the best."

And let's ignore the fact for a moment that Belichik is actually a defensive coach.

Basically, what New England does is whatever best suits their talents. I don't know how you can watch them this year and say anything other than they are the most adaptable team in history. They dumped the best deep threat in history, built their offense around rookie tight ends, and have been dinking and dunking everyone to death. I don't think I've seen anything like it.

And it also makes my "New England kinda stinks" posts from early in the year look really stupid.

---------- Post added January-6th-2011 at 12:49 PM ----------

LKB is clearly a brilliant guy and I agree with his posts 99% of the time. That being said, OF is killing it here. Completion percentage is one of the furthest things from being an independent variable in football. It's like saying that Jamaal Charles is the best running back in the league because he had the best YPA. Of course it's an indicator of QB quality but the benchmarking games seem pointless. Why even have the NFL combine if you'd never draft a QB with under 60% completion rate and under 6'1"? If it's all statistical and we can evaluate a QB purely through team-dependent statistics then there's really no need for the combine.

That's because you aren't fully grasping what I am trying to do.

The original topic is "How should we assess QBs?" The idea being, how do we assess QBs as they come out of college.

I've said repeatedly that there is no perfect formula. All I'm trying to do is create cut-off points.

Why am I obsessed with 60 percent?

Because of Joey Harrington, Akili Smith, Ryan Leaf, Chad Henne, Rex Grossman's senior year, Kyle Boller, and Cade McNown. In studying those guys, the one thing that truly stands out is that none of them could hit the bullseye in college. They were all sub 60 percent passers.

In researching the QBs that made it, all of them - except Cutler - were better than 60 percent passers.

The bottom line is if you did not complete 60 percent of your passes in college, there is 98 percent chance that you are not going to be good in the pros.

But do not confuse a 98 percent of NOT making it with a 98 percent chance of making it. I've acknowledged that it's not the only thing that matters. And I would still have drafted David Carr and Tim Couch and few other busts. But it eliminates a lot of bad eggs and only one good egg (and most of the people on this site think Cutler sucks anyway). All I'm doing is eliminating the players who are obviously going to suck. I still have to be intelligent within the smaller group I've created.

Why 6'1? Because the only stud QB in the league under 6'1 is Drew Brees. So, if you are under 6'1, there is a 98 percent chance that you are NOT making it. Do not confuse a 98 percent chance of NOT making it with a 98 percent chance of making it. All I'm doing is eliminating the players who are obviously going to suck. I still have to be intelligent within the smaller group I've created.

So, let's move onto IQ. Is Jason Campbell dumb? I don't know. But if I evaluate him and he is dumb, I'm not taking him. The only low Wonderlic score that has ever made it is apparently McNabb. So, if you score badly on your Wonderlic, there is a 98 percent chance that you are NOT making it. Do not confuse a 98 percent chance of NOT making it with a 98 percent chance of making it. All I'm doing is eliminating the players who are obviously going to suck. I still have to be intelligent within the smaller group I've created.

I'm willing to risk losing Brees and Cutler and McNabb if it means I don't draft Smith, McNown, etc.

My philosophy is essentially risk management. How do I reduce the chances of screwing up my team?

Once I have cut all the deadwood, then I will go to the combine and start worrying about release point and foot work and and stuff? I'm not going to fire the fat guy who played at Toledo in 1959. I'm just cutting down his pool of candidates. Because those fat guys are the ones who say, "My God...Akili Smith has the best arm angle I've ever seen. We can make him a superstar."

And then you draft a guy who looks great in a uniform, but can't play.

I say, "No...here are the six guys I'm interested in. Rank them for me."

It's not quite Moneyball where Billy Beane told his scouts "Here are the six guys I'm interested in and I already ranked them." But the philosophy is the same. Too many people misread Moneyball. They think it's a book about fat guys drawing walks. It's actually a book about inequalities in the market place and exploiting them. Fat guys who walked weren't drafted because they were the key to winning. They were drafted because they were undervalued. High school pitchers were over-valued so, the As drafted fat guys who walked and refused to draft high school pitchers. Closers were over-valued so the As developed closers and sold them high. Pitchers who struck people out were over-valued. Pitchers who did not walk batters were undervalued.

In football, the completion percentage in college seems to be undervalued. IQ may be under-valued. Height is not under-valued, but too many coaches seem to talk themselves out its importance. So, that's where to strike.

Spread option teams? Has any QB ever made the leap from spread option to efficient NFL QB within 3 years? I don't think so. I'll let the Broncos develop Tebow and prove that it can be done.

I worked at a hospital for four years. My job was new product evaluation and purchasing. Our philosophy was "Let the teaching hospitals like Georgetown be the first to test something. We want to be #2." We were not positioned to be cutting edge; it was too expensive and too risky. But as soon as Georgetown or Hopkins proved that a product worked, we wanted to be the second in the market to grab it. We were never going to end up on the cover of Time that way, but we were never going to inadverntantly kill half the NICU either.

[K.O. Johnny]

John Elway, Jay Cutler and Archie Manning are the three best QB athletes I've ever seen.

Archie played for the hapless Saints and had the worst won/lost percentage in NFL history for any QB with 100 or more starts. He isn't in the HOF because his team stunk. He was much more of an individual threat to defenses than either of his sons.

Athletic QBs can do more with less supporting talent than other QBs. By using Cutler's ability to move and throw on the run, Mike Shanahan could make his O-line look better than they really were (fewest sacks in the NFL in 2008 with eleven). Using Cutler's arm strength, Mike stretched the field vertically.

In the shoes of an offensive coordinator, I ask: What tools does this QB have that I can use to win games? That's how I grade QBs.

Cutler one of the best three QB athletes you've ever seen!! I think that's pushing it but I'll agree with Elway and Manning tried to make things happen with awful teams - but his interceptions are astronomical.

The most important question you ask is this:

In the shoes of an offensive coordinator, I ask: What tools does this QB have that I can use to win games? That's how I grade QBs.

The tools I'd say are co-operation and communication

I get the impression that there are two types of QBs - the vast majority are the ones who stand on the sidelines and are told "run this, if you see this, if you don't you audible this, then run this on this down and distance - they're little more than an extension of the OC's arm and as a consequence have no 'feel' for the game and no real input. I really think Jason Campbell is one of these.

But the really good ones treat the calling of plays as a collaborative effort, working with the OC - almost like an extra coach who can 'Throw a bit'. I get the feeling the McNabb failure is in part because it was a one way street from Kyle to McNabb. But if Elway got hurt and Gary Kubiak went in - you can be damn sure he'd be on the OC's shoulder like a parrot. Did you see that relationship with Mcnabb and Kyle?

[Lombardi's_kid_brother]

I have to come out say this: I like Jay Cutler. I advocated trading for him. I think he is already a good QB and with proper coaching that controls his worst impulses could be an excellent QB. But I have no earthly idea what OF sees in him.

I've never once watched him and said, "My God. It's like watching Steve Young reborn." I think Roethlisberger is a better "athlete." Everytime Cutler starts to scramble, his mechanics fall completely apart no one throws more late passes across the middel of his back foot from a weird arm angle.

To me, that's "athleticism" in a QB. When everything is breaking down around you, do you maintain your balance and mechanics? Jake Plummer could keep his body underneath him and throw a beautiful pass on a designed roll-out. In fact, he may have been the best I ever saw at that. The problem was, as soon as Lavar Arrington got a hand on his shoulder pads, his mechanics went to pieces.

Manning is by no means a great athlete, but he keeps it together as well as anyone. He rarely makes a throw off-balance - unless he has to run. Manning running and throwing is not always a pretty site.

Marino was probably the best ever at this. He had clay feet past the mid 80s, but he could have Dexter Manning munching on his spleen and still have perfect footwork. (Actually Marino is remembered too much for the last, I dunno, 26 years of his career. At Pitt, he could scramble and throw on the move as well as anyone. It's like he atrophied at age 27).

[PF Chang]

OK. I'm more cognizant of the "cutoff" concept you're pointing to now. That's a statistical way of doing things that I could get behind (evaluate in the sense that gives you the highest % chance of not busting).

---------- Post added January-6th-2011 at 12:11 PM ----------

I have to come out say this: I like Jay Cutler. I advocated trading for him. I think he is already a good QB and with proper coaching that controls his worst impulses could be an excellent QB. But I have no earthly idea what OF sees in him.

I've never once watched him and said' date=' "My God. It's like watching Steve Young reborn." I think Roethlisberger is a better "athlete." Everytime Cutler starts to scramble, his mechanics fall completely apart no one throws more late passes across the middel of his back foot from a weird arm angle.[/quote']

Oldfan LOVES Cutler but openly admits it is based on his own subjective analysis. But the guy throws too many damn interceptions and I don't think that can be explained away.

[Lombardi's_kid_brother]

OK. I'm more cognizant of the "cutoff" concept you're pointing to now. That's a statistical way of doing things that I could get behind (evaluate in the sense that gives you the highest % chance of not busting).

---------- Post added January-6th-2011 at 12:11 PM ----------

Oldfan LOVES Cutler but openly admits it is based on his own subjective analysis. But the guy throws too many damn interceptions and I don't think that can be explained away.

Yea, re-read the billion words I've written here. At no point have I said, Here is the Holy Grail!!!

I'm just trying to figure out a way to avoid huge mistakes.

The sports landscape is littered with can't-miss guys who missed and we still can't figure out why. (Again, read Moneyball. Billy Beane was, in fact, one of those can't miss guys. He was a five-tool player who was drafted in the First Round. And he couldn't play. It tells me that one day, Tim Couch will be the best GM ever).

NFL and NBA teams both consistently make the same mistakes when drafting.

1. They fall in love with measurables before looking at production.

2. They punish experience. (If you watch 40 games of a QB, you will find some stinkers. If you watch 4 games, you may not find any. Absence of evidence becomes evidence of absence).

This is how you get Akili Smith. Less than a season. No production. But God, he looked good at the combine. It too often goes, Combine then college numbers. I want to go College Numbers, then combine.

Except for running backs. College running backs with huge production terrify me. I think they are all used corn husks.

[gortiz]

No' date=' I don't. Because I don't think pass rush impacts completion percentage that dramatically. I think it impacts ypa and maybe interceptions. [/quote']

i agree ...

OL pass pro and DL pressure does not have that much impact on a QB's comp %. In fact neither does WR play, as weird as that sounds. At the end of the day some play callign and more importantly a QB's ability to read a D presnap and take an educated guess on who the open person on the first few reads of a passing play will be. Thow in mechanics and muscle memory as big factor. That is why I think comp %, over the course of a season ... is one of the better ways to meausre a QB.

Peopel use to laud Ted Williams' bat speed ... and while it was amazing, his ability to "out guess" pitchers was even more amazing. Same thing applies here ...sure bat speed was important, but Williams was such a student of the game, that is what made the difference. If he had a slow bat he would ended up with a .320 career average as opposed to his whoop ass .340 average.

---------- Post added January-6th-2011 at 12:34 PM ----------

At no point have I said' date=' Here is the Holy Grail!!! I'm just trying to figure out a way to avoid huge mistakes.[/quote']

has the 26-27-60 rule been discussed in this killer thread?... I'm really intrigued by this. seems like a pretty good performance indicator.

[Oldfan]

LKB, I'll deal with your Cutler issue only in this post.

LKB: It's not a great stat, but Cutler's QB rating over his career is:

88.5

88.1

86.0

76.8

86.3

If he was any more consistent, he would be a metronome. 2009 had three ridiculousy, embarrassingly awful games.

Your conclusion based on the QB rating is that Cutler has shown consistency in his performance between Denver and Chicago. My theory is that the QB rating is just coincidentally consistent in this case because it's not a good way to grade performance.

To prove my point I'll use the rankings at footballoutsiders.com. Since you are the one who first led me to that site, I think you'll agree that the stats guys there at least understand the problem even if they don't have it all figured out.

Cutler's 2008 ranking in Denver: #5 (Shanahan)

Cutler's 2009 ranking in Chicago: #36 (Turner)

Cutler's 2010 ranking in Chicago: #25 (Martz)

Neither the QB rating nor these rankings tell us much about Jay Cutler because they are measurements of the team quarterbacked by Jay Cutler not measurements of Cutler's individual performance.

---------- Post added January-6th-2011 at 12:58 PM ----------

...Oldfan LOVES Cutler but openly admits it is based on his own subjective analysis. But the guy throws too many damn interceptions and I don't think that can be explained away.

I admit it's a negative, however, I'm sure that most fans think that it's more of a negative than I do.

LKB, for example, says he thinks the TD/INT stat is important -- but he doesn't tell us how he uses it.

Does he realize that more than half of all INTs are caused by losing? A QB who has to force the ball trying to play catch up on the scoreboard will throw more INTs than one who plays with leads most of the time. From memory -- I think Cutler threw 18 INTs in 2008 and 12 of them happened with his team behind on the scoreboard. So, the QB on teams with lousy defenses can be expected to throw more picks.

Another point -- TDs are worth seven points and the average INT is worth four. So, let's assume that Cutler forces the ball for a TD for every time he forces and the pass is intercepted. If he could maintain that kind of ratio, we'd want him to keep forcing it. Now, we don't know that ratio, so we really don't know the extent of the problem.

---------- Post added January-6th-2011 at 01:15 PM ----------

has the 26-27-60 rule been discussed in this killer thread?... I'm really intrigued by this. seems like a pretty good performance indicator.

I checked out your link. It's the Lewin Forecast method I mentioned earlier with the Wonderlic feature added. It's my impression that statisticians are generally not impressed with David Lewin's approach.

[Thinking Skins]

Yes' date=' a short passing game versus a long passing game will impact YPA. But it will also make up for it in more receptions. Brady was not exactly heaving the ball downfield this year like he was in 2007. Yet, his YPA was fifth in the league and the second highest he ever put up (behind that insane 2007 campaign).[/quote']

YPA doesn't do it for me either because (staying with Brady), if he completes a 60 yarder to Wes Welker, then it was probably more like a 15 yarder with 45 yards after the catch. While if he completes a 60 yarder to Moss, then its probably a 50 yard completion with 10 yards after the catch. I think its pretty important to know how much of the pass is due to the QB's arm vs the WR's elusiveness.

[Lombardi's_kid_brother]

That's true, but the point you seem to be ignoring is that offensive coordinators affect the completion percentage for all QBs. So, the completion percentage reflects the QBs accuracy, both the quality and philosophy of the offensive coordinator, the quality of the receivers and other factors. The completion percentage is a useless number.

Sam Bradford is a more accurate passer than Peyton Manning. I arrived at that opinion by watching them throw a football. Your useless completion percentage stats won't support or deny that conclusion because the two players operate in completely different environments. Your stat has no more value in comparing QBs than the final times of a race between two sprinters in which one has to carry a 30-pound weight on his back.

This is FireJoeMorgan-level "I watch the games!" bull****.

So....the only way to assess a QB is to watch them....and the only way to accurately understand what they are doing is to be you. Because you haved decreed that Archie Manning and Jay Cutler are the greatest QBs ever and it's everyone's fault but theirs that they have not accomplished more.

[Thinking Skins]

What you're looking for doesn't exist. It's the Holy Grail for statisticians interested in Football. ASF's original projections for QBs was a variation of the Lewin Forecast which is still promoted on the Football Outsiders website. Not many statisticians take it seriously.

Here's the problem in a nutshell. Let's say that QB-A and QB-B both have equal talent potential. However, QB-A plays with support that allows him to perform at 90% of his potential while QB-B plays with support that allows him only to play to 50% of his potential. There's no way in hell you could possibly know that these two QBs have equal talent by grading their performances.

The only intelligent way to grade QBs is to do it as a scout might, by grading their tangibles. As to the intangibles, these are difficult to grade because they're intangible. Mostly, people hype or trash a QB on the intangibles because they can't be proven wrong.

I agree 1000%. Numbers are numbers and as a mathematician, I understand their power, but also their limitations. Trying to come up with a formula to predict a QBs success is BY DEFINITION using the past to predict the future. It doesn't take into account say how much a coordinator like Norv Turner can impact a QB (look at Alex Smith in his one year with Norv in SF) or the continuity of an offense, or simply the maturation of a QB (ala Mike Vick admitting that he didn't study film until this year).

---------- Post added January-6th-2011 at 01:35 PM ----------

I know you still like Campbell' date=' but I think it's interesting that his best year for completion percentage is the year where he was sacked the most. So, I don't think "pass rush" or "offensive line protection" or whatever necessarily matters that much. (Roethlisberger's best year for completion percentage was also the year he was sacked the most).[/quote']

You wouldn't agree that it impacts different QBs differently? I mean, some guys can extend plays, so when they are forced to scramble because the OL broke down, they're also forcing the coverage to hold that much longer, which can lead to more completions. Other guys, when the line is gone, they're just sitting ducks.

---------- Post added January-6th-2011 at 01:37 PM ----------

If I was drafting a QB out of college' date=' I would look at two numbers first and foremost. His college completion percentage and his IQ. The two biggest "intangibles" that kill college QBs are stupidity and laziness. And sometimes those go away after a while.[/quote']

But how do you measure a kid's laziness? How do you measure his coachability? How do you know if he'll be a locker room distraction? Are there stats for this?

---------- Post added January-6th-2011 at 01:47 PM ----------

I pretty much agree with you, stats are an indicator but don't tell the whole story. If you look at Brees' Int's this year you think "damn he's having a terrible year". But how many of his Int's have been tipped balls, busted routes, players falling down, etc? His supporting cast has kind of let him down this year, especially with Colston being injured.

This is another thing I was thinking about in terms of stats like completion percentage, tds, and ints. Drops are critical to a QB. There seems to be a unmentioned assumption that drops affect all qbs equally, but this isn't really the case. I'd figure that a QB at Vanderbelt during a 1 win season probably had more dropped balls than a QB at USC during a 10 win season. But were these balls dropped because he didn't put it in a good enough place and forced a WR to bend awkwardly to make the catch, or is it just a consequence of WRs not paying enough attention?

---------- Post added January-6th-2011 at 02:02 PM ----------

Third down percentage is an attempt to measure "clutch" and I don't think "clutch" actually exists. This has been shown across all sports - the players who do well in "clutch" situations actually perform at the same level as they always do. The problem with "clutch" is people's memories. David Ortiz is actually about a .300 hitter in close and late situations. He just got two of the biggest close and late hits in World Series history. So' date=' he is forever "clutch." Everyone remembers those two hits. No one remembers the times he failed.[/quote']

Would you say the same thing about TDs then, particularly red-zone TDs? All they are, really, is completions (that got lucky). Similarly, all interceptions are, really, is incompletions (that got unlucky).

---------- Post added January-6th-2011 at 02:08 PM ----------

One quick thought to explain my thinking on completion percentage. It's not a perfect stat obviously. And neither is batting average' date=' 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.[/quote']

Thats a big assumption. Maybe over the course of an NFL career these types of things will even themselves out, but even then I have my doubts. Every stat can be broken down a number of ways: batting average becomes singles vs doubles vs triples vs homeruns vs strikeouts vs foulballs vs walks. Completion percentage becomes completions over the middle, completions in the backfleid, completions deep, completions to the left, completions to the right, completions to the HBs, TEs, WRs. Incompletions become interceptions, batted balls, drops, throwaways, etc. All these things lead to more specific judgements on a hitter or a QB. Just because QB A and QB B have the same stats doesn't imply that their stats break down even remotely similarly.

---------- Post added January-6th-2011 at 02:20 PM ----------

I feel so bad missing most of this conversation. :-(

I'm not suggesting throwing Tom Brady out. I'm saying you treat him as an outlier - one of those things that make life interesting.

If your investments made a 5 to 7 percent return every year for 30 years - except 1997 because you got into tech stocks at the right time and got out at the right time so you made 26 percent - you wouldn't try to build your entire investment model around 1997. Crazy **** happened that year.

The way I read OF is that he wants to start with Tom Brady's 2007 and Randall Cunningham's 1998 and try to account for that first. I want to build from every year but 2007 and then try to figure it in last. When I get there' date=' I write down "Randy Moss" and move on with my life.[/quote']

I think Tom Brady is the perfect example of why its so hard (probably impossible) to predict QB behavior. This guy has gone through a number of changes as a QB. I mean his growth and development has been amazing. He went from a short throw type guy when he came into the league (with scouting reports which questioned his arm strength) to a big play guy when he teamed up with Randy Moss.

If Brady is an outlier, then I'd say that any QB who changes over the course of their career is an outlier (Steve Young, Mike Vick). This particularly makes it almost impossible to judge the successful QBs who are later round picks like a Gus Frerotte or a Tom Brady or a Matt Cassel, because generally the whole M.O. behind them is that they can be successful, but its gonna take time, but they're low risk high reward type guys.

[Oldfan]

This is FireJoeMorgan-level "I watch the games!" bull****.

So....the only way to assess a QB is to watch them....and the only way to accurately understand what they are doing is to be you. Because you haved decreed that Archie Manning and Jay Cutler are the greatest QBs ever and it's everyone's fault but theirs that they have not accomplished more.

When my debate opponent resorts to strawman arguments, I take it as a concession speech gone awry.:pfft:

[Lombardi's_kid_brother]

When my debate opponent resorts to strawman arguments, I take it as a concession speech gone awry.:pfft:

I don't know how to debate "I know it when I see it."

We've had an intelligent discussion for most of these pages. And suddenly, your conclusion is "Archie Manning is the third best QB ever because I liked him."

Fine. Two can play this game. Dan Pastorini was the best QB of the 1970s. Disprove my theory.

[Oldfan]

I don't know how to debate "I know it when I see it."

We've had an intelligent discussion for most of these pages. And suddenly' date=' your conclusion is "Archie Manning is the third best QB ever because I liked him."

Fine. Two can play this game. Dan Pastorini was the best QB of the 1970s. Disprove my theory.[/quote']Stop with the drama. I responded to another post with an opinion necessary as a preface to answering a question posed by another poster. I didn't "suddenly" ruin this intelligent discussion anymore than your random attempts at humor ruined it.

[LD0506]

Awwww don't let us all down now, it's been going so well!

[ZoEd]

Wow, excellent post LKB. I agree 100% with what you're saying and we kind of use the same terminology in Tech Training in the military. To qualify for a certain Air Forse Specialty Code you have to have certain numbers in the ASVAB.

For example, the Structures Career Field in Civil Engineering you're required to have a 47 Mechanical score. This isn't saying that a person with a 42 Mechanical score can't be an outstanding Structures Airmen, but it helps "weed" out the ones who will potentially be busts. That doesn't mean that everyone with a 47+ Mechanical score will go on to success either. But lets face it, you have to find something to help you discriminate or you'll end up spending tons of $$ taking uncessessary risks.

Same thing goes for football I guess, of course you'll miss the boat on a good QB from time to time but you need to narrow it down to a group worthy of taking that risk and then start looking at the other factors.

Excellent analysis LKB

[PF Chang]

ZoEd,

Correct me if I'm wrong but it seems like from your example that you're describing a situation where the score is based 100% on the candidate himself. % percentage is a different story, which is what OF is trying to say.

[Lombardi's_kid_brother]

ZoEd,

Correct me if I'm wrong but it seems like from your example that you're describing a situation where the score is based 100% on the candidate himself. % percentage is a different story, which is what OF is trying to say.

No one has ever explained why all the elite QBs in the NFL had great completion percentages in college and why so many of the busts did not. If it means nothing, shouldn't there be a mix? Shouldn't there be a few 52 percent guys who turned into 65 percent guys in the NFL?

Also, shouldn't it vary wildly for QBs who move from team to team? Why does Jay Cutler's stay the same? Why does Jason Campbell's stay the same?

There is probably a degree of inter-connectedness between completion percentage and team performance. But you aren't showing ANYTHING that supports the notion that team performance is the over-riding factor.

As much as I like Moneyball, I thought the Blind Side missed the mark in some ways. It discussed how Bill Walsh could make anyone a highly accurate passer. And it used Steve DeBerg as an example. DeBerg was in the 40s his rookie year. Walsh made him a 60 percent passer in one year.

It has to be the scheme, right?

Then why did DeBerg remain a 60 percent passer throughout his long NFL journey?

Seriously, show me one example of a QB whose completion percentage changed dramatically due to a new scheme. The only one I'm even remotely aware of is Michael Vick. But that might be explained by the "Trying/Not Trying" thing. He and Randy Moss have a lot of the same traits.

(Randy Moss might be the only Hall of Famer who played hard in less than 30 games).