The Almighty Quarterback Bandwagon Runs Out of Control

[Oldfan]

This line of thinking also emphasizes the impact of coaching on the NFL. The best coaches know how to spread their points around the team such that they are getting maximum value out of each position or unit.

When using 100% to represent the entire team strength, we quickly realize that the coach can reduce the value of the QB position -- which makes other positions worth more. Or, the reverse, he can inflate the value of the QB which reduces the value of the other ten positions on the offense.

So, while Belichik reduces the value of the QB position, Shanahan inflates. Thus, we can say that RG3 has much more impact on game outcomes than Brady, but we can't say that Shanahan's scheme will win more games. In order to do that, we would need to weigh both offenses, not just the QBs.

---------- Post added November-29th-2012 at 01:42 PM ----------

...through this season, id probably put rg3s value at nearly 30%. (rough estimate).

30% is much too high; but it's not so high that you qualify for the lunatic fringe of the bandwagon.

[Lombardi's_kid_brother]

Thinking about it further, it occurs to me that you could possibly predict the outcome of games based on assigning values to the percentage impact of each position (or player).

For instance, Jay Cutler or Ben Rothlisberger may play in schemes where they have a potential 10-12% impact on their teams success. This is a high number relative to the other positions on the team and we can say that they regularly match that percentage; meaning that they add 10-12 "points" out of 100 to their teams chances of winning. Jason Campbell and Byron Leftwich may have a lower impact ceiling (say 8-10 points) so unless their coaches choose to move those points into a different phase of the game via scheme, they are potentially lost to inferior talent.

I don't disagree with any of the theory here.

The problem is: these numbers are all being pulled out of someone's ass.

What you are all trying to come up with is a WAR stat for football. Football Outsiders has that with DVOA. I'm not completely sold on DVOA, but at least there is a methodology to it.

All anyone on this thread is doing is saying:

"The QB is responsible for 10 percent of a team's success."

"No...it's responsible for 11.5 percent of a team's success."

"No...it's 7.8 percent."

This is stupid. So very very stupid.

If somebody would bother to show me a methodology...or at least act like one existed...I would not think this is stupid.

And whoever brought up Mark Sanchez is ridiculous. Sanchez has always stunk. His completion percentage has never gotten out of the mid 50s, when in the modern NFL, you have to be at 60 percent to even justify your presence on the field. Rex Ryan was able to survive with him for two years the same way you can sometimes win big at blackjack by splitting tens or hitting on 16 against a 2. It was luck.

No one ever wants to acknowledge how much of pro football is luck. The sample size in pro football is ridiculously small so a run of abnormal luck can easily artificially inflate your image or decrease it. The most successful coaches and players tend to be the ones that put themselves in a position to take advantage of these runs of luck.

Tim Tebow's run of nonsense last year is an all-time example of this. The Bears' run to the Super Bowl with Grossman is another.

This only happens in football. If Thebo Sefalosha suddenly shot 60 percent from three point range for four or five games, no one would say, "You need to build a team around his three-point shooting." If someone in baseball hits .600 for 8 games, no one immediately wants to sign him for a 5 year, $100 million deal.

In football, five games can make you Matt Cassell or Rob Johnson or - christ - Todd Collins. Remember Todd Collins? We were pining for Todd Collins once upon a time.

[The Robert Griffin Experience]

How much does the QB affect the defense? Or rather, how much does the offense affect the defense? For example, an offense that relies on the QB to put up points consistently or can control the ball through accurate passing (if not both) may improve the defense's performance by 5%. If the QB's impact on the offense alone is 30%, then wouldn't that mean that he would get an extra 1.5% for how he affects the defense? I think in your calculation, you assume the QB has 0 impact on the defense and I really doubt that's the case. Even if it's not a high number, I'm quite sure that impact is greater than zero.

If somebody would bother to show me a methodology...or at least act like one existed...I would not think this is stupid.

We are largely speculating, but I think to pull that off, you'd need PFF-style grades on every player on every play, then derive a DVOA or something similar from from that.

[Lombardi's_kid_brother]

We are largely speculating, but I think to pull that off, you'd need PFF-style grades on every player on every play, then derive a DVOA or something similar from from that.

At least you are admitting like you are speculating and not acting like your "facts" were whispered to you by George Halas' ghost.

[Gibbs Hog Heaven]

At least you are admitting like you are speculating and not acting like your "facts" were whispered to you by George Halas' ghost.

There's a GHOST of George Halas?

Man I'm glad I'm not a Bears fan. That's one tough MF I wouldn't want spooking me.

Hail.

[Jethrodsp]

When using 100% to represent the entire team strength, we quickly realize that the coach can reduce the value of the QB position -- which makes other positions worth more. Or, the reverse, he can inflate the value of the QB which reduces the value of the other ten positions on the offense.

So, while Belichik reduces the value of the QB position, Shanahan inflates. Thus, we can say that RG3 has much more impact on game outcomes than Brady, but we can't say that Shanahan's scheme will win more games. In order to do that, we would need to weigh both offenses, not just the QBs.

For sure.

My point was to indicate that the difference between the success of teams throughout the league isn't necessarily the talent on the team or even the quarterbacks so much as it is each coach's ability to accurately gauge what the impact ceiling is for each player and how likely they are to reach that impact ceiling each time out on the field.

That being said, the more potential impact points a coach can rely on out of one player, the easier it becomes to maximize points across the rest of the roster; which is probably why we see so many Super Bowl trophies in the hands of really good, consistent performers at the quarterback position.

[Oldfan]

...This is stupid. So very very stupid.

If somebody would bother to show me a methodology...or at least act like one existed...I would not think this is stupid.

Does it bother you in the least that some obviously intelligent posters have come in to critique the methodology that you claim doesn't exist?

I know I must have asked you this before but -- can you give me a better idea of how much you think the average QB position is worth? I understand that you think that 10% of a team's strength (100%) is too low' date=' but how much difference are we talking about?

---------- Post added November-29th-2012 at 02:09 PM ----------

At least you are admitting like you are speculating and not acting like your "facts" were whispered to you by George Halas' ghost.

When you can't debate your opponent' date=' ridicule his position?

---------- Post added November-29th-2012 at 02:13 PM ----------

For sure.

My point was to indicate that the difference between the success of teams throughout the league isn't necessarily the talent on the team or even the quarterbacks so much as it is each coach's ability to accurately gauge what the impact ceiling is for each player and how likely they are to reach that impact ceiling each time out on the field.

That being said, the more potential impact points a coach can rely on out of one player, the easier it becomes to maximize points across the rest of the roster; which is probably why we see so many Super Bowl trophies in the hands of really good, consistent performers at the quarterback position.

I understood you and agree. I didn't make clear that I was adding a different slant onto your thoughts on coaching adjustments.

[Jumbo]

I was going to post this (below), and in the first version I had included "of the clueless people like LKB" right after "some" in the first sentence but had edited it out.

Then I just decided to hold off on posting it at all for very defined reasons. Usually, whenever I do that, the reasons I was holding out will be removed, and usually by exactly the person (when there is one) that I was considering in evaluating the matter.

This was my still-withheld, even after edit, post :

Maybe these last few posts will help clue in some people to what OF is saying---and note his model is incomplete, newborn, and very open to modifications that will enhance efficacy (even seeks such), while resisting unsupported, logically faulty, or outright stupid challenges.

OF's most basic contention is a more sophisticated (if still embryonic in applicable detail) version of something most people accept already---that the QB gets more credit for success than is often merited. And his selection of assigning numeric values is a point of departure, but still with informed reasoning behind it. Calling it stupid or ass-pulled as a critique, when you're the one posting more "stupidly-and-with-attitude", LKB, is no worthwhile counter to his framing. It shouldn't be that hard to find more reasoned counters if his thoughts are all thaat shaky.

A solution might not be possible; but you are in rare company among football minds in that you understand the problem and can ignore bogus solutions.

This is another of OF's long-standing positions in argument (and I mainly agree with it). QB impact tends to be often and significantly exaggerated by fans and many commentators, and popular rating systems are quite flawed. That DOESN'T mean they (QBs) can't BE significant OR very impacting. Durrr.

Nor does it mean OF has found another Golden Ratio. But he applies serious thought and effort. Speaking very generally, I think he goes too far in minimizing the QB effect, and sometime soon I may detail some reasons why.

If you post again in this thread, you need to have zero attempts at wit or sarcasm and be 100% debating any targeted contentions in substantive fashion. And no need to PM me. I still love you, but I will kick you out of the house.

(kidding)

[Lombardi's_kid_brother]

Does it bother you in the least that some obviously intelligent posters have come in to critique the methodology that you claim doesn't exist?

I know I must have asked you this before but -- can you give me a better idea of how much you think the average QB position is worth? I understand that you think that 10% of a team's strength (100%) is too low, but how much difference are we talking about?

I don't know. Probably a lot. Anecdotally, the difference between Ben Roethlisberged and Charlie Batch appeared to be 6000 percent.

But I think the entire idea of any player being measured as a percentage of the team's overall success is flawed from the beginning.

I would assume that a QB is as important to winning one game as a pitcher would be in baseball. The difference is a QB plays every game where as a pitcher plays 1/5 of all games. So whatever a starting pitchers value is times 5 is what a QB is worth.

Verlander was third in baseball last year in WAR, meaning that the Tigers would have lost 7 more games with a league average starter there. Replacing him with the baseball equivalent of Charlie Batch would be much much worse of course.

I have no earthly idea how to quantify this, but I suspect that RGIII will be worth something like 2 or 3 wins per year to the Redskins. In a league designed for every team to be 8-8, that's incredible. If someone can create a formula that shows me this wrong, I would be very willing to change my tune.

I think his value is inflated right now, because the rest of this Redskins' team is awful. I'm fairly certain that this is a 1-win team with Grossman or someone like Tannehill starting right now. Again, football is weird because it is such a small sample size. I'm not sure you can get any really useful data out of one season. Maybe not even two. It's a frustrating exercise when people aren't just spouting random numbers.

The question is not how much of the team's value does the QB make up. The question is how many wins does a certain QB provide over another QB. How many more wins per year does RGIII provide over Cousins.

And this is not Mike and Mike nonsense. The answer could be 1.32 or something like that.

I mean ultimately does it matter how much of the team's overal success is provide by whoever Arizona's QB is this week? No, the question is, what would Arizona's record be with a league average QB. And what would Arizona's record be with Tom Brady or Drew Brees. That's what matters.

---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------

Here is something fun for someone to do.

This stat would mean absolutely nothing, I think, and what I am suggesting is certainly the wrong way to quantify it.

But I would like someone with time to create a spreadsheet that takes every QB in the league who has played meaningfuly minutes and calculate this:

Completion Percentage * Yards Per Attempt * 1/2 (TDs/Interceptions) * 100

I have a sneaking suspicion that somewhere down that path lies a pretty interesting number.

[Oldfan]

For what it's worth... I like LKB a lot. He's a very intelligent poster. But when I attack his belief in the Almighty Quarterback, he flips out. I think I could attack his religious views and get a more reasoned response.

(written before seeing his 184)

[skinnyskins]

23.2% -- the 33 pass attempts as % of 142.4

18.6% -- after reduction for 20% coaching factor: techniques, playcalling, scheme

09.3% -- after reducing for 50% value of protection and receivers

9.3% -- value of the average QB position

My primary source for stats:

http://www.teamrankings.com/nfl/rankings/

How did you come up with 20% for coaching and 50% for protection and receivers? They appear arbitrary.

[Lombardi's_kid_brother]

And I will go a step further with this WAR concept.

It applies everywhere.

London Fletcher in his prime is worth more in terms of wins to a team than a league average MLB. I suspect it is smaller than a QB's value, but it matters. This is not revolutionary. What made Cal Ripken so revolutionary was that he could meet the standards for defensive play at the SS position while providing corner inflielder hitting numbers. He was so much more valuable than a Tim Foli-type shortstop that it was unfair.

[zoony]

I don't disagree with any of the theory here.

The problem is: these numbers are all being pulled out of someone's ass.

What you are all trying to come up with is a WAR stat for football. Football Outsiders has that with DVOA. I'm not completely sold on DVOA' date=' but at least there is a methodology to it.

All anyone on this thread is doing is saying:

"The QB is responsible for 10 percent of a team's success."

"No...it's responsible for 11.5 percent of a team's success."

"No...it's 7.8 percent."

This is stupid. So very very stupid.

If somebody would bother to show me a methodology...or at least act like one existed...I would not think this is stupid.

.[/quote']

:ols:

[Jumbo]

Good stuffs, LKB.

I need to find time to add something I think about evaluating a QB, and it sure doesn't lend itself easily to numerical value or objectively quantifiable measurement.

Gets into that hoary (and murky) turf of "intangibles", but I think it's actually is a tangible, but is not fungible. :pfft:

[S.T.real,lights,out]

Im really confused....

You're saying the QB pos is the most important pos on the field right?? I would think we would all agree with that...:whoknows:

[Jumbo]

Im really confused....

You're saying the QB pos is the most important pos on the field right?? I would think we would all agree with that...:whoknows:

Well, there's certainly more being said. Some miss it willfully. I don't think that's you. But in the end, it's just guys talking football. Not all such is the same in detail or form or merit, nor needs to be. One thing you can say about it all...it's all here.

So....

[S.T.real,lights,out]

Well, there's certainly more being said. Some miss it willfully. I don't think that's you. But in the end, it's just guys talking football. Not all such is the same in detail or form or merit, nor needs to be. One thing you can say about it all...it's all here.

So....

It seems like a lot of OF's threads kind of go that way...

Kind of the reason i like them...a little late to the party on this one tho.

---------- Post added November-29th-2012 at 03:23 PM ----------

Why?

The Steelers outgained the Ravens by 100 yards, the Ravens lone touchdown came on special teams by Jacoby Jones.

In a 20-14 loss do you think a team improves their chances to win if they don't fumble the ball away 5 times?

I ask again: Is it possible for a team to be non-competitive on offense and still be a competitive team?

Sure it is....we have done it for years....

[Oldfan]

...I have no earthly idea how to quantify this' date=' but I suspect that RGIII will be worth something like 2 or 3 wins per year to the Redskins. In a league designed for every team to be 8-8, that's incredible. If someone can create a formula that shows me this wrong, I would be very willing to change my tune. [/quote']We aren't as far apart as you think. Replacing Rex with Robert jacks up two factors at once. It raises both the position value and the talent factor. Two or three wins is a reasonable estimate for the improvement if all other factors remained the same.

[Jumbo]

We have plenty of threads of mundane premise featuring "me too!1!! forever and ever!!11 over and over!!1" action, or that don't emphasize all that much thinking--let alone traditional critical thinking skills---in here.

Always somewhat disappointing to me that with the amount of that other inventory so consistently plentiful, we still get people acting like a little of the other, even if taking unusual perspectives at times, needs to be dragged down or impugned. But it's the internet and wtf do I know?

[Oldfan]

How much does the QB affect the defense? Or rather, how much does the offense affect the defense? For example, an offense that relies on the QB to put up points consistently or can control the ball through accurate passing (if not both) may improve the defense's performance by 5%. If the QB's impact on the offense alone is 30%, then wouldn't that mean that he would get an extra 1.5% for how he affects the defense? I think in your calculation, you assume the QB has 0 impact on the defense and I really doubt that's the case. Even if it's not a high number, I'm quite sure that impact is greater than zero.

I allow the teamwork factors to offset.

My assumption is that teamwork is always a two-way street. The offense affects the defense, the defense affects the offense. A good QB can make his receivers look good; good receivers can make the QB look good.

Bear in mind, I'm talking about a 32 team average for QBs. Individual teams might have factors that don't offset. Some defenses don't help their offenses; some offenses don't help their defenses.

---------- Post added November-29th-2012 at 04:24 PM ----------

...But I would like someone with time to create a spreadsheet that takes every QB in the league who has played meaningfuly minutes and calculate this:

Completion Percentage * Yards Per Attempt * 1/2 (TDs/Interceptions) * 100

I have a sneaking suspicion that somewhere down that path lies a pretty interesting number.

What are you attempting to measure? If you want to know whether quarterback A's performance for team X was better of worse than quarterback B's performance playing for team Y' date=' then the formula will answer your question. But, we already have 87 formulas that will do the same thing.

Now, if you really want to know whether A or B is the better QB, your formula is as useless as the others.

---------- Post added November-29th-2012 at 04:27 PM ----------

How did you come up with 20% for coaching and 50% for protection and receivers? They appear arbitrary.

They are not arbitrary. They are reasonable estimates.

You may question their reasonableness, of course; but no one else has, which I find significant.

[skinnyskins]

[/color]They are not arbitrary. They are reasonable estimates.

You may question their reasonableness, of course; but no one else has, which I find significant.

That does not answer the question of how those specific numbers were derived.

And just because nobody else has questioned it, it does not make the numbers absolute. You of all people can appreciate that

This is important because it reduces the impact of the QB dramatically, according to your model.

[Oldfan]

That does not answer the question of how those specific numbers were derived.

And just because nobody else has questioned it, it does not make the numbers absolute. You of all people can appreciate that

This is important because it reduces the impact of the QB dramatically, according to your model.

I gave you satisfactory answers to your questions. I have nothing more to say on your point.

[skinnyskins]

I gave you satisfactory answers to your questions. I have nothing more to say on your point.

"Reasonable estimates" is the same as educated guess. Which is not really all that scientific. In earlier posts you have stated that it is impossible to determine the impact the coaches have on the performance of a team, yet here you have given coaches an exact number at 20%. You then go on to say that the offensive line and receivers contribute 50%. Asking why you chose those numbers is important.

You don't have to respond to this, but just because you create models with numbers does not make them anymore justified than simply saying, "I think the QB only contributes X amount to any given offense" If you can provide a scientific model that answers how you came up with those variables then fine, I will accept that. But, as the model stands now, adding 'estimates' to a model and coming up with 9.3% does not cut it for me, nor should it be accepted by others until all variables in the model can be accounted for.

[Oldfan]

...In earlier posts you have stated that it is impossible to determine the impact the coaches have on the performance of a team, yet here you have given coaches an exact number at 20%...

You're confused. I have in previous posts said, for example, that it's hard to determine how much a coach like Haslett is responsible for the defense's problem. That's a completely different concept than making a reasonable estimate of the weight of the coaching factor as part of the effort I offered in the OP.

Reasonable minds judge reasonable estimates. The fact that no one, including you, has said that my estimates are too high or too low is significant. That should tell impartial readers that those estimates are not way out of line.

I didn't present this as "scientific." I don't even know what you mean by that word in this context.

[skinnyskins]

You're confused. I have in previous posts said, for example, that it's hard to determine how much a coach like Haslett is responsible for the defense's problem. That's a completely different concept than making a reasonable estimate of the weight of the coaching factor as part of the effort I offered in the OP.

Reasonable minds judge reasonable estimates. The fact that no one, including you, has said that my estimates are too high or too low is significant. That should tell impartial readers that those estimates are not way out of line.

I didn't present this as "scientific." I don't even know what you mean by that word in this context.

It's impossible to give an accurate percent for the impact of the head coach vs. the QB, vs the rest for passing offense so I can't give an estimate. However, your estimate is influenced by your own bias against the importance of QBs so I would bet that your numbers are too high. If you can provide clarification for how you arrived at exactly 20% and 50% it would really help determine if this is a 'reasonable' estimate or not.