The Dolphins Make Me Cry.com - Forums

Bleacher Report has rated the top 1000 players in the league and released the midseason results (or at least some of them).

http://bleacherreport.com/articles/2673471-nfl1000-ranking-the-top-1000-players-at-midseason

The Dolphins have two players in the top 50 (which is about what you'd expect), Reshad Jones at #16 (highest rated safety) and Ndamukong Suh at #33 (5th rated DT).

Tannehill was rated as the 10th best quarterback, which I find quite reasonable, but will probably elicit complaints from certain members of this forum...

Landry hit the #3 spot at WR (next best is Parker at #70).

Biggest surprise is perhaps Ajayi, who is only rated #37, but then he didn't really explode until the last two games included in the data - and he gets dinged for blocking and a bit for receiving.

Neither Sims nor Jordan grade out well at tight end...

The entire offensive line except Pouncey gets severely shafted, even Tunsil and Albert don't reach inside the top 20. Tunsil grades out as both the best left tackle and guard on the Dolphins. Pouncey scores a nice 8th position, which is hard to fault.

Billy Turner and Dallas Thomas ranked #101 and #102 out of 103 ranked guards, btw. How bad must Austin Blythe be at #103???

Wake is the only DE who gets even close to respectable grades (#23 our of 79 4-3 DEs).

None of our DTs other than Suh do well. At all.

Things look pretty grim at OLB, where Koa Misi is the top Dolphin at #46 among 4-3 OLBs.

Alonso receives a pretty harsh grade at ILB, which is pretty surprising IMHO.

To the surprise of absolutely no one, our best cornerback (Maxwell) resides at #86.

The honor of our defensive backfield is somewhat saved by our safeties with Reshad Jones the best strong safety and Abdul-Quddus the 10th rated free safety.

I think it is funny when anybody uses bleacher report to prove a point. What a shit  site.

On the other hand I'm willing to give anybody a listen if they provide a clear and convincing rationale for their viewpoint, and I don't see that in the article linked in the OP.  We're left with only a personal opinion, or an authoritative opinion, at best, and even then the grounds for considering the opinion "authoritative" aren't provided.

I used to think that way but then they were bought by TBS/Turner Broadcasting for $170 mill or so. They were less of a joke then and became legit.

I'm unsure if they still let fans write for them or if they hire full time real reporters.

They hired skilled people (including at least one associated with football outsiders) to look at every single play of every single player and judge what went on. That encompasses a lot more information than is possible to glean from "yards per attempt". Of course, it also (potentially) introduces more subjectivity, but don't kid yourself -- all non-raw stats are fundamentally subjective in their construction (e.g. passer rating has a subjective weighing of the components that go into it).

Personally, if someone with a good understanding of the game looks at every single play of every single quarterback and grades them according to some fixed criteria, I will certainly tend to put a lot of faith in that, even if I don't know the exact criteria used.

Then why is it so discrepant with the objective data, and what is that supposedly authoritative individual's explanation for that?

Your opinion about the validity of that person's opinion should ride on his reconciling that discrepancy (which in this case is a rather large one), not on his opinion alone.

If there is no basis provided for that reconciliation, then you're left with the strong possibility that the person's view is erroneous, and no way to reject that notion.

Tell me why Tannehill is at the bottom of the league in DVOA and QBR, despite your ranking him 10th overall.

You say that as if DVOA and QBR are objective data.

They're light years more objective than the information in the link in the OP.

QBR certainly isn't. The decisions underlying construction of passing DVOA are completely subjective, but at least the individual components are objective, although it suffers from the same issue as the ranking data here (or that of Pro Football Focus) in that the "formula" is not publicly available and you are forced to trust the entity that provides the answer -- there's no methodology provided and no way to reproduce the results.

You'll have to pardon me if I find it hard to believe that one authoritative person, or even a small group of authoritative people, can watch all of the quarterbacks in the league free of inherent bias, and apply the same measurement criteria to every QB in the same way, without significant, systematic error (to say nothing of random error).  My lord, talk about non-reproducible results!

Call me a cynic. ;)

You have just described QBR - a statistic you cited approvingly less than 2 hours ago.

QBR used to have heavy subjective elements before Brian Burke was hired away from his own analytics website and onto ESPN's staff.  Now thanks to Burke's influence, the "division of credit" among teammates that used to be done completely subjectively is done in the following manner:


The result is that the only subjective element in that process is the human judgment (made by two people who must agree) of whether the QB was under pressure.

http://www.espn.com/blog/statsinfo/post/_/id/123701/how-is-total-qbr-calculated-we-explain-our-quarterback-rating

That sounds to me like two authoritative people watching all of the quarterbacks in the league free of inherent bias, and applying the same measurement criteria to every QB in the same way, without significant, systematic error (to say nothing of random error).

Isn't the proper term for that "non-reproducible results"?

For a mere one element of the overall statistic, yes, it's subjective and subject to non-reproducibility.  However, here are the many objective components of the statistic that are reproducible, and that no human could possibly consider or apply equally across QBs when evaluating them with eyesight alone:


And when it comes to DVOA, all of that sort of analysis is done, and the strength of the opposing defense is accounted for and applied to the measurement of the quarterback's performance.  No human could come anywhere close to that sophisticated an analysis.

One other point about the bolded portion above -- a statistic can be considered "objective" even if some of its components are weighted on subjective grounds, as long as those subjective weights are applied across all QBs.

QB rating for example has subjectively weighted components, but QB rating is calculated in the same manner for all QBs, which then lets us compare QBs to each other objectively, using QB rating.

Again, I find it very hard to believe that we can ascribe a sufficient degree of reliability and validity to the QB rankings made by an individual or a group of individuals (a la the link in the OP) when there is no way of knowing whether the criteria used to evaluate QBs were applied equally among them.

In other words, a subjective evaluator could indeed bear down and generate a lot of information about just one QB (via film study), but that doesn't allow us to compare QBs to each other with any reliability or validity unless all QBs were evaluated by that evaluator with the same subjective criteria and rigor, free of bias or significant error.  And if we can't determine how or whether that was done, how can any league-wide ranking of them (i.e., the link in the OP) possibly be considered valid?  Its validity is simply unknown and unable to be determined.

This of course is the glaring weakness of the entirely subjective approach, and when the alternatives are statistics as sophisticated as DVOA and QBR, that involve the objective application of so many relevant variables, the choice is a no-brainer in my opinion.

My motive here is simply to know what's accurate and valid about the team, so I don't set myself up for disappointment.  Obviously the best and most rigorous information available is going to facilitate that knowledge as well as possible.

If DVOA and QBR indicated that Tannehill was one of the league's best, I'd consider him such, and go ahead and get my hopes up.  But I'm not getting my hopes up on a false premise so I can be let down later on, when I should've known better.

DVOA and DYAR makes no attempt to distinguish the different components of the offense, except for a split into running plays and passing plays. When applied to quarterbacks, for example, perfectly thrown footballs dropped in the endzone (or on the way into the endzone **cough** Stills **cough**) are counted as negative plays for the quarterbacks. Nothing that actually happens on the play, other than the single line that is generated in the play-by-play, is taken into account. E.g. there is no difference between an interception thrown directly to an opposing receiver and a missed assignment from an offensive lineman allowing a hit by a defender that causes the ball to go way off target. A turnstile "tackle" allowing a blind side sack resulting in a fumble is graded the same as a botched snap or a fumble caused by the quarterback simply dropping the ball untouched. A short screen pass to a running back or tight end that is subsequently taken for 50 yards and a touchdown is graded exactly the same as a perfect backshoulder throw 50 yards downfield.

The result is that DVOA and DYAR are excellent measures of the offense as a whole, but should be applied very cautiously to individual players or components of an offense. Comparing a quarterback+offensive-line+receiver-combo from week to week for a given team is still valid -- and a very useful tool. However, comparing the DVOA of "a quarterback" from one team with the DVOA of "a quarterback" from another team is not likely to yield a valid result. "A quarterback" is in quotes here since no actual attempt is made to isolate the QUARTERBACK's performance from that of the rest of the offense and, as such, the metric isn't really measuring quarterback performance, despite the name.

If you look specifically at quarterbacks, DVOA will result in a overly positive score for quarterbacks who throw high-percentage screens to running backs and tight ends capable of churning out tons of yards after catch. As a whole, YAC becomes extremely important and quarterbacks without receivers capable of generating YAC will receive significantly lower grades than those who have such receivers. All indications are that YAC are mostly dependent on the receiver himself and how he is used, but fairly independent of the quarterback (I note that ESPN has acknowledged this and disregards actual YAC and instead adds an "expected YAC"). Quarterbacks with solid offensive lines will also produce higher DVOA ratings as the "fault" of a sack (and any fumble that might occur) rests fully on the quarterback.

The (subjective) decision has also been made to exclude interceptions (and fumbles) from the "situational awareness" (time remaining, score, down, distance) otherwise used in deriving DVOA, with a few exceptions (notably on fourth down inside 2 minutes and some rather arbitrary adjustments depending on length of throw, punishing interceptions on short throws more than those on long throws). This punishes risk-taking, which would otherwise increase win-probability in games where teams are down multiple touchdowns (outside last two minutes). Interceptions stemming from mistakes by the offensive line are also punished MORE severely than throws that are simply bad quarterback decisions, since hits on the quarterback (or arms) are likely to cause shorter interceptions (shorter "throws") than if the quarterback were not hit and just threw a "clean" interception.

I'm going to make a quick back-of-the-envelope calculation here, partly because I haven't bothered yet, but also because it applies to QBR as well, which I'll get to at some point, although this post is ballooning in length as is.

According to ESPN and other sites (link (http://www.espn.com/blog/statsinfo/post/_/id/110911/a-case-of-the-drops-how-quarterbacks-were-and-werent-hurt)), about 4% of passes are listed as dropped (3.6% for 2015). So far this year there have been a total of 6127 completed passes league wide, resulting in a total of 411 (passing) touchdowns, so 6.7% of all completed passes result in touchdowns. Assuming receiver drops are distributed randomly among passes, a dropped pass would also have a 6.7% chance of being a dropped touchdown pass. League average pass attempts (9666/32 = ) 302 results in about 12 dropped passes per team so far this year. With (12*0.067 = ) 0.8 of those expected to have gone for a touchdown. The Dolphins thus far have 3 dropped touchdown passes, including the Stills drop.

Dropped passes may or may not be a receiver specific, or just bad luck, but Tannehill has clearly gotten the wrong end of that particular stick.

The same goes with every single one of the other deficiencies in DVOA (when used to compare quarterbacks across teams) I listed above.

That's *remarkable*.

I will note that I am a huge fan of Football Outsiders and DVOA, but as the saying goes "lies, damned lies, and statistics". In this case, you have to remember what it is you are actually measuring; and what you are not measuring.

In other words, I honestly believe that DVOA massively underestimates Tannehill's performance this year (and, to a somewhat lesser extent in that I'm assuming he doesn't always suffer from such a ridiculous number of dropped touchdown passes, in other years).

mod edit: fixed broken link

They are objective data.  Measurable calculated data is by definition objective.  Just like shoe size of a WBA is objective data.  What is subjective is if you feel those measurements are a good or bad way to judge a QB.

I appreciate your in-depth analysis, but it's hard to believe that Tannehill's ability is being underestimated a great deal by the most sophisticated statistics we have available when there has been no season in his career that he's been any better than average with regard to them, and instead he's typically been below average with regard to them.

This year may be an anomaly with regard to the ability of DVOA to accurately represent his play, but when you look at how he has fared with regard to DVOA over his career, adjusting for this year's anomaly is likely to bring him from the depths of the league to somewhere between average and below average.

It's very unlikely that making such an adjustment would vault him from the depths of the league to being ranked 10th overall, a la the subjective material in the OP.

Whether or not a quarterback is under pressure is not objective.

Indeed it's not, but it's easier for me to imagine a group of people's being able to define it in such a way that they can reach sufficient agreement about when it's happening and when it isn't, and then apply that definition consistently across QBs and situations, than it is for me to imagine a group of people's being able to take the entirety of QB play and do the same thing.

One thing being measured is relatively narrowly defined, simple, and manageable, and the other is very broad, complex, and dynamic.

I suspect that if you took a sample of let's say 50 passing plays and had three people here decide which ones Tannehill was pressured on, you'd get a far greater degree of agreement than if you asked three people here to simply evaluate Tannehill in his entirety by assigning him a percentile rank among the league's QBs.

^^^

I suspect if one were to put a variety of QB stats out there, without the players' names attached, your opinion of Tannehill would be quite different.

Try me.

You ought to know by now that I've done my homework.

This year is a bit extreme with the statistical outliers of dropped touchdowns, but DVOA consistently underrates quarterbacks with bad offensive lines, quarterbacks with receivers who don't generate YAC, and quarterbacks who play from behind a lot. That isn't specific to Tannehill, although it has plagued Tannehill in each of his first four years. Part of that is personnel, but part is probably also that they keep changing coaches (and therefore blocking schemes, not just for the o-line, but the running backs and tight ends as well).

As for QBR, I disagree with your assessment of it being objective. The biggest issue is that we just don't know enough about how it is generated. I've read all the descriptions on ESPN before, but I've reread them a few times now and there's enough wiggle room to drive a fleet of trucks through in their descriptions. Oh, and you'll appreciate that they frequently refer to QBR as a measure of quarterback efficiency.

Apart from the obvious bit about determining pressure, it's not at all clear how the "division of credit" works. The only substantive statement is that it is based on analysis of thousands and thousands of plays. It might be a fixed ratio (with a boolean depending on pressure or not pressure), but that isn't actually stated. If it isn't fixed, then a whole boatload of subjective judgement comes into play. On the other hand, if it is fixed, it effectively averages out things that are important for the individual plays. The hope is that these aspects of the play average out over a game or a season, but without giving any supporting evidence that's a bit of a leap of faith.

The section on division of credit explicitly states that's what happens on completed passes, leading one to believe that no division of "credit" (blame) happens negative plays. However, some of the older descriptions of QBR (from when ESPN introduced it) included judgement calls on drops, underthrows, and overthrows, among other things. These are clearly subjective to some degree, but also important to include. It's entirely unclear if they are still used in QBR since ESPN won't show us the recipe.

Below is a chart with the 13 quarterbacks who have passed for 13000 or more yards in their first four years (I added Brady to the bunch manually, as he had 3 passes or something his first year). The data is from Pro Football Reference.


(http://i57.photobucket.com/albums/g221/fyodor_/13k-passing-qbs.png)

Sacks, wins and losses give too much away, so that isn't included ;). For reference, about half the players on the list are pretty close to .500 in W/L during their first 4 years.

The quarterbacks are, in alphabetical order:

Drew Bledsoe
Tom Brady
Andy Dalton
Joe Flacco
Jeff Garcia
Andrew Luck
Peyton Manning
Dan Marino
Cam Newton
Carson Palmer
Matt Ryan
Ryan Tannehill
Russell Wilson

Thanks.  This ought to be fun.

I'm well aware of what Tannehill has done in the way of passing yardage, which is a volume statistic, as opposed to an efficiency statistic.

The issues we're going to face in this particular exercise are that 1) passing yardage isn't anywhere near as predictive of winning as passing efficiency (YPA, or Y/A as it's referred to in the table) or overall play (QB rating), and 2) we have a mixture of time periods, both before and after 2004, when the rules were changed to favor the passing game.

The average QB rating in the league has steadily increased over the years, and jumped a great deal after 2004.  This year for example it's almost 90.

What stands out to me in that table is that you have an absolute stud with a QB rating of 101.8 and a YPA of roughly 8.  I know who that is, and I know he compiled those statistics after 2004.  Had he done it prior to 2004 it would be mind-boggling.

But what is the recipe used by the folks in the OP, and how do we know it was applied consistently?  If it's the details of the recipe you're after, are you truly comfortable ascribing more validity to their conclusions than you are to the conclusions arrived at via QBR and DVOA, given the knowledge we do have of how they're formulated?

The other issue we have with the material in the original post is that, despite Tannehill's being ranked 10th overall among quarterbacks, he isn't a standard deviation better than the average of the top 32 QBs listed along any of the dimensions by which they were rated.

According to the website, those dimensions were the following:


Tannehill's Z-score (or in plain terms, the extent to which he deviated from the average of the top 32 QBs listed) for each of those dimensions was the following -- a value of 1.0 or more represents at least a standard deviation higher than the average:

Accuracy:  0.99
Arm Strength:  0.79
Pressure:  -0.02
Decision-Making:  0.06
Overall:  0.58

So despite being ranked 10th overall (and rankings can provide misleading information for just this reason), Tannehill is not significantly better (as defined by at least a standard deviation, in my view) than the average quarterback among the top 32 listed in any area in which they were rated by these supposed authorities.

(Granted he's very close when it comes to "Accuracy," and so close in fact that we'll go ahead and give him that.  He's also fairly close in the area of "Arm Strength."  He isn't close "Overall," however.)

This sort of finding is much more consistent than the 10th overall ranking would make him appear with regard to how he's been evaluated by the sorts of sophisticated statistics we've discussed in this thread.  Again, the 10th overall ranking can be misleading in this case.

In fact, the only QBs among the top 32 listed who were at least a standard deviation better than the average of the group with regard to the "Overall" rating were the following:

Aaron Rodgers
Cam Newton
Sam Bradford
Tom Brady
Philip Rivers
Drew Brees
Ben Roethlisberger
Andrew Luck

Everyone else among the top 32 was either non-significantly different from the league average, or significantly below it.

There are actually TWO guys that stand out in that table, significantly higher than the others in both passer rating and YPA. One is the one you're thinking of, of course, the other compiled his stats quite a while before 2004, making them frankly astonishing.

As for yards being a volume statistic, I'm well aware of that. What I was looking at was quarterbacks who were asked to (and managed) throw a whole heck of a lot in the beginning of their careers. Quarterbacks that were, one way or another, asked to deliver far more than your average NFL quarterback. In that company, Tannehill doesn't really stand out much, neither good nor bad. What can be said is that the rest of those guys either had great careers or look like they will have great careers. Not necessarily HOF careers, of course, but still pretty darn good careers. Looking at the guys who are done with theirs, there's really one that sticks out in a bad way and he's perhaps the "bottom" projection for Tannehill. As long as we get a guy like the one who followed that guy, I'm fine with that ;).

Is number 11 Rodgers?  I would rank 11 the best than 13.  Even though 1 and 2 has a high TD and rate I would rate them lower than most because my subjective opinion is int% is the most important stat. I would prefer a player like 5 or 9 over 1 or 2

Rodgers didn't make the list for lack of production. Only players with more than 13k yards in their first 4 years (5 for Brady) are on it.

If you WERE to take Rodgers and remove his pre-starter years, he'd be so far ahead of everyone else it's ridiculous. 17k yards, interception rate of 1.8%, touchdown rate of 6.4%, 105 passer rating.

Your emphasis on low interception rate made me think of Brett Favre. He had a first year like Brady where he didn't really play and so didn't make the list. Had I though of him, I'd have added him manually, even though he doesn't come with a pun like Brady did.

For reference, Favre would have slotted in at #6 with nearly identical stats to #5 across the board.

Considering my original goal was to look at rookie quarterbacks asked to throw insane amounts in the first part of their career, perhaps I shouldn't have added Brady (or any others who sat their first year). Certainly, spending 3 years developing behind Favre isn't equivalent to starting from day 1.

Oh, and #13 on the list is a bit of an odd player, since his first NFL season came when he was 29...

#13 has to be Weinke or Weeden right?

One statistic that should be very valuable in theory is the ratio of YPA to INT rate, given that YPA in theory is partly a measure of risk-taking in the passing game, and INTs are more likely as a function of increased risk-taking.

The resulting measure would indicate how well a QB is able to exhibit the best of both worlds of the game manager (the guy who doesn't lose games by throwing INTs), and the game winner (the guy who takes sufficient risk to lead the team to victory with his arm).

What we've seen from Tannehill thus far this year was increased aggressiveness through the first five games, which led to an inordinately high INT rate for him (seven in five games), followed by the reduction of his role to that of a game manager over the last three, which has produced a three-game win streak.

Theoretically Tannehill's ratio of YPA to INT rate would be smaller than that of someone like Russell Wilson or Aaron Rodgers, who play more aggressively in the passing game but nonetheless throw interceptions at a relatively low rate.

With Tannehill on the other hand, you have to rein him in to being a game manager so you don't lose games via his turnovers.

Dolfanalyst, what is your verdict on Tannehill's position on that chart?

Weinke had 1 year as a starter, and didn't manage even the average yardage needed that year.

Weeden hasn't started a full season yet, getting close his first season (15 games), bouncing around the league the following 3 years, before flushing out.

I'm unable to see the chart where I am right now, but assuming Tannehill is at or near the top, and assuming the chart is ordered according to passing yardage, his position on the chart is a function of having thrown a comparatively higher number of passes.

This is what happens when you have a comparatively low YPA -- you can still amass a high degree of yardage, but it takes you more passes to do so, obviously.

And the problem there, again, is that YPA is far more strongly correlated with winning than is passing yardage.  Efficiency is much more important than volume in this area.

That's more than a bit disingenuous considering that more than 29 players in a 43 player population are, by definition, within +-1 standard deviation if you assume the population follows a normal distribution.

Ergo, one would expect everyone from about 8 to 36 to be between +-1 standard deviation.

If your claim is all players from 8 to 36 on the list are essentially interchangeable (in terms of performance), then you are, quite frankly, nuts.


Sometimes I think you have a good grasp of statistics, but then you write stuff like this. BY YOUR VERY OWN DEFINITION, with a list of 43 players, all but about 7 or 8 will be with below one standard deviation above the average.

Circular logic much?

First, I'm stopping at the top 32 players, because there are 32 teams in the league.


You don't think there are distributions of human variables in the world in which most people cluster around the average, and there are relatively small tails of the distribution in either direction?  Someone has to be "nuts" to believe that?

This is essentially "the bell curve," which is the most common frequency distribution found in the world! :)


I think the problem here is your grasp of statistics if you truly believe it's impossible theoretically that the distribution of QB ability/performance in the NFL is such that there are eight starting QBs who are significantly above average, 16 who are within the average range (less than 1 SD in either direction from the mean), and another 8 who are below average.

Do you really believe that? ???

(http://www.updconsulting.com/wp-content/uploads/2010/10/graph-1-01.jpg?w=300)

I would just like to go on record as saying that the only ranking I give a crap about is where the team ranks in the standings.   ;D

I don't consider YPA to be nearly as important as int rate.  You throw one int and the drive is over.  Unless you are behind and it is late in the fourth quarter it doesn't matter if you can get down the field in 2 plays or it takes 20.  In fact taking 20 plays to get down the field when you are up is better than doing it in 2.

Granted it is a subjective view, but my view is protecting the ball is most important.  I would take a RB that avgs 3 yards per rush and fumbles once per 40 carries over someone who avgs 4 yard per carry and fumbles once every 35 carries.   

Turnover margin is one of the most important variables in football.  In fact, teams that win the turnover battle and the YPA battle (between QBs) win 95% of the games in the NFL.

The Dolphins beat the Jets last week largely because of the 2-0 turnover margin they enjoyed, because they surely didn't win the YPA battle.  In fact they lost it by a full 2 yards, which is a very large margin in that area.

What percentage of games do you win if you win the turnover battle but lose YPA?  I am willing to bet it is north of 70%

Nice image, though a bit on the large side...

Anyway, I'm far more familiar with gaussian functions that I could possibly wish for. But thanks.

You were being completely disingenuous (that's giving you the benefit of the doubt) in the way you presented your data. As #10 in a normal population of 43, Tannehill is by definition not more than 1 standard deviation above the mean. Yet your presentation made this seem like a shock and proof that Tannehill really wasn't significantly better than average. Not that this prevented you from expressing disbelief at the very notion that he could be #10 to begin with, of course.

Let me be very, very clear here: You do NOT need to be more than 1 standard deviation better than average for it to be significant in sports. As a trivial proof of this, take a group of runners. It doesn't matter HOW MUCH better you are than the guy behind you, a fraction of a second is all it takes. Certainly, there are individual variances that come into play and football is more complex than running, but teams filled with +1 SD players would crush teams filled with AVERAGE players. Not every game, but in the VAST majority of games. Certainly not insignificant ;).

So you're saying that it doesn't really matter that much that Tannehill has a relatively pedestrian YPA since he has such a low INT%? ;)

Then you should've done the analysis I just did when looking at the data to begin with, and determined for yourself that Tannehill, despite being ranked 10th overall, was nonetheless in the average range.


Your point here rests on the idea that the sample should extend beyond 32 quarterbacks.

Surely if we do that, then yes, the sample will include more than just the typical number in the league who are significantly below average, which will in turn make the ones above average appear better by decreasing the overall mean of the group.

But is that what you're saying?  We should include backup QBs in the analysis, as well?

Actually what you've just described are the grounds for relegating a QB to the role of a game manager, which is precisely what the "QB whisperer" coach the front office is paying so much money has done after watching him play a mere five regular season games.

No, that's incorrect. I don't dispute that performance most likely follows a normal distribution. That's the assumption we all make when dealing with these numbers all the time. Thus it's completely ridiculous for you to come out and say on the one hand that, well, he couldn't possibly be the 10th best. And then, on the other hand, say that, well, 10th best is really not significantly different than average.


No. No. No.

I'm saying that if you're 1 standard deviation behind Usain Bolt, you may be pretty f'n fast, but you'll get smoked every time.

If we take the 32 starting QBs 8 will be in the top 25%, 8 in the bottom and 16 would be in "average range". 10 would in fact be the second best among the 16 average qbs

I didn't say he couldn't possibly be the 10th best.  What I said was that the ranking is inconsistent with the statistics we've mentioned in the thread, and that there is an insufficient basis for deeming the ranking to be valid, since we're unaware of the methodology used and whether it was applied consistently across quarterbacks.

The issue that 10th best isn't significantly better than the average QB, when using that ranking information on the webpage, is yet another issue to contend with, in addition to the others mentioned.


Certainly.  And likewise, if you're 1 standard deviation behind the best QB in the league, you may be pretty f'n good, but you won't give your team nearly the probability of winning.

And now you're back into rankings (i.e., "second-best"), which again can be misleading.

The proper way to do this is to determine how deviant from the league norm Tannehill is.

In the case of these rankings of overall performance, he's little more than a half a standard deviation (0.58) from the mean of the top 32 QBs ranked.  That ain't all that great.

If we're assuming that the population follows a normal distribution for the metric we're measuring (which is likely the case), then rankings map directly to z-scores. You calculate the standard deviation with the explicit assumption that the distribution is normal. If the rankings DON'T map directly to z-scores (in a significant way), then the distribution likely wasn't normal to begin with.

In other words, for a normal distribution rank and z-score are equivalent, by definition, so rank isn't any more or less misleading than looking at standard deviations.

For such small populations (e.g. 32 teams) there are bound to be slight irregularities, so the mapping may not be completely exact, although I have to say it fits really well for most of the metrics I've looked at.

Assuming similar variances in individual performance, being half or a full standard deviation above average results in the same win-probability as being half or a full standard deviation behind the best. In other words, half a standard deviation is very significant. Your specifically stated that any difference of less than a full standard deviation was insignificant. That is something that I disagree with completely and utterly -- and, your original statement to the contrary, you seem have come around (see quoted text above). If so, let's just leave the statistical aspects of this discussion behind. I have more problems with ESPN's QBR and am not at all convinced it is the least bit more objective than the scores I linked in the OP. These issues are systemic and, in my opinion, largely invalidate QBR as a meaningful metric when comparing dissimilar quarterbacks. I'll leave it at that teaser for now, though. I'm sure QBR will pop up again soon ;).

The problem is that the rankings are perceived by the average fan to be interval data, with no knowledge of the size of the intervals between each datum.

Really what you're saying here highlights yet another problem with these rankings, in that we don't know how they correlate with winning, which would determine the expected increase in win probability for every SD increase in the measure used to make the ranking.

What I've been doing here is generously superimposing the correlations with winning we know exist for other measures (which are in the .50 range, give or take), and which give meaning to 1 SD changes either way in QB statistics.

Hell, these rankings could be correlated with winning at 0.00, in which case a change of a standard deviation (or more) would mean nothing!  Or, they could be correlated with winning at 1.0, in which case much smaller differences in SD would be associated with far bigger increases (or decreases) in win probability.

Either way, we would need to know how any measure correlates with winning to determine the meaning of a 1 SD change in the measure in either direction.

Well, then let that be a lesson to "the average fan" ;)

(The point in your subsequent post opens a whole 'nother can o' worms and I'm just going to leave it at "very small sample size" for the time being.)

So Fyo when ya revealing who is who?

That should be after Dolfanalyst announces his verdict for Tannehill's position on the chart, right?

I think he's already said what he's going to say about that, so I'll go ahead and put up the uncensored image:

(http://i57.photobucket.com/albums/g221/fyodor_/top-passing-qbs-uncensored.png)

The takeaways for me are:

- Dan Marino was really good. I know we all know this, but to compile those stats in that era is astonishing.
- Russell Wilson has gotten off to one heck of a start to his career. I don't get to see that many Seahawks games, but maybe I should.
- It's just fun to see a list headed by Peyton Manning, Dan Marino, and Ryan Tannehill.
- Quarterbacks that were asked to throw a BOATLOAD from day 1 generally turned out to be very good quarterbacks.

This sounds like it's reached the "just for fun" stage, and just for fun I'll try to give this a little more perspective here.

During the Tannehill era (2012 to 2016), teams with a QB rating between 84.2 and 86.2 (roughly his career QB rating) are 36-45 -- a 45.5 win percentage.  That translates to roughly a 7-9 regular season record.

During the same era, teams with a QB rating between 100.8 and 102.8 (roughly Russell Wilson's career QB rating) are 39-22 -- a 63.5 win percentage.  That translates to roughly a 10-6 regular season record.

During the Dan Marino era alluded to here (1983 to 1986), teams with a QB rating between 94.2 and 96.2 (roughly Marino's overall rating during that period) were 36-15 -- a 70.6 win percentage.  That translates to roughly an 11-5 regular season record.

Yet Tannehill has only slightly less passing yardage than Marino did during these time periods, and a good bit more than Wilson.

So again, if we focus exclusively on (or said differently, cherry-pick -- and I'm not saying that was fyo's intention here) passing yardage alone, we can easily go astray in terms of what's most important with regard to winning, and come away with a misperception of the quality of a QB.

Again, my intent here (and in all of these sorts of analytical discussions about Tannehill and QBs) is to refrain from setting myself up for disappointment by making a misguided conclusion in this way.

My takeaway Dolphin fans that hate on Tannehill should STFU.  Is he the single best QB in the history of the NFL? Absolutely not.

But, he does match up pretty favorablely with other very good/great QB. 

A common refrain among those making the case for Marino was that fins didn't have much of a running game and put the game on Dan.  Well nobody on this list has more attempts, that speaks volumes about what share of the workload falls to Ryan each game.

He has a lower int % than Brady, a player known for his focus on ball security.

Reaffirms my opinion that Bledslow wasn't all that great.  Clearly the worst of the list. Tannehill is head and shoulders better than a QB that won the Patriots two AFCCGs

How so?

Unless your one and only stat you look at is y/a he looks good.  He has a better passer rating than Manning he has a better int rate than Brady or Marino.  Three guys that are either in the hall or will be.

You could say the same things about Jeff Garcia, and a whole host of other QBs who went on to have mediocre or worse careers, but who aren't included in that table because they didn't pass for enough yardage during their first four years in the league.

Four time probowl is mediocre or worse carrear, keeping in mind the majority of starters never go even once.  Your problem is you are setting the bar way too high.  He is not in the Peyton, Rodgers, Brady class of QBs.  But he is very good.  Probably the second best QB the Dolphins have ever had.

I'm setting the bar that high because quarterback play is more important now than it was in the days of Dan Marino, Drew Bledsoe, and Peyton Manning's first four years.

In fact there are folks who believe the parity in the league has been significantly undermined by the rule changes that favor the passing game, which make the teams with the best QBs much more likely to win than the other teams in the league.

And parity is what made this league grow in popularity by leaps and bounds.  When the team you root for can go 6-10 one year and make the playoffs and have a chance at the Super Bowl the next, your interest and investment as a fan is much greater than if you're relatively hopeless about your team's chances for success.

If the game tends to revolve around just one player (the QB), and that player on your team is inadequate, it's far easier to become hopeless and lose interest.

This is probably why we discuss Tannehill so much!  On some level we're aware that our fortunes rest largely on this one guy, and we're all doing our best to stay hopeful and remain interested.

I guess that reliance on just one guy is why NEP went 0-4 during Brady's suspension and had a losing season in 2008 when he was injured....ummm scratch that.

I guess no point in really discussing football with you if you feel Tannihill is mediocre or worse and inadequate and/or believe that the QB is the only position that matters.

Tannehill is a very good but not great QB, Seahawks and Giants have both recently proved you can win a Super Bowl with a good qb when facing a team with a great one.

It sounds like the differences in the ways we think about these things would make any further discussion unproductive.  Thanks anyway, however. :)

I had more time this morning and went ahead and calculated the correlation between win percentage and the "Overall" QB measure on the page linked in the OP.  That correlation is 0.32.

What that means is that it would take lots of movement (up or down) by a QB with regard to that measure to produce an appreciable effect on winning.

In fact, based on the strength of the relationship between this QB measure and winning (again a correlation of 0.32), the Dolphins' record would be roughly the same as it is now (4-4) if Tannehill were ranked anywhere between #3 and #21 in the league on that list.

He would have to be ranked #1 or #2 in the league, or be ranked #22 or lower, to have produced a different Dolphins record (better or worse) thus far this season.  Every ranking in between -- according to this measure -- would likely be associated with the same record the team has now.

So, if you happen to be sold on this measure, realize that Tannehill is going to have to somehow climb to being measured as either the best or the second-best QB in the league to produce a meaningful, positive effect on the Dolphins' record.

^ That is is many ways unsurprising. Since the other measures of QB skill include other effects, like receiver quality (YPA, drops), offensive line, etc., a change in those measures represents an improvement across all those components. Improving quarterback + offensive line + receivers is, unsurprisingly, going to result in a stronger overall improvement (and thus correlation with winning) than simply changing any one of those components.

The conclusion, if one believes the quarterback performance metric to be valid, is that simply changing quarterbacks isn't likely to result in a significant improvement overall for the Dolphins, unless that quarterback is elite.

That seems a perfectly reasonable conclusion, IMHO.

Only in this case, "elite" is defined as comprising only two of the league's quarterbacks, and as of now those quarterbacks -- according to this measure -- are Aaron Rodgers and Cam Newton.

They are not Matt Ryan, Tom Brady, Russell Wilson, Derek Carr, Ben Roethlisberger, Drew Brees, or Philip Rivers.

In fact, according to this measure, in terms of a QB's effect on his team's win percentage, Ryan Tannehill is interchangeable with those folks.

So to summarize.....there are only two QBs active in the NFL that if replaced Tannehill would change the win percent for the Dolphins....this leads to one of two conclusions either Tannehill ain't the problem and fans should stop hating on him or said fan has unreasonable expectation that anything but the top 2 qbs are acceptable.

If by your own analysis you are saying the fins and pats traded Brady and Tannihill the Pats would still be 7-1 and Fins would still be 4-4 you should look elsewhere to improve the team.

...or the measure of quarterback play used to determine all of that is actually invalid, and consequently none of that is true.

I'm a bit pressed for time this weekend, but I took a quick look at your correlation calculations and wasn't completely able to reproduce them. I seem to recall you noting that you limited your analysis to 32, since that's the number of teams. I ran the numbers using the "top" quarterback for each team with multiple quarterbacks listed, while keeping the total win% (alternative would be to either discard games or separate e.g. NE into two, the latter being the better choice, IMHO, but then I'd have to look up a bunch of games and I don't have the time right now). Anyway, wasn't able to completely reproduce your results for the standings as they were before Thursday's game.

There's actually a bit of an interesting situation this year, which highlights the uncertainty when running statistics on what are fairly small sample sets: The correlation between SCORING and WINNING, which is historically a fairly solid 0.7 (slightly higher than even turnover margin), This season (prior to the TNF game) it's at about 0.4.

For reference, the correlation between scoring and the quarterback totals in the linked article (OP) is 0.57.

Here are the data I used to make the calculation:

Player   Team   Overall   WIN %
Aaron Rodgers   GB   80.7   0.5
Cam Newton   CAR   79.3   0.375
Sam Bradford   MIN   78.8   0.625
Tom Brady   NE   78.3   0.875
Philip Rivers   SD   78   0.444
Drew Brees   NO   77.9   0.5
Ben Roethlisberger   PIT   77.2   0.5
Andrew Luck   IND   76.8   0.444
Matt Ryan   ATL   74.6   0.667
Ryan Tannehill   MIA   74.6   0.5
Eli Manning   NYG   73.4   0.625
Tyrod Taylor   BUF   73.3   0.444
Matthew Stafford   DET   73   0.556
Carson Palmer   AZ   73   0.438
Dak Prescott   DAL   73   0.875
Derek Carr   OAK   71.9   0.778
Russell Wilson   SEA   71.7   0.688
Jay Cutler   CHI   71.7   0.25
Marcus Mariota   TEN   71.6   0.444
Alex Smith   KC   71.3   0.75
Jameis Winston   TB   69.1   0.375
Carson Wentz   PHI   68.9   0.5
Colin Kaepernick   SF   68   0.125
Trevor Siemian   DEN   67.1   0.667
Brian Hoyer   CHI   66.3   0.25
Joe Flacco   BAL   66.3   0.556
Cody Kessler   CLE   66   0
Andy Dalton   CIN   65.6   0.438
Case Keenum   LA   64.4   0.375
Kirk Cousins   WAS   63.3   0.563
Brock Osweiler   HOU   60.9   0.625
Ryan Fitzpatrick   NYJ   59.6   0.333
Blake Bortles   JAX   61.9   0.25

I did include both Jay Cutler and Brian Hoyer, but if I remove Hoyer and stick with only the "top" QB as you did, then the correlation as I calculate it actually drops from 0.32 to 0.29.

The partial correlation between the "Overall" QB measure in the OP and win percentage, controlling for team points scored per game, is 0.06.

...Or quite simply stats are merely just stats and do not adequately indicate QB play nor do they adequately predict success or failure.

If that's true, then what does it mean that Peyton Manning has the most passing yardage and touchdown passes in the history of the league?  What does it mean that Aaron Rodgers has the highest career QB rating of all time in the NFL?  What does it mean that Dan Marino threw for over 5,000 yards and 48 TDs at a time when that was unheard of?  Nothing?  Are these statistics meaningless, and not indicative of these players' talent?

Don't throw the baby out with the bathwater.  Just because one statistic is garbage doesn't mean they all are.

That's hardly surprising, since the quarterback doesn't play defense.

If you'd stop skipping the descriptions of what these statistical tests do, you might recall that the what "controlling for team points scored per game" means is that what is calculated is the correlation between winning and QB performance while keeping points scored constant. In other words, if it doesn't result in more points scored, the performance of the quarterback doesn't matter much in winning games. I think that's a reasonable statement that most would agree with and does not preclude a causal relationship between quarterback performance and winning.

That was exactly my point, however, that despite the diminished relationship between scoring and winning this year that you pointed out in the post of yours I quoted, the relationship between QB play and winning, while controlling for team points scored, is what would be expected.

In other words, QBs are still doing what they do with regard to winning, despite that points scored and winning are less related than usual.

The overall point is that we should still expect a measure of QB play to be correlated with winning, despite that points scored are less correlated with winning than usual.  Thus, the QB measure in the OP doesn't "get a pass" this year in terms of determining its validity.  There should still be a relationship between QB play and winning expected, as in other years.

I disagree. If the result of QB play is, primarily, to score points, then the correlation is entirely subsumed by the correlation between scoring and winning. I.e. if A -> B -> C (A = QB play, B scoring, C winning), then any reduction in correlation strength between B and C will result in a reduction between A and C as well.

The overall point, as I see it, is that we're simply dealing with small sample sizes. Scoring points hasn't suddenly become less important than in previous years.

Anybody that WATCHED Marino play knew that he was special. Same goes for Peyton Manning and Aaron Rodgers. You don't need stats to tell you this and what the stats DON'T tell you is WHY they are special. Marino wasn't special because he threw a lot of TD passes, the reverse is true he threw a lot of TD passes because he was special and the reason he was special was because of the fire in his gut, his ability to elevate his play, his desire to be the best and of course his god given talents as well. Ask Marino how to play QB and he'll tell you "pick a guy a let it fly". That's not a joke, that's the way he played from the seat of his pants. He wasn't some master tactician, the game just made sense to him, he just saw things happening before they even happened. I think there's a lot of truth to that in Peyton Manning as well, it's well documented his ability to break down tape and study opponents, but in the heat of the moment, you don't think, you just act. It's instinct. These are things you can't teach, some people are just born to be QB's, but most are not. For those that aren't they can achieve a level of play through a lot of hard work and repetition, but they will NEVER be like Marino and Manning because they can't. It's just not in them and stats won't tell you that, watching them play and realizing that the stats don't define these players, that they transcend the stats will.

Do you think that because it's possible to see how well (or not well) a QB can play, the statistics that he compiles are meaningless?

Surely you don't think it has to be one or the other, that either eyesight or statistics are valid, and not both?

The issue there (the bolded portion) is that you're proposing a very simple causal path that discounts things such as YPA, interception rate, third down conversion rate, time of possession generated by sustained drives, etc., that have less than isomorphic relationships with scoring, but nonetheless have, especially when aggregated, significant relationships with winning, in that they prevent opposing teams from scoring.  Consequently you can have causal paths between QB play and winning that don't involve scoring.

The overall issue really comes down to whether you believe the interchangeability between Tannehill and Matt Ryan, Ben Roethlisberger, Drew Brees, Derek Carr, Philip Rivers, and Tom Brady implied by the QB measure in the OP is due to 1) a decrease in the relationship between scoring and winning in 2016, or 2) the simple invalidity of the QB measure, because the QB measure involves the same sample size (eight or nine games) as does the diminished relationship between scoring and winning in 2016.

Are there elements that the offense (and the quarterback) can provide that help winning without scoring points? Sure, but there's nothing that even remotely compares in magnitude to SCORING. It is far, far away the most important aspect of an offense. Everything else is nice, but the effect is hard to pick out from the swamping signal of points scored.

That is some serious GIGO going on.  You use Brady's qbr but the teams win % including games he wasn't even at the stadium.  Any conclusion drawn from that is pure and utter trash.

Well, feel free to do the ultra-time consuming work of paring down win percentage by QB starts, for all the teams in the league, free of charge, and we'll make sure and sit here in the bleachers and critique your work. :D

Not wasting my time. just because you did a bogus analysis doesn't require me to correct it, and even so running the data on half a season would be statistically meaningless even if done correctly. 

No, it doesn't require you to do anything, but you might want to show a modicum of appreciation for the folks putting the work in and doing the best they can with it, rather than sitting up on your perch and providing only blowhard critique.

But you didn't do anything worth appreciating, using bogus data to hate on Tannihill when the REAL data says he is not the problem, is not worthy of appreciation.

Again, I think the differences in the ways we think about these things are going to make any further discussion unproductive. :)

Perfect example here today of how what you see and the statistics that represent it can be one and the same.

Tannehill's having one of the best games of his career, and his YPA is a whopping 10.2, and his QB rating 146.1 at the end of the third quarter.

If anyone's interested, Tannehill's game against San Diego was the 10th-best game of the year in terms of ESPN's Total QBR, which adjusts for the strength of the opposing defense.

http://www.espn.com/nfl/qbr/_/type/player-game