Bill Barnwell snipes at Allen, and gets shredded in response

[The Frankish Reich]

6 hours ago, JoPoy88 said:

For those curious:

 

“Total QBR was developed by a team at ESPN Stats & Information Group including Jeff Bennett, Dean Oliver, Alok Pattani, Albert Larcada, and Menlo College professor Ben Alamar. The group also received input from ESPN analysts Trent Dilfer, Jon Gruden, and Ron Jaworski. Total QBR was developed based on analysis of 60,000 NFL plays between 2008-2010, and was unveiled on August 5, 2011. The formula was modified in 2012 and again in 2013.”

It's not a very good measure, but I actually applaud the effort. Folks: don't be so defensive about Josh! He's moving up in all of these rating systems.

-Traditional QB rating: now 22nd overall as Barnwell says, but look who's in a virtual tie with him: none other than Mr. Brady himself.

-ESPN's Total QBR: now 24th. The formula obviously puts way too much weight on completion percentage and way too little on rushing yardage - Mahomes is basically tied with Lamar Jackson for the lead. 

- fivethirtyeight.com's ELO rating: Josh is ahead of TB12!  OK, this is my favorite "one stat captures everything" metric, and not just because Josh is in the lead. Overall he ranks behind only Jackson, Mahomes, Rodgers, Wilson, Prescott, Stafford, Watson. Generally fits with my eyeball test ...

 

[DCOrange]

3 hours ago, HappyDays said:

 

He was terrible against the Pats but clearly better than last year in every other game. In every other conceivable stat - passer rating, YPG, ANY/A, DVOA, YPA, sack percentage, TD%, INT%, completion % - he has improved this year. The only thing that has gotten worse is his rush YPG and YPC. This leads me to believe that QBR way over values rushing from a QB. There is no other explanation for that metric saying he is worse this year than last year. Mind boggling stuff.

Nah, that's frankly incorrect. The Jets game was terrible. Eagles and Titans were bad. Bengals was pretty bad. And the Patriots he pretty much singlehandedly lost the game. He was also quite good when he returned from his injury last season.

 

He's certainly been better statistically this season, and the past 7 games for him have been a significant step up from what he's previously been, but the beginning of the year still happened and he was bad/sometimes terrible then.

Edited November 30, 2019 by DCOrange

[thebandit27]

30 minutes ago, DCOrange said:

Nah, that's frankly incorrect. The Jets game was terrible. Eagles and Titans were bad. Bengals was pretty bad. And the Patriots he pretty much singlehandedly lost the game. He was also quite good when he returned from his injury last season.

 

He's certainly been better statistically this season, and the past 7 games for him have been a significant step up from what he's previously been, but the beginning of the year still happened and he was bad/sometimes terrible then.

 

This isn't far off from your Calvin Ridley take, just FYI.

 

And I only reference it because it's the only other time where I remember you being just way out in left field 

[HappyDays]

32 minutes ago, DCOrange said:

The Jets game was terrible. Eagles and Titans were bad. Bengals was pretty bad.

 

I don't agree with any of this. There were games last year, especially early on, where he looked like he didn't belong in the NFL. He looked like that once this year. You're not going to convince anyone who has watched every snap he's taken that he is overall worse this season than he was last season. It's an indefensible position both statistically and on the tape. I never paid much attention to QBR but this seals it for me. I don't trust any stat that gets it that wrong.

[BADOLBILZ]

3 hours ago, dave mcbride said:

 

His ypc is down in large part because of the increased rate of kneel downs. For real.

 

 

And dives for short yardage.   They have been numerous and have had them on back-to-back plays in the Dallas game(though the second wasn't intended to be one).

[Mike in Horseheads]

Hmmm that link didn't work for me, wonder if I'm blocked. Doesn't matter the guy trashes the Bills and Allen and everyone of his ? takes.

[DC Tom]

15 hours ago, DC Tom said:

 

Because QBR is garbage.

 

Any "statistical measure" that includes "clutch-weighted expected points added on penalties" is a joke.  What the hell does that even mean?  Clutch expected penalties?

 

It's such a stupid statistical measure that I just described it to my wife, whose math skills don't go beyond "2+2=potato," and her response was "What????"

[stinky finger]

23 hours ago, Binghamton Beast said:

 

I'm not wearing my glasses. 

 

Is this the neck of a 12 string gee-tar?

[DCOrange]

19 minutes ago, HappyDays said:

 

I don't agree with any of this. There were games last year, especially early on, where he looked like he didn't belong in the NFL. He looked like that once this year. You're not going to convince anyone who has watched every snap he's taken that he is overall worse this season than he was last season. It's an indefensible position both statistically and on the tape. I never paid much attention to QBR but this seals it for me. I don't trust any stat that gets it that wrong.

If you took just his past 7 games, he’d be slotted in as the 13th best QB in the league. I think that’s an accurate barometer of how he’s playing right now. But those first few games in which he either cost the team the game or tried to lose it for 3 quarters before saving the day still happened and that’s weighing his rating down at the moment. If he continues playing the way he has recently, his QBR will be much higher this season than it was last year. 

[The Frankish Reich]

16 hours ago, DC Tom said:

 

Because QBR is garbage.

 

Any "statistical measure" that includes "clutch-weighted expected points added on penalties" is a joke.  What the hell does that even mean?  Clutch expected penalties?

The concept isn't stupid at all. Let's say QB A, over the course of a season (beware small sample sizes!), draws defensive PI penalties at a higher than QB B. Let's say that this amounts to one additional PI call per game, with the average PI penalty being for 20 yards. At the end of the season, it's as if QB A passed for 16 x 20 = 320 more yards than QB B. (Notice how easy I kept the math ... even your wife would get this.) That's significant, and that's not being measured by traditional "counting" stats (total passing yards, etc.) or even by some rate stats (completion percentage; after all, QB A essentially "completed" 16 more passes in the season). 

The execution is a different matter. Here football stats are in their infancy compared to baseball stats; part of that is the inherently individualistic nature of baseball (hitter vs. pitcher) as opposed to football. So the question then is this: "is drawing PI penalties a repeatable skill for QBs, or is it essentially luck, or a combination of luck and having good receivers?" That's what we find out when measuring things year-by-year. If QB A consistently draws defensive PI penalties at a significantly higher rate year after year after year, we call that a "skill." A good example from baseball is "pitch framing" stats for catchers in baseball. We've learned that certain catchers are way, way better (or worse) than average in "stealing" called strikes from umpires; it's not sheer luck; it's a skill. It changed how MLB GMs evaluated catchers; guys that were significantly better than average at pitch framing add value, sometimes the equivalent of a few wins a year when compared to their poorer competitors, and free agent salaries began to reflect that.

Math anxiety is not your friend; it is something to be overcome.

[DC Tom]

2 minutes ago, The Frankish Reich said:

The concept isn't stupid at all. Let's say QB A, over the course of a season (beware small sample sizes!), draws defensive PI penalties at a higher than QB B. Let's say that this amounts to one additional PI call per game, with the average PI penalty being for 20 yards. At the end of the season, it's as if QB A passed for 16 x 20 = 320 more yards than QB B. (Notice how easy I kept the math ... even your wife would get this.) That's significant, and that's not being measured by traditional "counting" stats (total passing yards, etc.) or even by some rate stats (completion percentage; after all, QB A essentially "completed" 16 more passes in the season). 

The execution is a different matter. Here football stats are in their infancy compared to baseball stats; part of that is the inherently individualistic nature of baseball (hitter vs. pitcher) as opposed to football. So the question then is this: "is drawing PI penalties a repeatable skill for QBs, or is it essentially luck, or a combination of luck and having good receivers?" That's what we find out when measuring things year-by-year. If QB A consistently draws defensive PI penalties at a significantly higher rate year after year after year, we call that a "skill." A good example from baseball is "pitch framing" stats for catchers in baseball. We've learned that certain catchers are way, way better (or worse) than average in "stealing" called strikes from umpires; it's not sheer luck; it's a skill. It changed how MLB GMs evaluated catchers; guys that were significantly better than average at pitch framing add value, sometimes the equivalent of a few wins a year when compared to their poorer competitors, and free agent salaries began to reflect that.

Math anxiety is not your friend; it is something to be overcome.

 

You're an idiot.  Not the least of which for telling a physicist he has math anxiety.

 

I'll point out the considerable flaw in your above "analysis" later, when I'm not typing on a phone.  Or maybe I won't, and just let you live with your head up your ass.  

[The Frankish Reich]

2 minutes ago, DC Tom said:

 

You're an idiot.  Not the least of which for telling a physicist he has math anxiety.

 

I'll point out the considerable flaw in your above "analysis" later, when I'm not typing on a phone.  Or maybe I won't, and just let you live with your head up your ass.  

Well maybe it's statistical analysis anxiety. I eagerly await your full keyboard response.

[DC Tom]

2 minutes ago, The Frankish Reich said:

Well maybe it's statistical analysis anxiety. I eagerly await your full keyboard response.

 

I've published five papers on statistical physics. 

[RaoulDuke79]

26 minutes ago, The Frankish Reich said:

The concept isn't stupid at all. Let's say QB A, over the course of a season (beware small sample sizes!), draws defensive PI penalties at a higher than QB B. Let's say that this amounts to one additional PI call per game, with the average PI penalty being for 20 yards. At the end of the season, it's as if QB A passed for 16 x 20 = 320 more yards than QB B. (Notice how easy I kept the math ... even your wife would get this.) That's significant, and that's not being measured by traditional "counting" stats (total passing yards, etc.) or even by some rate stats (completion percentage; after all, QB A essentially "completed" 16 more passes in the season). 

The execution is a different matter. Here football stats are in their infancy compared to baseball stats; part of that is the inherently individualistic nature of baseball (hitter vs. pitcher) as opposed to football. So the question then is this: "is drawing PI penalties a repeatable skill for QBs, or is it essentially luck, or a combination of luck and having good receivers?" That's what we find out when measuring things year-by-year. If QB A consistently draws defensive PI penalties at a significantly higher rate year after year after year, we call that a "skill." A good example from baseball is "pitch framing" stats for catchers in baseball. We've learned that certain catchers are way, way better (or worse) than average in "stealing" called strikes from umpires; it's not sheer luck; it's a skill. It changed how MLB GMs evaluated catchers; guys that were significantly better than average at pitch framing add value, sometimes the equivalent of a few wins a year when compared to their poorer competitors, and free agent salaries began to reflect that.

Math anxiety is not your friend; it is something to be overcome.

 

11 minutes ago, DC Tom said:

 

I've published five papers on statistical physics. 

[CincyBillsFan]

30 minutes ago, The Frankish Reich said:

The concept isn't stupid at all. Let's say QB A, over the course of a season (beware small sample sizes!), draws defensive PI penalties at a higher than QB B. Let's say that this amounts to one additional PI call per game, with the average PI penalty being for 20 yards. At the end of the season, it's as if QB A passed for 16 x 20 = 320 more yards than QB B. (Notice how easy I kept the math ... even your wife would get this.) That's significant, and that's not being measured by traditional "counting" stats (total passing yards, etc.) or even by some rate stats (completion percentage; after all, QB A essentially "completed" 16 more passes in the season). 

The execution is a different matter. Here football stats are in their infancy compared to baseball stats; part of that is the inherently individualistic nature of baseball (hitter vs. pitcher) as opposed to football. So the question then is this: "is drawing PI penalties a repeatable skill for QBs, or is it essentially luck, or a combination of luck and having good receivers?" That's what we find out when measuring things year-by-year. If QB A consistently draws defensive PI penalties at a significantly higher rate year after year after year, we call that a "skill." A good example from baseball is "pitch framing" stats for catchers in baseball. We've learned that certain catchers are way, way better (or worse) than average in "stealing" called strikes from umpires; it's not sheer luck; it's a skill. It changed how MLB GMs evaluated catchers; guys that were significantly better than average at pitch framing add value, sometimes the equivalent of a few wins a year when compared to their poorer competitors, and free agent salaries began to reflect that.

Math anxiety is not your friend; it is something to be overcome.

 

It's not about "math anxiety", it 's about applying statistical & mathematical analysis to a chaotic system.  The three biggest issues involving the use of analytics as a tool in football compared to baseball are:

 

*  the enormous variability involved in EVERY NFL play.  The fact that on any given play there are a minimum of 40 human beings involved (22 players; 5 refs; 10 - 15 coaches) creates a huge # of potential outcomes.

 

*  the much smaller sample size in football means that the impact of the variability is magnified.  There are over 160 baseball games in a season compared to 16 football games. 

 

*  the number of subjective measurements involved.  When conducting controlled clinical trials or running lab experiments most of the data subjected to statistical analysis is obtained either through objective measurement or at least highly controlled subjective measurement.

 

Given these facts, it's particularly foolish to apply analytics to young players where the sample size is ridiculously small.

 

 

[Aussie Joe]

5 hours ago, stosh64 said:

Here is barnwell the clowns latest article.

"Don't count out the Dallas Cowboys yet: Why they could make an NFL playoff run" 

espn

Shocking view point, I know...

ETA a little relevant snippet,

What a clown.

Exactly..,

 

And if the game is closer than maybe the Bills keep trying to score instead of killing clock in the last quarter..

[DC Tom]

2 minutes ago, CincyBillsFan said:

 

It's not about "math anxiety", it 's about applying statistical & mathematical analysis to a chaotic system.  The three biggest issues involving the use of analytics as a tool in football compared to baseball are:

 

*  the enormous variability involved in EVERY NFL play.  The fact that on any given play there are a minimum of 40 human beings involved (22 players; 5 refs; 10 - 15 coaches) creates a huge # of potential outcomes.

 

*  the much smaller sample size in football means that the impact of the variability is magnified.  There are over 160 baseball games in a season compared to 16 football games. 

 

*  the number of subjective measurements involved.  When conducting controlled clinical trials or running lab experiments most of the data subjected to statistical analysis is obtained either through objective measurement or at least highly controlled subjective measurement.

 

Given these facts, it's particularly foolish to apply analytics to young players where the sample size is ridiculously small.

 

 

 

And it doesn't help that many of those subjective measures aren't subject to accurate modeling.  How should it impact a QB's rating when an obvious PI that saves a TD isn't called, is challenged, overturned, but still doesn't grant the TD?

[Saxum]

26 minutes ago, DC Tom said:

 

And it doesn't help that many of those subjective measures aren't subject to accurate modeling.  How should it impact a QB's rating when an obvious PI that saves a TD isn't called, is challenged, overturned, but still doesn't grant the TD?

 

That is pretty simple.  Teams get rated on zebra friendliness (I did not say it was a number but probably a multi value polynomial formula including things such as market size, income to NFL, ratings, number of cameras in game, etc) and that that factors into weight of failed calls.   Some teams definitely train players to play to the limit of zebra team.

[The Frankish Reich]

1 hour ago, DC Tom said:

 

I've published five papers on statistical physics. 

Well, then, let's hear about what's wrong with the concept of factoring in a QB's ability to draw PI calls ...

[The Frankish Reich]

1 hour ago, DC Tom said:

 

And it doesn't help that many of those subjective measures aren't subject to accurate modeling.  How should it impact a QB's rating when an obvious PI that saves a TD isn't called, is challenged, overturned, but still doesn't grant the TD?

With all due respect (hey, maybe you scored an 800 on the GRE verbal section), I don't think you read my post.

I noted that first we need to determine whether a particular stat measures a skill as opposed to randomness. That's why I used the example of pitch framing in baseball. We thought it was random -- the ump, the reputation of the pitcher (Greg Maddux always got the close ones), etc., etc. But when some smart analysts looked at it more closely, they recognized a pattern. It was repeatable. Every year Bengie Molina (an otherwise pretty much replacement level catcher) far exceeded expectations on called balls/strikes. The same thing happened with the great Batting Average on Balls in Play theory -- we discovered that pitchers have an extremely limited ability to avoid base hits on balls put in play. What we thought was a skill was, in fact, just noise.

EDIT: by the way, that response -- amounting to "what about bad refereeing" -- was rather underwhelming ...

 

Edited December 1, 2019 by The Frankish Reich