Once and for all....PFFffffffft

[Royale with Cheese]

1 minute ago, Einstein said:

 

to be honest, I disagree with several of PFF grade. Ed is just one of them.

 

but keep in mind that two NFL coaches may disagree with each other on their own grading.

 

so, in my opinion, that’s not that unusual.

 

The issue is, you can't argue hard for PFF when you yourself disagree yourself with several of their grades.  You can't pick and choose what is a legit grade or not.

 

When a website has both Ed Oliver and Jeffrey Simmons outside of the top 50 for interior defensive lineman, their legitimacy has to be questioned.

There is not a GM, coach or scout that would have both of them on the outside of the top 50 or anywhere close to it.  Jordan Phillips top 15?

[Airseven]

PFF is one grading system. It's informed entertainment. 

[Mango]

1 hour ago, eball said:

Love this.  Many of us have been "skeptical" (I'm being kind) to the grading system PFF uses to assess NFL players. 

 

 

 

I don't disagree on this topic, but JJ Watt is also a dummy. 

I have no other value to offer to this conversation. 

[Royale with Cheese]

6 minutes ago, Mango said:

 

I don't disagree on this topic, but JJ Watt is also a dummy. 

I have no other value to offer to this conversation. 

 

Ed Oliver is out of the top 60 interior DL according to PFF.....unless I'm reading it wrong.

[WideNine]

1 hour ago, eball said:

Love this.  Many of us have been "skeptical" (I'm being kind) to the grading system PFF uses to assess NFL players. 

 

 

 

Thank you JJ!!!

 

Some of us have been preaching that to the analytics choir here to no avail for years.

 

Those tools are just tools (no pun intended)... Coaches look at a lot of other factors when it comes to player evaluation.

 

 

 

[BullBuchanan]

I mean, most coaches are awful and the majority of the rest are painfully mediocre former athlete dinosaurs. Is it a surprise their grades don't correlate 1:1 with empirical analysis?

17 minutes ago, Royale with Cheese said:

 

The issue is, you can't argue hard for PFF when you yourself disagree yourself with several of their grades.  You can't pick and choose what is a legit grade or not.

 

When a website has both Ed Oliver and Jeffrey Simmons outside of the top 50 for interior defensive lineman, their legitimacy has to be questioned.

There is not a GM, coach or scout that would have both of them on the outside of the top 50 or anywhere close to it.  Jordan Phillips top 15?

A grade is a grade based on inputs and a model. It doesn't care what you believe. What a result of a model means is up for interpretation, but the grade itself really isn't so long as the input is accurate to the rubric.

[WideNine]

3 minutes ago, BullBuchanan said:

I mean, most coaches are awful and the majority of the rest are painfully mediocre former athlete dinosaurs. Is it a surprise their grades don't correlate 1:1 with empirical analysis?

 

Would like to see pffffft coach an NFL team.

 

Yeah, that would work.

 

 

 

[Buffalo716]

28 minutes ago, Beck Water said:

I'm not a scout, but I'm not sure that's the way "most NFL coaches grade" and would welcome some input on this - any input from guys who've coached at the college or HS level @HoofHearted @Buffalo716?

 

Either way, I think the main point is that the coaches (and scouts) are typically people who have a deep and nuanced knowledge of the game.  They're people who can diagnose the defensive coverage based on subtle cues like footwork and body position.  We saw this kind of thing with Tony Romo when he started his broadcast career, where he could accurately read the defense and predict what the offense was going to do and the network told him to tone it down.  The ability of these people with a lifetime football background, playing and coaching, to deduce the play call, or the coverage, or the protection call, after looking at film, is on a different level.

 

It's my understanding that PFF uses a lot of graders in Ireland and now India whose understanding of the game is taking a few training sessions.  Now they claim that only 10% of their employees finalize grades or something like that, and that everything is reviewed, but 🤷‍♂️?

 

Some critiques of PFF:

https://sportank.com/media/post/is-pff-reliable-are-they-all-that-credible

 

I picked these because neither are out to just dismiss or trash PFF.

You typically will watch film with your position group

 

There are two angles... Sideline Angle and end zone angle 

 

The sideline angle will keep every player in view throughout the entire play

 

The end zone angle is the box angle... And it does not come off of the box... Basically used for the front seven and oline

 

So an offensive line coach... Will be responsible for grading his lineman

 

He will typically use the box angle

 

And I guess it differs from coach to coach... But a lot of coaches, especially offensive line coaches will use a three-point system

 

Just a + or - for assignment, technique and effort

 

You know how many snaps they play and you multiply that by 3 and that's how many total possible pluses you can have

 

So if you played 50 snaps ... You could have up to 150 pluses... After breaking down execution technique and effort if you wind up with 140 +

 

That would be a really good day....

 

Compared to if you only had 80 pluses... You got a lot to work on

 

And then you could further break down your technique and execution to see what you need to work on or what you excel at... Obviously seeing effort is the easiest thing on film

 

And concerning when you don't bring it

Edited January 17 by Buffalo716

[BullBuchanan]

2 minutes ago, WideNine said:

 

Would like to see pffffft coach an NFL team.

 

Yeah, that would work.

 

 

 

Why would you want PFF to coach a team? What point is it that you think you're making?

[WideNine]

1 minute ago, BullBuchanan said:

Why would you want PFF to coach a team? What point is it that you think you're making?

 

That was sarcasm.

 

 

[Punch]

It's already been pointed out in this thread, but NFL teams pay for and utilize PFF analysis. Many, if not most, NFL teams also employ their own analysts that compile and review similar data.

 

I don't wholeheartedly endorse PFF by any means, but the type of analysis they provide should be weighted appropriately in concert with more traditional coaching and scouting analysis. Any kind of data analytics is only a piece of the puzzle. Even PFF supports this notion. 

[WideNine]

1 minute ago, BullBuchanan said:

Typically sarcasm has to have relevance to the point being made.

 

Well it was.

 

Your point regarding dinosaur coaches not being in alignment with empirical analysis.

 

The relevance being let the champions of empirically incomplete and often flawed analysis (pfffft) try to coach an NFL team successfully.

 

 

 

[BullBuchanan]

Just now, WideNine said:

 

Well it was.

 

Your point regarding dinosaur coaches not being in alignment with empirical analysis.

 

The relevance being let the champions of empirically incomplete and often flawed analysis (pfffft) try to coach an NFL team successfully.

 

 

 

What is your evidence to suggest the analysis is flawed?

[WideNine]

5 minutes ago, BullBuchanan said:

What is your evidence to suggest the analysis is flawed?

 

They are never flawed, just ask them. They just run into anomalies and statical outliers like Josh Allen.

 

 

[BullBuchanan]

1 minute ago, WideNine said:

 

They are never flawed, just ask them. They just run into anomalies and statical outliers like Josh Allen.

 

 

what anomalies do they run into? How is Allen a statistical outlier and if he is how does that make the analysis flawed?

[hondo in seattle]

1 hour ago, Punch said:

PFF loves Josh Allen, FWIW. 

 

Not to defend PFF, but "literally putting a coaches grade next to a PFF grade" doesn't really prove much considering how many meatheads are coaching in the NFL. Coaches are constantly wrong when evaluating talent and situation. 

 

Sure, coaches are wrong sometimes.  Talent evaluation is an inexact science.  But comparing dedicated professionals to the amateurs at PFF is a stretch.  

 

Coaches have expertise in football.  PFF's analysts have expertise in statistics.  PFF's graders are young guys with no credentials and boring lives who sit at home watching out All-22 handing out subjective, uninformed grades.    

 

Coaches, on the other hand, consider the assignment, the situation, and what the player was coached to do when giving out a grade.   

 

Yes, coaches sometimes evaluate guys poorly.   But let's remember what an analytics guru once famously said, "If Josh Allen succeeds, the Bills will have outsmarted basically all regular humans and the entirety of math itself."

 

I'm not saying there's no place for analytic sites like PFF.  I trust they do the math right.  But I'll trust the football professionals to evaluate football.   If the coaches' grades don't line up with PFF, my hunch is PFF needs to get better.  

 

 

[Punch]

Just now, hondo in seattle said:

 

Sure, coaches are wrong sometimes.  Talent evaluation is an inexact science.  But comparing dedicated professionals to the amateurs at PFF is a stretch.  

 

Coaches have expertise in football.  PFF's analysts have expertise in statistics.  PFF's graders are young guys with no credentials and boring lives who sit at home watching out All-22 handing out subjective, uninformed grades.    

 

Coaches, on the other hand, consider the assignment, the situation, and what the player was coached to do when giving out a grade.   

 

Yes, coaches sometimes evaluate guys poorly.   But let's remember what an analytics guru once famously said, "If Josh Allen succeeds, the Bills will have outsmarted basically all regular humans and the entirety of math itself."

 

I'm not saying there's no place for analytic sites like PFF.  I trust they do the math right.  But I'll trust the football professionals to evaluate football.   If the coaches' grades don't line up with PFF, my hunch is PFF needs to get better.  

 

 

 

No kidding. In a vacuum, a single PFF grade vs. a single coaches grade doesn't prove much, which is pretty plainly all I'm suggesting. In a later post in this thread, I largely agree with what you posted.

 

Meanwhile, PFF grades are "molded by the NFL" as our old friend Matthew Coller tweeted (I'll reiterate for the third time, I do not necessarily wholeheartedly endorse PFF):

 

 

 

[WideNine]

1 hour ago, hondo in seattle said:

 

Sure, coaches are wrong sometimes.  Talent evaluation is an inexact science.  But comparing dedicated professionals to the amateurs at PFF is a stretch.  

 

Coaches have expertise in football.  PFF's analysts have expertise in statistics.  PFF's graders are young guys with no credentials and boring lives who sit at home watching out All-22 handing out subjective, uninformed grades.    

 

Coaches, on the other hand, consider the assignment, the situation, and what the player was coached to do when giving out a grade.   

 

Yes, coaches sometimes evaluate guys poorly.   But let's remember what an analytics guru once famously said, "If Josh Allen succeeds, the Bills will have outsmarted basically all regular humans and the entirety of math itself."

 

I'm not saying there's no place for analytic sites like PFF.  I trust they do the math right.  But I'll trust the football professionals to evaluate football.   If the coaches' grades don't line up with PFF, my hunch is PFF needs to get better.  

 

 

 

Well, those that prostrate themselves before the alter of data analytics can still hope.

 

Perhaps someday in the future they will be suffiently rendered down to enough bits and bytes that they will not actually have to try anything because their predicted performance will be provided in a KPI report.

 

 

Edited January 18 by WideNine

[Royale with Cheese]

1 hour ago, BullBuchanan said:

I mean, most coaches are awful and the majority of the rest are painfully mediocre former athlete dinosaurs. Is it a surprise their grades don't correlate 1:1 with empirical analysis?

A grade is a grade based on inputs and a model. It doesn't care what you believe. What a result of a model means is up for interpretation, but the grade itself really isn't so long as the input is accurate to the rubric.

 
Okay so you with PFF grade that Ed Oliver has been below average this year. 

58 minutes ago, BullBuchanan said:

What is your evidence to suggest the analysis is flawed?

They have Jordan Davis as a top 10 DT and he’s been terrible this year.

 

Ed Oliver, below average.

Edited January 17 by Royale with Cheese

[mike22nc]

I’m a bit confused by some people in the thread considering PFF grades to be analytics. Analytics is, in general, the collection, calculation and interpretation of objective statistics. 
 

PFF grades are not analytics in my opinion. They are more of a scouting tool, which are subjective and are an interpretation and evaluation of each play on the field. They surely can have value if the evaluation is done well. Do they have the full context when determining their grades? Of course not, just like many statistics don’t take every factor (or even most factors) into account.

 

From what I see, the PFF grading is better and more accurate for certain positions than others. The offensive and defensive lines are particularly difficult to judge and I don’t place much store into the grades provided. Everyone is bringing up Oliver and I would need to jump on that bandwagon. The numbers backup what we’ve been watching all year, and having him outside the top 10 much less the top 50 of the interior lineman is egregious.