PFF Grades - LBs

[atlantacowboy]

I thought my post was dripping with sarcasm, but I guess you failed to pick up on it

Sarcasm does not translate unless the reader is familiar with your opinions. I am not. My apologies.

 

[GORICO]

I don't pay for access to PFF, but here's a link to the LB grades so far for the season. In the link, the top 3 are shown, then the rest are in some random order. Just click the player's name and you will get their overall grade and rank.

https://www.pff.com/nfl/grades/position/lb

Parsons is #3 overall
Jaylon is #20 overall
LVE is #67 overall
Neal is #66 overall

A couple that I also found interesting:
Myles Jack is #51 overall
Anthony Hitchens is #63 overall

pleasant surprise for Jaylon?

 

[TheMarathonContinues]

pleasant surprise for Jaylon?

It is...and I know people highlight the 20 yard play they hold him responsible for but he really was a reason why Jalen Hurts did run down our throats.

 

[GORICO]

It is...and I know people highlight the 20 yard play they hold him responsible for but he really was a reason why Jalen Hurts did run down our throats.

20 overall is much better than expected

 

[GORICO]

It is...and I know people highlight the 20 yard play they hold him responsible for but he really was a reason why Jalen Hurts did run down our throats.

was hoping LVE might be graded higher...but then seeing Jack and Hitchens...maybe because early in season and it takes longer for defenses to catch up

 

[plasticman]

They are actually very open about their scoring system. I see you've taken a real deep dive into their web site.

I am talking about a specific mathematical formula and not a general description of the elements that are involved.

For example, the passer rating formula used by most NFL statistics is:

C=Completions
A=Attempts
P=Passing yards
T=Touchdowns
I=Interceptions

Passer Rating = 5 X (C/A-.3) + .25 X (P/A-3) + (T/A X20) + 25 X (2.375-I/A)

I did not see such a formula in their description of how they determine their scores.

Apparently, you took a "deeper dive" than I did so perhaps you could enlighten me?

 

[atlantacowboy]

I am talking about a specific mathematical formula and not a general description of the elements that are involved.

For example, the passer rating formula used by most NFL statistics is:

C=Completions
A=Attempts
P=Passing yards
T=Touchdowns
I=Interceptions

Passer Rating = 5 X (C/A-.3) + .25 X (P/A-3) + (T/A X20) + 25 X (2.375-I/A)

I did not see such a formula in their description of how they determine their scores.

Apparently, you took a "deeper dive" than I did so perhaps you could enlighten me?

There is no "mathematical formula" which is why it is a superior system. It is far more nuanced. They grade every player on every play. The grade is obviously subjective and that's a good thing unless you think a QB's grade should be dinged the same for a drive killing INT compared to an INT on a desperation Hail Mary at the end of the half. Your QBR formula makes no distinction between those INTs. There are also several layers of graders many of whom are former pro coaches and scouts. So, its not guys like us making the call from our couches. These are pros looking at the same pro film NFL coaches use.

https://www.pff.com/news/pro-how-pff-grades-quarterback-play

 

[ItzKelz]

Just DAL:
3: Parsons
20: Jaylon
66: Neal
67: LVE

ItzKelz be like......

 

[ItzKelz]

I find it interesting how people's bias clouds their judgment on how players are performing. PFF isn't perfect by any means, but it's a decent tool and is fairly accurate in my opinion. It's limited data at this point and some of our players rankings are bound to change. For example, Tony Pollard is ranked #1 among RB's at the moment and I have a hard time seeing that hold up all year.

Jaylon said watch the tape.....cause he ball so hard

 

[ItzKelz]

I don't mind PFF, directionally it seems to backup a number of opinions. But how did Dak not score a better grade missing just 5 passes on Monday?

fumble for a TD for the other team

 

[plasticman]

There is no "mathematical formula" which is why it is a superior system. It is far more nuanced. They grade every player on every play. The grade is obviously subjective and that's a good thing unless you think a QB's grade should be dinged the same for a drive killing INT compared to an INT on a desperation Hail Mary at the end of the half. Your QBR formula makes no distinction between those INTs. There are also several layers of graders many of whom are former pro coaches and scouts. So, its not guys like us making the call from our couches. These are pros looking at the same pro film NFL coaches use.

https://www.pff.com/news/pro-how-pff-grades-quarterback-play

Far more nuanced?

Okay, but at some point, a score is assigned to each type of completion per your description. What score is assigned a Hail Mary , what score is assigned a drive killing INT, etc. they do not clearly define a score given for each type of scenario.

Also, it seems as though they are saying that the type of scenario is determined by an individual, coach, expert, whatever. In what way do they assure it is consistent?

If they do not have a consistent description on exactly how they score then why would I ever trust it?.

You are correct that the passer rating has flaws and I am in favor of a more fair scoring system. However, anytime a score is based on a judgement by a human, no matter their expertise, there is the opportunity for bias, whether they realize it or not.

The current passer rating for a single game should never be used to judge a QB. However, a seasons worth of passes would absorb both the Hail Nary's and the drive killing INT's due to a much larger data pool. the larger the data pool, the more relevant the description.

Please notice that I did not attempt to make some sarcastic comment regarding your explanation. I am not trying to impress, i am not trying to belittle your effort or understanding on the subject matter. I believe in mutual respect and courtesy on this forum.

 

[atlantacowboy]

Far more nuanced?

Okay, but at some point, a score is assigned to each type of completion per your description. What score is assigned a Hail Mary , what score is assigned a drive killing INT, etc. they do not clearly define a score given for each type of scenario.

Also, it seems as though they are saying that the type of scenario is determined by an individual, coach, expert, whatever. In what way do they assure it is consistent?

If they do not have a consistent description on exactly how they score then why would I ever trust it?.

You are correct that the passer rating has flaws and I am in favor of a more fair scoring system. However, anytime a score is based on a judgement by a human, no matter their expertise, there is the opportunity for bias, whether they realize it or not.

The current passer rating for a single game should never be used to judge a QB. However, a seasons worth of passes would absorb both the Hail Nary's and the drive killing INT's due to a much larger data pool. the larger the data pool, the more relevant the description.

Please notice that I did not attempt to make some sarcastic comment regarding your explanation. I am not trying to impress, i am not trying to belittle your effort or understanding on the subject matter. I believe in mutual respect and courtesy on this forum.

0 is a neutral score. So, a QB throwing a Hail Mary would not be dinged and get a neutral score for the play. There is never going to be a mathematical formula that perfectly describes a football players performance. Too many variables and judgement calls. Measuring only the result of the plays without context does not give an accurate picture of how the player actually performed. I've seen games where Romo had a terrible QBR b/c he had 3 INTs..... one bounced off the wrs hands. Another the wr ran the wrong route on a timing throw. And the 3rd was a desperation heave. He actually played well but you would never know it looking at his QBR.

The "data pool " theory holds true for PFF as well. The more data, the less relevant the bias. Its a more accurate b/c they go into more detail. The scoring is not that controversial. The QB throws an INT. The scouts look at the actual play to score it, not just the result of the play. How is that not superior?

 

[kumizi]

Subjective stats like PFF are useful if people understand how to use them. They dont really mean too much after 3 games. They're like defensive stats in baseball. Even over one full season, several players scores can be way off. The bigger sample size you have, the more useful they are.

Jaylon actually looked decent in weeks 2 and 3 so it doesnt surprise me. I dont believe he ends up in the top 30-40 by season's end though.

Similar for Pollard. He's been very good. I've been vocal how good he is but he isn't going to end up a top 7-8 running back at seasons end.

After 8-10 games, I start paying attention to PFF.

After ~30 or so games (or 2 full season), I start making definitive judgements on players based on PFF.

For example, PFF leads me to believe Connor Williams is about average and not nearly as bad as Cowboys fans make him out to be. We have a big sample size on Connor. Similarly, unless I see improvement soon, PFF leads me to believe Biadasz is bad. He's played 15 games so far. I'll give him the rest of the season to improve.