PFF IS GARBAGE FOR THIS

[Rocky P]

Talk about gaslighting

[Tuco Ramirez]

2 minutes ago, Rocky P said:

Talk about gaslighting

Yeah, PFF is ai garbage. Time to throw was 3.79 seconds but he was running for his life at less than 2 seconds. Plus our DTs were consistently stonewalling double teams. They couldn’t move a feather with a tank. 

[4thandFive]

Film Guy thinks PFF is accomplished by AI.

[Ben Blankenship]

I have a friend who works for PFF and I just asked him about this and this is what he said. So here's some insight into PFF. 

"They consider it winning the matchup if you block them for more than 2.5 seconds and anything that happens past 3.5 seconds puts no fault on the OL because it is unrealistic to block for that long. Also only one Texas DL had an above average pass rush win % (Watson) and the others had solid win % but nothing crazy. So basically they can have good grades for blocking pretty well because they weren't given any blame to some of the stuff that took a while to happen. Those grades in the tweet also aren't the 5 starters. Moe Jr got that grade on 2 snaps. The 5th starter was Perry who got a 47 grade. And the DL can still get some + grade for getting sacks after which is why Simmons still got a 92 but the OL isn't blames for that stuff. So probably some is on the QB. Faizon had a 40.7 which was the lowest on the team so I'm guessing that's what PFF would say happened. Faizon also had an average time to throw of 3.79 which is very high and pressure to sack rate of 43.5%, also very high"
 

[Rocky P]

1 minute ago, Ben Blankenship said:

I have a friend who works for PFF and I just asked him about this and this is what he said. So here's some insight into PFF. 

"They consider it winning the matchup if you block them for more than 2.5 seconds and anything that happens past 3.5 seconds puts no fault on the OL because it is unrealistic to block for that long. Also only one Texas DL had an above average pass rush win % (Watson) and the others had solid win % but nothing crazy. So basically they can have good grades for blocking pretty well because they weren't given any blame to some of the stuff that took a while to happen. Those grades in the tweet also aren't the 5 starters. Moe Jr got that grade on 2 snaps. The 5th starter was Perry who got a 47 grade. And the DL can still get some + grade for getting sacks after which is why Simmons still got a 92 but the OL isn't blames for that stuff. So probably some is on the QB. Faizon had a 40.7 which was the lowest on the team so I'm guessing that's what PFF would say happened. Faizon also had an average time to throw of 3.79 which is very high and pressure to sack rate of 43.5%, also very high"
 

i would challenge the 3.79 seconds to throw.  Texas was in his lap all game long, and just because you are in front of the player if he pushing you all the way back into the QB, is that really winning?

[Here for the Wins]

Several sacks were because Faizon left the pocket and ran into Simmons who wasn’t really rushing but playing contain.  That’s on the QB not the Oline.  Believe the RBs were graded as atrocious in the pass pro too - actually it was only Bishop.

Sanders grade now sits at 72.1.  Moe only played 2 pass plays.  Perry played 58 and was graded at 47.2 yet not listed here.

The average time to throw is likely solid.  It does not mean you cannot or need not apply context.  When “kept clean” that time is 3.27 seconds.  That’s more time than most take throughout a game when including pressured.  When “under pressured” it was 4.37, also a high number but not as high relatively speaking.  Almost 100% of the time, the under pressure rate results in a higher average time to throw than the kept clean time, at least in my observations.

[Ben Blankenship]

I would agree with that. Just because he is engaged on a block doesn't mean he is winning the block. I would not consider that a win.

[Here for the Wins]

5 minutes ago, Rocky P said:

i would challenge the 3.79 seconds to throw.  Texas was in his lap all game long, and just because you are in front of the player if he pushing you all the way back into the QB, is that really winning?

Remember when Arch rolled left versus UTEP, then rolled right then threw the pick?  That was probably 7 seconds or more from snap to release.  Thats what would get thrown in that bucket.  It’s not indicative of when pressure happened but actual time from snap to release however that occurs.

[ArizonaLonghorn]

17 minutes ago, Ben Blankenship said:

I have a friend who works for PFF and I just asked him about this and this is what he said. So here's some insight into PFF. 

"They consider it winning the matchup if you block them for more than 2.5 seconds and anything that happens past 3.5 seconds puts no fault on the OL because it is unrealistic to block for that long. Also only one Texas DL had an above average pass rush win % (Watson) and the others had solid win % but nothing crazy. So basically they can have good grades for blocking pretty well because they weren't given any blame to some of the stuff that took a while to happen. Those grades in the tweet also aren't the 5 starters. Moe Jr got that grade on 2 snaps. The 5th starter was Perry who got a 47 grade. And the DL can still get some + grade for getting sacks after which is why Simmons still got a 92 but the OL isn't blames for that stuff. So probably some is on the QB. Faizon had a 40.7 which was the lowest on the team so I'm guessing that's what PFF would say happened. Faizon also had an average time to throw of 3.79 which is very high and pressure to sack rate of 43.5%, also very high"
 

Far too nuanced for your average Fan Site Expert, alas!

Burn it down!

[GoHorns1]

Proof once again that PFF is absolute joke in grading.

[ArizonaLonghorn]

From the free side of their site (I think) - "How PFF grades pass protection".   Not what might seem intuitive to the casual viewer like me, but it's worth understanding what and why they are doing even if you disagree.

I just started a subscription to PFF earlier this year and am still learning how they grade ... it's frustrating at times but usually there's a method to their madness.

[Texasborn91]

If RT didn’t grade out in dark dark red I believe none of it. 

[Nate C]

PFF includes scramble time in time to throw lol

[Ben Blankenship]

10 minutes ago, Texasborn91 said:

If RT didn’t grade out in dark dark red I believe none of it. 

Their right tackle had a 47 grade. Which honestly might be way to high after that performance

[Here for the Wins]

34 minutes ago, Nate C said:

PFF includes scramble time in time to throw lol

So they should cut it off before he throws it?  So you’d call it “average time to………not throw just yet.?”

[Nate C]

9 minutes ago, Here for the Wins said:

So they should cut it off before he throws it?  So you’d call it “average time to………not throw just yet.?”

No… You shouldn’t include scramble time in time to throw. When the pocket collapses and you have to scramble, that’s different than pocket time. For example,  when Colin strip sacked Faizon and the ball went out of bounds, they would rule it “7 seconds time to throw” when the pocket collapsed in two seconds. The Extra 5 seconds he was running for his life.

[Here for the Wins]

1 minute ago, Nate C said:

No… You shouldn’t include scramble time in time to throw. When the pocket collapses and you have to scramble, that’s different than pocket time. For example,  when Colin strip sacked Faizon and the ball went out of bounds, they would rule it “7 seconds time to throw” when the pocket collapsed in two seconds. The Extra 5 seconds he was running for his life.

Why not?  
 

They have lots of information.  It is broken down multiple ways.  Even as a user of the information, you can disregard whatever piece you find not useful.  That average time to throw can be broken down multiple ways.  Did you know that?  Because often as it’s presented is in the context of every pass.  But it is available when “kept clean” as a standalone so inherent in that is how long a QB sits in the pocket.  So now imagine when collecting that data point only on non-scrambles or just collect it every time then sort it out.  Probably easier to collect every play - snap to throw equals no ambiguity.  And the thing is that every game is collected consistently.

So I ask you now, did you know it can be broken down separately?

[Gerry Hamilton]

I’m not a PFF guy … never have been.

Raheem Biles this year so far is another reason why. 

[justwinbaby]

Always thought PFF is garbage

[Nate C]

13 minutes ago, Here for the Wins said:

Why not?  
 

They have lots of information.  It is broken down multiple ways.  Even as a user of the information, you can disregard whatever piece you find not useful.  That average time to throw can be broken down multiple ways.  Did you know that?  Because often as it’s presented is in the context of every pass.  But it is available when “kept clean” as a standalone so inherent in that is how long a QB sits in the pocket.  So now imagine when collecting that data point only on non-scrambles or just collect it every time then sort it out.  Probably easier to collect every play - snap to throw equals no ambiguity.  And the thing is that every game is collected consistently.

So I ask you now, did you know it can be broken down separately?

It’s not a stat people have been using to judge the performance of Faizon. So when you look at “average time to throw” as a whole, without considering the above, it makes Faizon look worse than he really was. 

[horns96]

2 hours ago, Ben Blankenship said:

I have a friend who works for PFF and I just asked him about this and this is what he said. So here's some insight into PFF. 

"They consider it winning the matchup if you block them for more than 2.5 seconds and anything that happens past 3.5 seconds puts no fault on the OL because it is unrealistic to block for that long. Also only one Texas DL had an above average pass rush win % (Watson) and the others had solid win % but nothing crazy. So basically they can have good grades for blocking pretty well because they weren't given any blame to some of the stuff that took a while to happen. Those grades in the tweet also aren't the 5 starters. Moe Jr got that grade on 2 snaps. The 5th starter was Perry who got a 47 grade. And the DL can still get some + grade for getting sacks after which is why Simmons still got a 92 but the OL isn't blames for that stuff. So probably some is on the QB. Faizon had a 40.7 which was the lowest on the team so I'm guessing that's what PFF would say happened. Faizon also had an average time to throw of 3.79 which is very high and pressure to sack rate of 43.5%, also very high"
 

Is your friend's name "Claude"?

I'm just kidding, but it strains credulity that, conservatively, PFF could watch every play of all 100+ NCAA games in 24 hours and grade every 100 total players in each game and provide meaningful grades for those thousands of players.   Do they employ 1000+ skilled tape watchers?

I ask this sincerely, as it just doesn't make any sense.

Edited 1 hour ago by horns96

[RattlerHorn]

See for yourself who the blame should be on.

 

[ArizonaLonghorn]

40 minutes ago, horns96 said:

Is your friend's name "Claude"?

I'm just kidding, but it strains credulity that, conservatively, PFF could watch every play of all 100+ NCAA games in 24 hours and grade every 100 total players in each game and provide meaningful grades for those thousands of players.   Do they employ 1000+ skilled tape watchers?

I ask this sincerely, as it just doesn't make any sense.

I also wondered about the numbers, the sheer tonnage of data they collect and analyze in such a short time, so I dug into it a bit.  138 FBS teams and these are graded first.  So a max # of 138 games if each played the same day against a non-FBS team (which is unlikely, but a hard ceiling).  69 games if all 138 played other FBS teams but there are usually a few teams with bye weeks or the games are spread across different days.  And the games are spread across 3 or 4 viewing windows during the day so they can finish say an early game while the 2nd window game plays, then work on the evening game.  So probably need roughly half as many graders as games played on the busiest day.

You ask "Do they employ 1000+ skilled tape watchers?"  Check this PFF Player Grades link (free), scroll down to the "Who is doing the Grading?" block and it says they have "600 full or part-time analysts, but less than 10% of analysts are trained to the level that they can grade plays. "  So 60-ish graders and another 9-10 support guys per grader.  

What do the support guys do?  I guess they do the grunt work of chopping up the videos so the grader can simply grade, maybe IDing players and formations, breaking plays down into run vs pass etc.  I've exchanged messages with a guy who works part time for PFF and he said he was responsible for one task on every play - with multiple guys doing this it would free up a grader to simply evaluate by position.

How much actual time is taken up by active plays in a 60 minute game?  I've heard 12-15 minutes, surely less than 20 minutes.  So if the All-22 is broken down right they can probably grade two games a day.  You mention "100 total players in each game" but they would only do 22 per play and somewhere in that link I gave you it says most plays grade as neutral for most of the players so they can probably rip thru it pretty fast.

I was skeptical at first but I can see how they can pull it off.  

Hope this helps.  Not trying to carry water for PFF, just trying to understand it better.  This was something I dug into a little (I had the same concerns you raise) and now I can see how it makes sense that they could do it that quickly with 60 trained graders.  Whether you agree with their ranking results is a different topic.

[Ben Blankenship]

1 hour ago, horns96 said:

Is your friend's name "Claude"?

I'm just kidding, but it strains credulity that, conservatively, PFF could watch every play of all 100+ NCAA games in 24 hours and grade every 100 total players in each game and provide meaningful grades for those thousands of players.   Do they employ 1000+ skilled tape watchers?

I ask this sincerely, as it just doesn't make any sense.

Haha they do employ a lot of people I believe. They grade the games live and by quarter. And they have at least 2 people per game. One charting and one reviewing. I don’t know much more beyond that. 

[horns96]

5 minutes ago, Ben Blankenship said:

Haha they do employ a lot of people I believe. They grade the games live and by quarter. And they have at least 2 people per game. One charting and one reviewing. I don’t know much more beyond that. 

Appreciate the insight.