I don't recall a time they explained a grade they gave a performance. Story is Jordan Love & CJ Stroud had 99.9% identical stats in their respective playoff games.
PFF gave Love the highest grade of the week, a 92.5, while they gave Stroud a 77.8. JJ Watt called them out, obviously for the Texans connection. But he made a great point.
This got PFFs attention and they put out an long video on YouTube explaining as well as being very defensive and pissy about what Watt said. These guys are douche bags.
I had never even know PFF had a show until this situation. They come off very unlikable and it wouldn't surprise me if their feelings were involved in the grading. They don't even definitively say what comprises the grades. A lot of "I think" "probably" "lets pretend" hypothetical answers.
This is why I can't take them seriously. No one knows their formula, apparently not even them, AND they do a terrible job explaining.
Downfall is a bit much. Remember PFF is a for profit company trying to make money like everyone else. They have a pretty damn good membership with a really devout base. I doubt they will be going anywhere anytime soon.
If anything this is very good for the industry. Most people couldn't tell you the difference between a linear and a logarithmic regression, let alone understand what they are talking about when they talk about things like machine learning (it's just statistics rebranded) and analytics. Which is fine. You don't really need to understand how a car works for it to be useful to you. Which is how we always try to get people to see data. All models are wrong but some are useful. And that's the real key. Use. Think of it like hurricane spaghetti models for example. All wrong, but the usefulness is gtfo of dodge.
At the end of the day do these paywall analytics really provide any more usefulness than a madden rating? What matters is their users believe they do. And there's something about secret sauce analytics (PFF, DVOA, etc) that makes people feel they have an extra edge over those with casual analytics (EPA, QBR) or just old fashion stats (PPG). So yeah, I don't see PFF going anywhere anytime soon.
Using stats to judge any of this stuff is inherently flawed - especially with quarterbacks.
You have a play where the QB does everything right, but his linemen screw the pooch, and his receivers all stand around with their thumbs up there asses, so the guy miraculously avoids 3 pass rushers, scrambles away, and tosses the ball out of bounds. Freaking brilliant QB play for that situation. On the sheet it's an incomplete pass.
You have a play where the receiver gets behind the safety and the QB hits him in stride. 20-yard touchdown. But the only reason it's 20 is because that's how far away they were. The exact same play could have been 70, or 99, or 750 if they moved the goal line out there. But put that 20 on the stat sheet.
And of course the dump offs. QB slings the ball out to the running back in the flat. Easiest toss and catch that anyone could ever make. Running back makes 9 guys miss and goes 80 yards. Stat sheet rings up all 80. Even worse if it's one of those glorified hand-off tosses that goes for 80.
At the end of the day it's players that win you the game on the field. Honestly have no clue what our EPA was during our championship runs and I don't really care. As BB said, stats are for losers.
Any sort of grading tool will be flawed to some extent. I think PFF can be a decent tool if people understand it’s not the holy bible of football. Also, it gets annoying when people flip flop on PFF grades based on bias. Like the Bucs subreddit was using PFF to point out AWJ is an elite safety (he is, but not the point) but then when someone pointed out that PFF graded Kancey as a below average player then the grades all of a sudden are bullshit. People only get mad when the PFF grade doesn’t support their bias, just like JJ Watt throwing a fit because the QB of his former team didn’t get graded good enough. Who cares.
Last edited by Miller4Prez64 on Thu Jan 18, 2024 1:28 pm, edited 1 time in total.
Miller4Prez64 wrote: ↑Thu Jan 18, 2024 1:22 pm
Any sort of grading tool will be flawed to some extent. I think PFF can be a decent tool if people understand it’s not the holy bible of football. Also, it gets annoying when people flip flop on PFF grades based on bias. Like the Bucs sub was using PFF to point out AWJ is an elite safety (he is, but not the point) but then when someone pointed out that PFF graded Kancey as a below average player then the grades all of a sudden are bullshit. People only get mad when the PFF grade doesn’t support their bias, just like JJ Watt throwing a fit because the QB of his former team didn’t get graded good enough. Who cares.
You are 100% correct here. We see examples of this on the IABL pages all the time. But it adds to the discussion and, imo, the overall fun of talking about the sport all of us in here love.
Miller4Prez64 wrote: ↑Thu Jan 18, 2024 1:22 pm
Any sort of grading tool will be flawed to some extent. I think PFF can be a decent tool if people understand it’s not the holy bible of football. Also, it gets annoying when people flip flop on PFF grades based on bias. Like the Bucs sub was using PFF to point out AWJ is an elite safety (he is, but not the point) but then when someone pointed out that PFF graded Kancey as a below average player then the grades all of a sudden are bullshit. People only get mad when the PFF grade doesn’t support their bias, just like JJ Watt throwing a fit because the QB of his former team didn’t get graded good enough. Who cares.
Watch the video and then you'll understand why it's flawed. The PFF analysts struggled mightily to explain the grading. They did a ton of talking and breaking down and it was all based on "well let's just say....." The problem is and always has been that they don't admit their formula is mostly nonexistent and at its best based off subjective analytics. PFF treat their grades as the holy Bible, so that's what gets people pissed off.
Miller4Prez64 wrote: ↑Thu Jan 18, 2024 1:22 pm
Any sort of grading tool will be flawed to some extent. I think PFF can be a decent tool if people understand it’s not the holy bible of football. Also, it gets annoying when people flip flop on PFF grades based on bias. Like the Bucs sub was using PFF to point out AWJ is an elite safety (he is, but not the point) but then when someone pointed out that PFF graded Kancey as a below average player then the grades all of a sudden are bullshit. People only get mad when the PFF grade doesn’t support their bias, just like JJ Watt throwing a fit because the QB of his former team didn’t get graded good enough. Who cares.
Watch the video and then you'll understand why it's flawed. The PFF analysts struggled mightily to explain the grading. They did a ton of talking and breaking down and it was all based on "well let's just say....." The problem is and always has been that they don't admit their formula is mostly nonexistent and at its best based off subjective analytics. PFF treat their grades as the holy Bible, so that's what gets people pissed off.
I fully understand that it’s flawed. Nobody should take them as serious rankings of a player’s ability or talent. But if it’s so flawed, why do people get so up in arms about their players not being graded better? The Pro Bowl is one thing because that actually can affect bonuses, legacy, HoF status but nobody uses PFF seriously for anything like that.
Part of the "flaw" in objective grading is that if Jordan Love fits a ball into a window 3 feet wide for a 25 yard TD to his fourth read and Stroud throws a ball to a guy on his first read with no one within 15 yards of him who walks in for a 25 yard TD, the stats say those plays were equally good from the QBs. Anyone watching it might tell you, well, actually that was a very easy play for Stroud and a very difficult play for Love. PFF tries to give you the later analysis, not the former.
Ken wrote: ↑Thu Jan 18, 2024 3:45 pm
Part of the "flaw" in objective grading is that if Jordan Love fits a ball into a window 3 feet wide for a 25 yard TD to his fourth read and Stroud throws a ball to a guy on his first read with no one within 15 yards of him who walks in for a 25 yard TD, the stats say those plays were equally good from the QBs. Anyone watching it might tell you, well, actually that was a very easy play for Stroud and a very difficult play for Love. PFF tries to give you the later analysis, not the former.
This. Anyone who's bothered watching their youtube shows would know this. People are just ignorant as to what PFF sets out to do and they don't like it when their grades don't match up with expectations.
PFF isn't perfect, but it's still the best metric we have for trying to objectively grade players on a play by play basis without focusing on results but rather the process.
In the situation I described, IIRC, Stroud would get a low value for doing what is expected of him... hitting a wide open guy, maybe a 0.1. And Love would get a high value like a 0.9 for the play that the stats say are equal. So you should expect the grading to diverge significantly from equal stats.
Another example is that Baker's TD to Trey Palmer might have been worth fewer PFF points than the deep incompletion to Mike that should have been caught for a 55-yard TD. Sure, the stats say one was a long TD and one was an incompletion, but on Palmer's TD, Baker threw a 3 yard out to a wide open guy. On the incompletion, he put a dime on Mike's hands in bounds after the ball traveled 60 yards in the air. The latter was a far better throw despite the results, and so I would bet that his grade on the Palmer TD was like a 0.1 and his pass to Evans was like a 0.5.
Ken wrote: ↑Thu Jan 18, 2024 4:30 pm
In the situation I described, IIRC, Stroud would get a low value for doing what is expected of him... hitting a wide open guy, maybe a 0.1. And Love would get a high value like a 0.9 for the play that the stats say are equal. So you should expect the grading to diverge significantly from equal stats.
Another example is that Baker's TD to Trey Palmer might have been worth fewer PFF points than the deep incompletion to Mike that should have been caught for a 55-yard TD. Sure, the stats say one was a long TD and one was an incompletion, but on Palmer's TD, Baker threw a 3 yard out to a wide open guy. On the incompletion, he put a dime on Mike's hands in bounds after the ball traveled 60 yards in the air. The latter was a far better throw despite the results, and so I would bet that his grade on the Palmer TD was like a 0.1 and his pass to Evans was like a 0.5.
Every play is graded on -2 to +2. In this case, Baker would get 0.5 for "a well thrown short pass" to Trey Palmer. The Evans drop would have probably been +1.0 or +1.5 as +2.0 per PFF would be "Eli's pass to Mario Manningham on the game winning drive in the Super Bowl." While it was very deep accurate pass, the one from Eli needed elite placement. They have their play by play scoring very well documented on their site.
Let's take your Stroud vs Love TD comparison. Being 25 yards in the air, I would guess that would be at worse +0.5 but by distance could be +1.0. Loves pass doesn't hold up to the distance and difficulty of the Eli pass. So +1.0 to +1.5 due to difficulty. Since different people grade different games you see how someone could rate one way and the other the other way.
The problems that most people have is that all these +/-2 scores are added up into a 0-100 score and they don't tell how it is converted. They both had 21 passes and neither were sacked and both scrambled once. So that would have to be a lot of 0.5-1.0 easier/shorter throws for Stroud to have that big of a difference in scores. At first look that doesn't make sense. The second is as I mentioned above, the numbers are subjective based upon the person watching the game and personal bias may be involved.
I am not saying that they are wrong, but using a 0-10 scale as you guessed would make a much more sense than shaving of .5 point here and there than their scoring methods.
What I don’t understand about PFF is are you really telling me a human has objectively graded every player, in every NFL as well as every NCAAF game on every single play?
I don’t want to do the math but that seems like a lot of man hours to do in general, let alone to remain consistent and objective throughout all of it.
Kress wrote: ↑Thu Jan 18, 2024 12:35 pm
Using stats to judge any of this stuff is inherently flawed - especially with quarterbacks.
You have a play where the QB does everything right, but his linemen screw the pooch, and his receivers all stand around with their thumbs up there asses, so the guy miraculously avoids 3 pass rushers, scrambles away, and tosses the ball out of bounds. Freaking brilliant QB play for that situation. On the sheet it's an incomplete pass.
You have a play where the receiver gets behind the safety and the QB hits him in stride. 20-yard touchdown. But the only reason it's 20 is because that's how far away they were. The exact same play could have been 70, or 99, or 750 if they moved the goal line out there. But put that 20 on the stat sheet.
And of course the dump offs. QB slings the ball out to the running back in the flat. Easiest toss and catch that anyone could ever make. Running back makes 9 guys miss and goes 80 yards. Stat sheet rings up all 80. Even worse if it's one of those glorified hand-off tosses that goes for 80.
Exactly this. We've seen it a million times. Jameis' 30/30 season comes to mind. He would just wing it down there and trust Evans to beat the entire secondary and make the catch. Many times, Evans did. A 50 yard TD where Winston throws hope and prayer up and Evans goes over 3 guys to make the circus catch rates exactly the same as Mahomes dodging the entire defensive line and throwing a 50 yard laser while flying sideways that hits his gone in stride and in the hands. '
Facing a reeling Philadelphia Eagles team that didn’t have wide receiver A.J. Brown, the Tampa Bay Buccaneers knew that stopping the run would be immensely important to secure a victory.
The Buccaneers taking an early lead in the wild-card matchup surely changed the Eagles’ play-calling agenda, but Tampa Bay still managed to allow just 42 rushing yards on 15 carries. This was as big a reason as any that the Buccaneers are still playing while the Eagles head home.
Perhaps the most intriguing statistic to come out of Monday night’s game is the fact that former All-Pro linebacker Devin White played just one run-defense snap and 26 total snaps. A closer look at his production in the run game shows that this is no coincidence.
Devin White Career Run-Defense grades
Year Grade Rank*
2019 46.9 81/100
2020 38.3 82/96
2021 28.2 91/94
2022 36.7 83/90
2023 30.9 87/89
*among qualified LBs
White finished outside the top 80 in run-defense grade in each of his first four seasons, and he appears very likely to do so again this year. In fact, White owns a career 27.3 run-defense grade that ranks second worst of any linebacker who has played at least 100 snaps since he was drafted in 2019.
There is no doubt White brings solid pass-rush ability and makes some splash plays in coverage, but he’s never found consistency against the run.
Tampa Bay, despite some favorable traditional numbers, was struggling in run defense this season, woes that could’ve been the team's undoing. Coincidentally, things started to turn around in Week 13, which was the beginning of a stretch where White missed time due to injury.
Buccaneers Run-Defense Grades
Metric Weeks 1-12 Weeks 13-WC
Run-Defense Grade 52.9 (30th) 68.9 (12th)
Run-Defense Grade (LBs) 46.6 (32nd) 74.2 (9th)
White missed three games down the stretch and has played only 54 run-defense snaps since his return in Week 16. He’s been utilized in a manner that suits his strengths.
Over his past four games, White has rushed the passer 61 times, including 13 snaps against the Eagles, creating 10 pressures in the process. He also owns a league-leading 90.2 coverage grade over the past four weeks.
But if White has been excelling on passing downs, who has spearheaded Tampa Bay’s turnaround in the run game?
Look no further than former fifth-round pick K.J. Britt. The third-year man out of Auburn started playing significant snaps in Week 14. While he’s almost certainly not the next coming of Bobby Wagner, Britt has been playing the type of fundamentally sound football that the Buccaneers needed.
Britt, on an admittedly small sample size, has played like an above-average linebacker. His overall grade leads the team, while his run-defense grade leads all players in their front seven.
He’s also been adept in coverage, posting a 73.3 coverage grade while allowing only five catches for 40 yards across 83 coverage snaps.
As a result of Britt’s emergence, veteran stalwart Lavonte David has played his best football of the season, as well. Add it all up, and you have a linebacking corps that leads the league, by far, with an 84.2 overall grade since Week 14.
Britt’s willingness to play downhill in the run game has been the key. It’s not flashy, but he’s rarely out of position — a common problem for Devin White.
Tampa Bay now has three linebackers who are being utilized to their strengths. The combination of Britt's run-stopping prowess, David’s coverage skills and White’s utilization as a passing-down specialist has given the Buccaneers more defensive options.
Those three against a fantastic Detroit offensive line, which leads the league in overall grade, is quietly a fun matchup to look forward to this weekend. If Tampa Bay pulls off the upset, those linebackers will likely be a primary reason
Sdbucs wrote: ↑Thu Jan 18, 2024 7:07 pm
What I don’t understand about PFF is are you really telling me a human has objectively graded every player, in every NFL as well as every NCAAF game on every single play?
I don’t want to do the math but that seems like a lot of man hours to do in general, let alone to remain consistent and objective throughout all of it.
That's an issue, and the major flaw I do agree with is that PFF never knows what a player is SUPPOSED to do on a play. A guy could be doing everything right but the guy next to him blocks the wrong assignment, and then he's left to block air and the running back gets blown up in the backfield. The guy who actually blocked someone messed up, but on tape it looks like the other guy is just jogging and doing nothing.
Obviously, I think that those things are outliers. Usually a seasoned observer could know generally what is intended, so I've always taken the grades as a guideline, not an actual ranking.
One thing I do appreciate about PFF though is more advanced ways to look at the game. I believe they were the ones who coined the term "Defeats" which is a play that essentially ends a drive. So a tackle is more valuable on third down short of the line to gain vs one that gave up the first. Lavonte was always a top Defeater in his prime.
