Thurman#1 Posted October 23, 2019 Posted October 23, 2019 2 minutes ago, GunnerBill said: There are a lot of advanced analytics that they like but it has become the PFF haters mantra that "oh the NFL teams don't care about their grades." Well run NFL teams use every source of information they can get. There is no salary cap on this stuff so they get their hands on whatever they can and then they sift through to work out the bits that they can use. Makes total sense.
Aireskoi Posted October 23, 2019 Posted October 23, 2019 The problem with stats is that they are not watching the actual games. You can learn a lot from watching the whole game, but PFF and FO (Football Outsiders) types of services don't have time. You learn much more by being at the game, talking to the kids, seeing how they react on the sidelines, with their teammates, good times and bad, you hear the ball whistle, you feel the impact of big plays vs. run of the mill passes.... Real scouting will trump these stat based analysis, we fans never get to see the whole picture like scouts do. That's how Beane and McDermott knew Josh was their man, they didn't only look at numbers. PFF and FO were surprised that Josh could run the ball, that's kind of a big miss for stat guys...but you wouldn't see that watching cut ups of every throw. 2
H2o Posted October 23, 2019 Posted October 23, 2019 13 hours ago, Reader said: Hey all, I know the general sentiment here about PFF, but as a guy who enjoys numbers and statistics I figured I would bring up a point anyways. As you know PFF tends to be quite critical of Allen where our forum debates to what extent that critique is warranted, something I'd mention though is to what extent PFF seems to neglect mentioning Allen's improvement these past couple of weeks. By my count, and I don't have elite so I can't go back and check, Allen's grade has trended up significantly since the Patriots game. Note the game by game scores are guesses based of Allen's overall grade, so technically they are almost non-guesses thanks to math. Week 1 score-57ish slightly below average Total rating 57ish Week 2 score-75ish strong game by Allen Total rating 66ish Week 3 score-52ish below average game for Allen, Total rating 60ish Week 4 score-34ish extremely poor game for Allen, Total rating 50ish Unsurprisingly we see Allen have an extremely bad game against the Patriots similar to everyone else who has played them. Week 5 score-70ish strong game by Allen Total rating 57ish Week 6 score-75ish strong game by Allen Total rating 60ish. Allen has played two great games by their measures and yet one hardly hears a peep about that. In addition to that Allen now has an average grade as a QB, especially important noting that last year his passing grade at the end of the season is poor. The Bills still have a lot of kinks to work out, but the fact is that Allen seems to be improving and if that is something he continues to do we have a lot to feel excited about even if the experts who see it hardly comment on it.
Putin Posted October 23, 2019 Posted October 23, 2019 11 hours ago, r00tabaga said: I saw PFF and stop'd reading.... Exactly !! Why keep re-posting this garbage ??
Stank_Nasty Posted October 23, 2019 Posted October 23, 2019 (edited) 6 hours ago, Thurman#1 said: Yeah, PFF didn't mention kumquats either. Damn them. They leave so many things out. Whenever you hear, "they left out," it's almost always something not necessarily relevant, but important to the kibitzer. Same in this case, IMO. And it really doesn't make much sense to say PFF doesn't know what they're doing. They do. And the proof of that is that nearly every NFL team buys their materials. They wouldn't do that if they didn't know what they are doing. PFF gets a lot of crap here and most of it is of the "I don't like the message, and here's the messenger standing right here ... wait, I've got an idea" variety. Agreed that Allen is trending up a bit the last two games, but this last game was a lot better than the Pats game but not really a very good game for him. He still has a long way to go, looks to me. So then you agree that heading into last week the Steelers are the 13th best team in the nfl and their were 19 teams in the nfl better than the bills including the broncos and titans? .... That right there tells me PFF has some serious flaws. It’s merely a small tool in the toolbox. .... also just last week PFF mentioned Allen in the top 5 of one of their made up stats “adjusted comp%”, only to throw their own stats credibility under the bus and claim he still hasn’t been accurate because his “perfect throw %” is low.... they’re discrediting their own freaking stats in order to find a way not to be wrong about Allen. That’s sort of ridiculous. Edited October 23, 2019 by Stank_Nasty
Thurman#1 Posted October 23, 2019 Posted October 23, 2019 (edited) 14 minutes ago, Stank_Nasty said: So then you agree that heading into last week the Steelers are the 13th best team in the nfl and their were 19 teams in the nfl better than the bills including the broncos and titans? .... That right there tells me PFF has some serious flaws. It’s merely a small tool in the toolbox. It's not perfect. Nothing is. It's just very good. The PFF power rankings use an ELO system, which is much affected by strength of opposition, or perceived strength of opposition, really. Power rankings are always going to differ, whatever your system, as you can't correctly adjust for strength of schedule, ball bounces, etc. Especially fairly early in the season, it's hard to know what's real and what's not. Did the Bills look like the, for example, 7th best team in the league against the Dolphins? Again, how come a very large majority of NFL teams are buying their stuff if they're not good. A small tool in the toolbox? Fair enough. Some serious flaws? In power rankings, everyone does, especially this early. Me, I think the Bills should be higher. But disagreeing with one thing about PFF, or rather one particular system PFF uses for fun to predict the future, doesn't show a major problem. Edited October 23, 2019 by Thurman#1
oldmanfan Posted October 23, 2019 Posted October 23, 2019 8 minutes ago, Thurman#1 said: It's not perfect. Nothing is. It's just very good. And power rankings are always going to differ, as you can't adjust for strength of schedule, ball bounces, etc. Especially fairly early in the season, it's hard to know what's real and what's not. Did the Bills look like the, for example, 7th best team in the league against the Dolphins? Again, how come a very large marjority of NFL teams are buying their stuff if they're not good. A small tool in the toolbox? Fair enough. Some serious flaws? In power rankings, everyone does, especially this early. You keep asking why teams buy their product. Perhaps you could give us an answer?
Thurman#1 Posted October 23, 2019 Posted October 23, 2019 (edited) 4 minutes ago, oldmanfan said: You keep asking why teams buy their product. Perhaps you could give us an answer? I already did, I think, but I'll try again. Hmm. Because they're good at what they do. And the teams recognize that. Edited October 23, 2019 by Thurman#1
DCOrange Posted October 23, 2019 Posted October 23, 2019 14 hours ago, Reader said: Hey all, I know the general sentiment here about PFF, but as a guy who enjoys numbers and statistics I figured I would bring up a point anyways. As you know PFF tends to be quite critical of Allen where our forum debates to what extent that critique is warranted, something I'd mention though is to what extent PFF seems to neglect mentioning Allen's improvement these past couple of weeks. By my count, and I don't have elite so I can't go back and check, Allen's grade has trended up significantly since the Patriots game. Note the game by game scores are guesses based of Allen's overall grade, so technically they are almost non-guesses thanks to math. Week 1 score-57ish slightly below average Total rating 57ish Week 2 score-75ish strong game by Allen Total rating 66ish Week 3 score-52ish below average game for Allen, Total rating 60ish Week 4 score-34ish extremely poor game for Allen, Total rating 50ish Unsurprisingly we see Allen have an extremely bad game against the Patriots similar to everyone else who has played them. Week 5 score-70ish strong game by Allen Total rating 57ish Week 6 score-75ish strong game by Allen Total rating 60ish. Allen has played two great games by their measures and yet one hardly hears a peep about that. In addition to that Allen now has an average grade as a QB, especially important noting that last year his passing grade at the end of the season is poor. The Bills still have a lot of kinks to work out, but the fact is that Allen seems to be improving and if that is something he continues to do we have a lot to feel excited about even if the experts who see it hardly comment on it. Just FWIW, I tried to track it this way last year...PFF retroactively goes back and adjusts grades from prior games. I’m not sure why or when, maybe some sort of adjustment related to the level of competition as they learn more about each team, but for example, there were times last year where the only way Allen or Edmunds could have possibly jumped up the way they did is if they received a grade higher than 100, which isn’t possible. Or there were cases where PFF tweeted out their exact grade for a given week but the math wouldn’t add up. 1
oldmanfan Posted October 23, 2019 Posted October 23, 2019 5 minutes ago, Thurman#1 said: I already did, I think, but I'll try again. Hmm. Because they're good at what they do. And the teams recognize that. How about specifics? Do you think teams actually use their grades on a player in a given game to base decisions on, for example, who to start and who to sit? Who to draft? Who to trade for? What data do they actually use from PFF? How do you answer the criticisms of their data, in that the grades they give players are complicated by not knowing the play call, not knowing the assignment, and such? Other wise, your statement that "they're good at what they do" comes across simply as you slavishly upholding another media-related site as faultless.
Stank_Nasty Posted October 23, 2019 Posted October 23, 2019 (edited) 34 minutes ago, Thurman#1 said: It's not perfect. Nothing is. It's just very good. The PFF power rankings use an ELO system, which is much affected by strength of opposition, or perceived strength of opposition, really. Power rankings are always going to differ, whatever your system, as you can't correctly adjust for strength of schedule, ball bounces, etc. Especially fairly early in the season, it's hard to know what's real and what's not. Did the Bills look like the, for example, 7th best team in the league against the Dolphins? Again, how come a very large majority of NFL teams are buying their stuff if they're not good. A small tool in the toolbox? Fair enough. Some serious flaws? In power rankings, everyone does, especially this early. Me, I think the Bills should be higher. But disagreeing with one thing about PFF, or rather one particular system PFF uses for fun to predict the future, doesn't show a major problem. I would debate it’s not “for fun” for them. It’s a formula they stand by and it had legitimately bad teams significantly ahead of good ones. thats hard to take seriously. it’s small tool. Not much more. Whether it’s a big one or small one teams will use it if it’s a tool. So I don’t know why so much weight is being put on that. Edited October 23, 2019 by Stank_Nasty
Stank_Nasty Posted October 23, 2019 Posted October 23, 2019 40 minutes ago, Stank_Nasty said: .... also just last week PFF mentioned Allen in the top 5 of one of their made up stats “adjusted comp%”, only to throw their own stats credibility under the bus and claim he still hasn’t been accurate because his “perfect throw %” is low.... they’re discrediting their own freaking stats in order to find a way not to be wrong about Allen. That’s sort of ridiculous. Quoting myself here so I don’t have to type it out again.... how am I supposed to take them seriously when they will throw their own formulas under the bus so easily just a o remain correct on a player narrative?
oldmanfan Posted October 23, 2019 Posted October 23, 2019 2 minutes ago, Stank_Nasty said: I would debate it’s not “for fun” for them. It’s a formula they stand by and it had legitimately bad teams significantly ahead of good ones. thats hard to take seriously. it’s small tool. Not much more. Whether it’s a big one or small one teams will use it if it’s a tool. So I don’t know why so much weight is being out on that. My guess is that teams buy it because it saves them money over hiring a group of interns to break down film all the time. It would be interesting to know what they actually take from the data. From perusing some Google documents about PFF it seems their player grades are targeted more at fantasy football fans who think it offers them some advantage. On the other hand it seems guys like Belichick don't place much stock in their grades since the pff personnel have no idea about play calls, responsibilities of players of a given call, etc.
D. L. Hot-Flamethrower Posted October 23, 2019 Posted October 23, 2019 1 hour ago, Aireskoi said: The problem with stats is that they are not watching the actual games. You can learn a lot from watching the whole game, but PFF and FO (Football Outsiders) types of services don't have time. You learn much more by being at the game, talking to the kids, seeing how they react on the sidelines, with their teammates, good times and bad, you hear the ball whistle, you feel the impact of big plays vs. run of the mill passes.... Real scouting will trump these stat based analysis, we fans never get to see the whole picture like scouts do. That's how Beane and McDermott knew Josh was their man, they didn't only look at numbers. PFF and FO were surprised that Josh could run the ball, that's kind of a big miss for stat guys...but you wouldn't see that watching cut ups of every throw. Yes, and that is why groups that are selling their analytics will NEVER give Josh credit. It will at best be a backhanded, forced compliment when there is no where else to go.
D. L. Hot-Flamethrower Posted October 23, 2019 Posted October 23, 2019 23 minutes ago, DCOrange said: Just FWIW, I tried to track it this way last year...PFF retroactively goes back and adjusts grades from prior games. I’m not sure why or when, maybe some sort of adjustment related to the level of competition as they learn more about each team, but for example, there were times last year where the only way Allen or Edmunds could have possibly jumped up the way they did is if they received a grade higher than 100, which isn’t possible. Or there were cases where PFF tweeted out their exact grade for a given week but the math wouldn’t add up. Wow I don't like that. I wonder how they justify that one?
GunnerBill Posted October 23, 2019 Posted October 23, 2019 1 hour ago, Aireskoi said: The problem with stats is that they are not watching the actual games. You can learn a lot from watching the whole game, but PFF and FO (Football Outsiders) types of services don't have time. You learn much more by being at the game, talking to the kids, seeing how they react on the sidelines, with their teammates, good times and bad, you hear the ball whistle, you feel the impact of big plays vs. run of the mill passes.... Real scouting will trump these stat based analysis, we fans never get to see the whole picture like scouts do. That's how Beane and McDermott knew Josh was their man, they didn't only look at numbers. PFF and FO were surprised that Josh could run the ball, that's kind of a big miss for stat guys...but you wouldn't see that watching cut ups of every throw. You think PFF don't watch the games? 2
buffalobillswin Posted October 23, 2019 Posted October 23, 2019 PFF doesn’t like my QB so I think they stink! 1
DCOrange Posted October 23, 2019 Posted October 23, 2019 1 hour ago, D. L. Hot-Flamethrower said: Wow I don't like that. I wonder how they justify that one? I'm guessing it's one of two things (or possibly both): 1. I know they have "initial grades" that are released like the day after the game but then they go back and assign final grades later on (I think they watch each game multiple times or something, not entirely sure of their process), so it's possible that the grades that they tweet out are sometimes the initial rating and therefore throw things off if you plug those in as the grades for a given week. 2. As I mentioned, I kinda wonder if there's some sort of adjustment that occurs based on the level of competition. Not sure; that doesn't sound like the way PFF generally explains their grading process, but something definitely occurs that throws the math off. 1 hour ago, Stank_Nasty said: Quoting myself here so I don’t have to type it out again.... how am I supposed to take them seriously when they will throw their own formulas under the bus so easily just a o remain correct on a player narrative? That's not throwing their own stats under the bus. Adjusted completion percentage is still just a completion percentage at the end of the day, and most logical people would agree that completion percentage does not necessarily mean a QB is accurate. PFF has advanced beyond that and now charts the final location of each pass relative to the receiver and defender (and I'm guessing/hoping they also account for whether or not the pass causes the receiver to slow down, if the receiver just takes a poor angle, etc.) to determine how accurate a pass is. It's two completely different things. Adjusted completion percentage is just telling you what a QB's completion percentage would be if you adjust for dropped passes, spikes, throwaways, and passes that are batted down at the line of scrimmage. Their newer accuracy ratings are actually indicative of how accurate a QB is.
Rock-A-Bye Beasley Posted October 23, 2019 Posted October 23, 2019 9 hours ago, TheFunPolice said: Last year we heard how Allen couldn't hit the short easy passes Now he can't hit the deep balls Lol What’s funny? He hasn’t hit a single one.
GunnerBill Posted October 23, 2019 Posted October 23, 2019 1 hour ago, DCOrange said: Just FWIW, I tried to track it this way last year...PFF retroactively goes back and adjusts grades from prior games. I’m not sure why or when, maybe some sort of adjustment related to the level of competition as they learn more about each team, but for example, there were times last year where the only way Allen or Edmunds could have possibly jumped up the way they did is if they received a grade higher than 100, which isn’t possible. Or there were cases where PFF tweeted out their exact grade for a given week but the math wouldn’t add up. I have done the same. And you are right there are times it doesn't quite seem to stack up. I think it is because they don't do it game by game. Rather they weight by snap counts. So say week 1 you play 40 snaps and then week 2 and week 3 you play 20 snaps. Your overall grade is more heavily weighted by your week 1 performance. But then week 4 you play 50 snaps. Suddenly the extent to which your week 1 grade is weighted in the overall is significantly reduced.
Recommended Posts