JESSEFEFFER Posted September 22, 2020 Posted September 22, 2020 An interesting take on him, these two games and his first two years. Funny how analytics people see the same thing and come to such different conclusions. Some voodoo math going on here. #1 vs. PFFs #12? Quick Reads on Josh 2
HappyDays Posted September 22, 2020 Posted September 22, 2020 What PFF does isn't analytics. They are subjectively grading players. Football Outsiders just takes raw data and contextualizes it for game situation and opponent. In their system a 5 yard pass on 3rd and 4 is graded better than a 10 yard pass on 3rd and 20, even though the 10 yard pass looks better in traditional statistics. 10 3
BuffaloHokie13 Posted September 22, 2020 Posted September 22, 2020 It's still very early in the year, but I've got him sitting at #2 through 2 weeks. I believe he's #2 in ANY/A as well, which I'd defer to until we're at least 4 games into the season. 1
Brennan Huff Posted September 22, 2020 Posted September 22, 2020 At the end of the day none of these analysts matter. We know what we have. We have the makings of a ***** legend. ESPN, PFF, CBS, etc none of them matter whether they have Allen ranked #1 or #32. WE ARE GOING TO THE ***** SUPER BOWL!!!! 4
Clyde Smith Posted September 22, 2020 Posted September 22, 2020 Ya'll don't like us but we winning should be our rally cry this season. 9 2 1 1
Bleeding Bills Blue Posted September 22, 2020 Posted September 22, 2020 (edited) 22 minutes ago, HappyDays said: What PFF does isn't analytics. They are subjectively grading players. Football Outsiders just takes raw data and contextualizes it for game situation and opponent. In their system a 5 yard pass on 3rd and 4 is graded better than a 10 yard pass on 3rd and 20, even though the 10 yard pass looks better in traditional statistics. https://www.pff.com/news/pro-how-pff-grades-quarterback-play They try to factor drops in - but the "difficulty and timing of the throw" piece is garbage. They basically don't count screens... which is stupid. We've run enough bad screens to know that there is something that makes them successful beyond it being an easy throw. They don't include how a good hard count can slow a pass rush. They don't factor how an audible or line adjustment can scheme a guy wide open. It's an easy throw so its worth less to them - despite good quarterbacking being the reason you had the time, or understanding of the defense. They try to value a QBs ability to buy time, but holding the ball is negative... despite no one being open... or a offensive lineman performing the assignment incorrectly. It's insanely subjective in the context of a football game. The QB is a leader of the unit, but a function of the offensive unit. There is so much context given to every play, primary reads to secondary reads, pre-snap to post snap adjustments. I actually appreciate what FO does though - it just contextualizes the stats. Not trying to re-invent the wheel. Edited September 22, 2020 by dneveu 1 3
34-78-83 Posted September 22, 2020 Posted September 22, 2020 2 minutes ago, dneveu said: https://www.pff.com/news/pro-how-pff-grades-quarterback-play They try to factor drops in - but the "difficulty and timing of the throw" piece is garbage. They basically don't count screens... which is stupid. We've run enough bad screens to know that there is something that makes them successful beyond it being an easy throw. They don't include how a good hard count can slow a pass rush. They don't factor how an audible or line adjustment can scheme a guy wide open. It's an easy throw so its worth less to them - despite good quarterbacking being the reason you had the time, or understanding of the defense. They try to value a QBs ability to buy time, but holding the ball is negative... despite no one being open... or a offensive lineman performing the assignment incorrectly. It's insanely subjective in the context of a football game. The QB is a leader of the unit, but a function of the offensive unit. There is so much context given to every play, primary reads to secondary reads, pre-snap to post snap adjustments. This x1000. To add they also don't factor in scheme or weather well enough.
mjt328 Posted September 22, 2020 Posted September 22, 2020 18 minutes ago, HappyDays said: What PFF does isn't analytics. They are subjectively grading players. Football Outsiders just takes raw data and contextualizes it for game situation and opponent. In their system a 5 yard pass on 3rd and 4 is graded better than a 10 yard pass on 3rd and 20, even though the 10 yard pass looks better in traditional statistics. Yep. Subjective is the key word. For instance... if a quarterback they like and respect happens to throw a completed ball into traffic, they will give that QB a positive grade for "trusting his playmakers" and "threading the needle" between defenders. If a QB they don't like (such as Josh Allen) does the same thing, he will get docked for making a "turnover worthy" throw. Their writers also make biased articles/statements in order to gain clicks and attention. Buffalo fans are notorious for defending their players, so PFF likes to give them low grades and rile them up. One of their writers admitted as much last season, when he claimed that Duck Hodges was better than Allen. Bottom line. Allen threw for over 400 yards, 4 touchdowns, 70% completion percentage, 0 turnovers, had a 4th quarter comeback to win the game, etc. They had him ranked outside the Top 10 for Sunday's games. Nothing else needs to be said. Their grades are worthless. 2 8
klos63 Posted September 22, 2020 Posted September 22, 2020 23 minutes ago, Clyde Smith said: Ya'll don't like us but we winning should be our rally cry this season. typical buffalo - please like us. I don't know what you watch, but I've seen endless praise for Allen. Many picked the Bills to win the division. But that doesn't fit the ' I don't get no respect' narrative. 1
Bleeding Bills Blue Posted September 22, 2020 Posted September 22, 2020 2 minutes ago, mjt328 said: Yep. Subjective is the key word. For instance... if a quarterback they like and respect happens to throw a completed ball into traffic, they will give that QB a positive grade for "trusting his playmakers" and "threading the needle" between defenders. If a QB they don't like (such as Josh Allen) does the same thing, he will get docked for making a "turnover worthy" throw. Their writers also make biased articles/statements in order to gain clicks and attention. Buffalo fans are notorious for defending their players, so PFF likes to give them low grades and rile them up. One of their writers admitted as much last season, when he claimed that Duck Hodges was better than Allen. Bottom line. Allen threw for over 400 yards, 4 touchdowns, 70% completion percentage, 0 turnovers, had a 4th quarter comeback to win the game, etc. They had him ranked outside the Top 10 for Sunday's games. Nothing else needs to be said. Their grades are worthless. I don't like hate what they do for WRs and CBs. Taking into account winning and losing routes where the ball isn't thrown can show you some interesting things. But at the end of the day... If you're in cover 3... you play behind the WR as you have no safety help. You will probably give up some underneath stuff unless you have really good eye discipline and closing speed. And cover 3 can look like man, just like cover 4 can look like man. and cover 6 can look like man. just because you are in a zone doesn't mean you aren't like... covering someone. this isn't madden, you don't just fade into a space of grass.
Coach Tuesday Posted September 22, 2020 Posted September 22, 2020 3 minutes ago, dneveu said: I don't like hate what they do for WRs and CBs. Taking into account winning and losing routes where the ball isn't thrown can show you some interesting things. But at the end of the day... If you're in cover 3... you play behind the WR as you have no safety help. You will probably give up some underneath stuff unless you have really good eye discipline and closing speed. And cover 3 can look like man, just like cover 4 can look like man. and cover 6 can look like man. just because you are in a zone doesn't mean you aren't like... covering someone. this isn't madden, you don't just fade into a space of grass. You are talking about PFF, not FO, right?
eball Posted September 22, 2020 Posted September 22, 2020 43 minutes ago, HappyDays said: What PFF does isn't analytics Thank you for posting this. PFF hires *anybody* to watch game tapes and assign arbitrary grades based upon what they see, with no knowledge of play call or contemplation of game situation. I know this because I actually "applied" to do this a few years ago on a lark as I thought it might be a fun hobby. There's no analytics going on.
ProcessAccepted Posted September 22, 2020 Posted September 22, 2020 I'm originally from Ireland. Here's an interesting fact that you may not know. We use the C word an awful lot over there. People like Sam Monson is the reason why. 2 4 2 1
Bleeding Bills Blue Posted September 22, 2020 Posted September 22, 2020 9 minutes ago, Coach Tuesday said: You are talking about PFF, not FO, right? Yes. FO would look at it contextual to the down and distance, not try and decipher the coverages etc. 1
Coach Tuesday Posted September 22, 2020 Posted September 22, 2020 Just now, dneveu said: Yes. FO would look at it contextual to the down and distance, not try and decipher the coverages etc. Right. They factor in both the result and the intended result. It's a much better model (though still not perfect). That said, I'm amazed to see Josh ranked #1 this week in QB DVOA, never thought I'd see the day.
marky Posted September 22, 2020 Posted September 22, 2020 1 hour ago, dneveu said: https://www.pff.com/news/pro-how-pff-grades-quarterback-play When you play a game of football, do you go through and count +1, -1, +0.5, 0, +2 on every play? To anyone who hasn't played, here's a clue what it's like. It feels like you are dying. Score early, and get out of there.
HappyDays Posted September 22, 2020 Posted September 22, 2020 1 hour ago, eball said: Thank you for posting this. PFF hires *anybody* to watch game tapes and assign arbitrary grades based upon what they see, with no knowledge of play call or contemplation of game situation. I know this because I actually "applied" to do this a few years ago on a lark as I thought it might be a fun hobby. There's no analytics going on. Yeah the whole point of analytics is to remove the human/subjective element of analyzing the game. PFF is the polar opposite of that. 1
JohnNord Posted September 22, 2020 Posted September 22, 2020 3 hours ago, JESSEFEFFER said: An interesting take on him, these two games and his first two years. Funny how analytics people see the same thing and come to such different conclusions. Some voodoo math going on here. #1 vs. PFFs #12? Quick Reads on Josh I love this paragraph. Aaron Schatz and FO were dunked on so bad their last past analysis of Allen. This is their way of acknowledging that they were probably wrong, without ever saying it. “We'll start with Josh Allen and the Bills. The consensus at Football Outsiders has always been pessimistic about Allen's chances for success. Our QBASE article in 2018 noted Allen's "horrifying" statistics at the University of Wyoming, and we referred to him as "a parody of an NFL quarterback prospect" in Football Outsiders Almanac that same year.”
Hapless Bills Fan Posted September 22, 2020 Posted September 22, 2020 30 minutes ago, HappyDays said: Yeah the whole point of analytics is to remove the human/subjective element of analyzing the game. PFF is the polar opposite of that. I'm not saying I disagree with you about the outcome. PFF's claim is actually that they are removing the human/subjective element of grading the QB's game - by having each play graded using subjective human observers. The same is true of a number of analytics systems. An early and much criticized one would be total QBR. Even Football Outsiders, while I like their system better as it is less "secret sauce" and closer to the actual data - has its subjective points as far as the game impact of a play. It's an interesting paradox.
Recommended Posts