Tweet
Originally posted by thelightningwill
View Post
-
Or Trout was that bad at RG. One of the worst performances Ive seen. Up there with Roundtree or Ellis.Hashtag thepowderblues -
More of a product of being on the right side IMO. He wasn't nearly as bad as when he played at LG during the season.
I still don't understand the decision to have him make his first ever start at RG in the playoffs.Comment
-
His play at LG got worse as the season went on but it was nowhere near as bad as his play at RG.Originally posted by Beerman View Post
I agree that I don't understand why they played him at RGHashtag thepowderbluesComment
-
Why is everything so black and white? Is it not possible (actually probable) that there are things that PFF and Outsiders do that is OK, even good, but things they do badly. The OL line stuff is crap, but stuff like yards after contact, yards after catch are NOT official stats, but I do think they do a good job. I think there is some use in collecting those, and I think that helps. I like that they are systematic, but again, I can go out in the world, measure something meaningful, and if I make those measurements badly, they mean little or nothing. Do whatever stats you want, but if you screw up the measures (in OL case the play by play grades), you aren't getting anything good out of it. (https://www.youtube.com/watch?v=G2y8Sx4B2Sk).
I never said Dunlap was awful. He is a good pass protector, and he really, really improved his blocking at the point of attack. He still struggled some with his backside protection. Troutman and Reinhart also did not do a great job of blocking on the backside. But it was often a case of our guys one does something well, the other does something poorly. And our TE really didn't block that well last season, front side or backside. I just pointed that Dunlap has some holes in his game. As does Clary, as does Fluker, as does Reinhart, and Hardwick, and really, every single guy who has ever played football, even the guys in the hall of fame. But to apply a blanket good/bad is just silly. I like the fact that they are probing for some truth and to find some objective measure of OL performance, but doing it badly just screws things up more then it helps.
My point is that most media pundits seem to put great stock in using the PFF numbers for OL, yet it's pretty clear that is really just due to the fact that there is a lack of options. Most media guys don't have the background or patience to watch a lot of games, or to know grade the average OL.
As I already stated, I think there is some value in the pass rushing stuff. I like yards after catch, and the yards after contact. I just think you need to verify it, which is something you have to do with all statistics anyway (and that comes from every stats teacher and fellow professionals who do this stuff for a living).
Again, the founder and many of their staff are British. And as one of the links points out, the founder has NEVER played football, which is a cardinal sin to many of us who have played. There are guys like Cossell, who can learn a lot by just watching film. Joel Buchsbaum was another great scout, who never played. But those guys spend a lot of time watching, spend a lot of time learning the game. And if they want to be transparent, then they need to let people see what they are doing. Buschbaum and Cossell both watched a lot of games, and used to cite a lot of examples. But PFF needs to let people see what they are doing if they are going to be truely credible. That is one thing the Sabermeterics guys in baseball had going for them, a rigorous peer review (and a sport that lends themselves to statistical analysis). Football, is always going to be a lot harder, and probably needs to rely a lot more on deterministic analysis, rather then stats.
Some pages talking about PFF
http://www.giants.com/videos/audio/B...7-3a3b5689b3d2 - talks some about their methodologies. First 10-14 minutes has that, the rest is some old giants stuff.
http://www.palmbeachpost.com/news/sp...s-ranki/nbLPG/ - points out that the guy is british and kind of full of himself, because every NFL team scouts itself
http://www.footballoutsiders.com/ext...n-and-charting - this uses stats to argue stats, and I disagree with some of it, but it is thought provoking.
http://blog.chron.com/texanschick/20...s-cornerbacks/ - similar to post above.Comment
-
Did anyone claim everything was black and white?Originally posted by Steve View PostHashtag thepowderbluesComment
-
Isn't he arguing the opposite. Maybe I misread his point.Originally posted by Steve View PostHashtag thepowderbluesComment
-
No, you did not miss a thing. In that post I said the exact opposite of what Steve is indicating that I said. I do not accept PFF's exact numbers as gospel, but I do think that when they have a player or players rated either very positively or very negatively, they probably are at least correct as to the general thumbs up or thumbs down proposition as to whether a player was good or not.Originally posted by oneinchpunch View Post
The rationale is that for very extreme analyses (either positive or negative) they would have to be very far off in their analysis for it to be wrong as at least a general indicator of whether or not a player was a bad or good player. While conceivable in theory, I think it is unlikely to be the case in reality.
In this case, Clary had the worst grade on the team at -18.3, including a stunningly bad -21.4 on running plays, while Dunlap was the second highest graded player on the team at +24.4, including a team best by a mile +16.5 on running plays, which was an incredible 13.2 points better than the next best offensive player on the team.
Applying the concept that I mentioned, I could believe that there could be some error as to both players, but in no way is it likely that Clary was somehow actually good against the run or that Dunlap was actually bad against the run since any mistake at all is just as likely to be for a player as it is against a player. Here, you would have to line up a bunch of errors all going in one direction in each case to get to the general result that Steve is suggesting. While some movement is possible, I think the movement that would be necessary for Steve's suggestion to be accurate is unlikely.Comment
-
There's really no point in repeating the same thing in repeated posts to every single persons response. We get your view on Clary. Repeating it over and over and over is what got you banned last time.Comment
-
I never cited the specific PFF numbers in this thread which really show just how much they would have to be off for Steve's theory about Dunlap and Clary to be correct.Originally posted by Beerman View Post
Also, if someone is going to say that I said X when I said -X, then it is definitely fair for me to point that out.Comment
-
Does it really matter arguing semantics? Do what you want, I'm just offering friendly advice, but it gets really old pretty much reading the same post (with different words) from you over and over. You make some valid points, but there's no need to keep arguing the point over and over after you made your initial point.Comment
-
If people are arguing that he didn't make his point, which you just said he did, wouldn't he be inclined to try to prove his point to the people that are calling him f'n stupid because of the point he's trying to make?Originally posted by Beerman View PostLast edited by oneinchpunch; 03-27-2014, 08:42 PM.Hashtag thepowderbluesComment
Comment