Coach Tuesday Posted November 30, 2004 Posted November 30, 2004 Maybe someone a bit more numbers-savvy (NOT Fezmid) can de-compute what they're talking about over at www.footballoutsiders.com . Here's a link to a table indicating TWO different valuation methods for all current NFL QBs: http://www.footballoutsiders.com/stats/qb.php They provide a category called "DPAR," or Defense-Adjusted Points Above Replacement - as I understand it, represents the number of points scored when a QB puts the ball in play, as compared with an average replacement. Bledsoe ranks 25th here, with a "9.8" rating (again, explanation would be helpful). Another category is called "DVOA," or Defense-Adjusted Value Over Average which, I think, a "value" per play as compared with an average replacement. Bledsoe ranks 28th here, with a negative 5.9% "value." I think that means, he has almost 6% less value per play as compared with an average replacement. They also have interesting tables re: offensive, defensive and ST efficiency - no surprise, we're near the bottom on offense, and top-5 in D and ST (#1 ST). Interestingly, and relevant to a debate I was engaged in in another thread today, our offensive efficiency rating moved up ZERO after the Rams game. (flame away)
Alaska Darin Posted November 30, 2004 Posted November 30, 2004 Maybe someone a bit more numbers-savvy (NOT Fezmid) can de-compute what they're talking about over at www.footballoutsiders.com . Here's a link to a table indicating TWO different valuation methods for all current NFL QBs: http://www.footballoutsiders.com/stats/qb.php They provide a category called "DPAR," or Defense-Adjusted Points Above Replacement - as I understand it, represents the number of points scored when a QB puts the ball in play, as compared with an average replacement. Bledsoe ranks 25th here, with a "9.8" rating (again, explanation would be helpful). Another category is called "DVOA," or Defense-Adjusted Value Over Average which, I think, a "value" per play as compared with an average replacement. Bledsoe ranks 28th here, with a negative 5.9% "value." I think that means, he has almost 6% less value per play as compared with an average replacement. They also have interesting tables re: offensive, defensive and ST efficiency - no surprise, we're near the bottom on offense, and top-5 in D and ST (#1 ST). Interestingly, and relevant to a debate I was engaged in in another thread today, our offensive efficiency rating moved up ZERO after the Rams game. (flame away) 138672[/snapback] Sounds about right on the surface.
SJ Bills backer Posted November 30, 2004 Posted November 30, 2004 wow - you mean to tell me that despite our win streak, new found offensive production, and the re-animation of the mummy called Drew that we're STILL ranked near the bottom of the league in O? Yeh, I know - its a team game. Still, its hard to defend Bledsoe's production given these numbers. Thanks for this new slant on an old problem
DeeRay Posted November 30, 2004 Posted November 30, 2004 Basically, I take it that this computation, calculation or whatever concludes what 90% of us have been saying for well over a year... Drew sucks. Really? Who would have known? We had no idea.
Dan Gross Posted November 30, 2004 Posted November 30, 2004 Maybe someone a bit more numbers-savvy (NOT Fezmid) can de-compute what they're talking about over at www.footballoutsiders.com . Here's a link to a table indicating TWO different valuation methods for all current NFL QBs: http://www.footballoutsiders.com/stats/qb.php They provide a category called "DPAR," or Defense-Adjusted Points Above Replacement - as I understand it, represents the number of points scored when a QB puts the ball in play, as compared with an average replacement. Bledsoe ranks 25th here, with a "9.8" rating (again, explanation would be helpful). Another category is called "DVOA," or Defense-Adjusted Value Over Average which, I think, a "value" per play as compared with an average replacement. Bledsoe ranks 28th here, with a negative 5.9% "value." I think that means, he has almost 6% less value per play as compared with an average replacement. No, DVOA relates to the average team in the NFL. Bledsoe's performance is 5.9% below that of an "average" QB (and it's 23rd, not 28th). A team/player that is ranked 16-20 is going to be just above/below 0. If you actually read about the stats you were spouting, you'd see that DPAR was created to show performance relative to a replacement player. DPAR attempts to say "this is how much an improvement the guy is over a replacement in his position." A replacement player generally has a DVOA of -13.3%. That means Bledsoe is 9.8 "points" better than a replacement (presumably Shane Matthews, judging from the fact that they try to take into account who is the replacement and their performance, where possible). They also have interesting tables re: offensive, defensive and ST efficiency - no surprise, we're near the bottom on offense, and top-5 in D and ST (#1 ST). Interestingly, and relevant to a debate I was engaged in in another thread today, our offensive efficiency rating moved up ZERO after the Rams game. (flame away) 138672[/snapback] Well, duh it didn't move up after the Rams game. The rating is weighted by the opponent, and our offense played against the worst rated (by their measures) defense in that game, so whatever numbers came out would be weighted down heavily. It's actually good that we didn't lose any position. I would expect to see a jump based on the Bills facing a much better defense in Seattle...(they haven't updated the team stats yet, and I'm guessing they won't until they crunch in the MNF figures). I'm still trying to figure out the info behind some of their numbers here...They list Bledsoe as having 326 passes for 1811 yards 13 TD's 16 TO's, as of the end of week 12. The official stats list him as having 301 attempts with 27 sacks for 328 total drop backs. I'm guessing that, since they don't count spikes against incompletions that the spikes factor in somehow, and maybe the botched snap figures in as an attempt...The official stats put him at 2011 yards passing with 158 yards lost due to sacks, and 8 yards rushing...no matter how you add/subtract you don't come up with 1811. The TD/TO numbers are consistent...
Fake-Fat Sunny Posted November 30, 2004 Posted November 30, 2004 The responses just above go into the detail of what the stats mean, but I think the general meaning is even clearer. It maakes little difference to me what the stats say in terms of comparing Bledsoe to an average replacement as the replacements in the real world we would have for Bledsoe are far less than average. Matthews would have been at home on the couch if not for injury and even the most ardent of JP supporters do not attempt to argue that he would have better production than Bledsoe, they instead argue lets have him make his mistakes now so he canlearn the game and be ready to compete next year. This notion is one which demonstrates that TD made an error in never getting anything resembling an NFL level back-up for Bledsoe.
Coach Tuesday Posted November 30, 2004 Author Posted November 30, 2004 I'm still trying to figure out the info behind some of their numbers here...They list Bledsoe as having 326 passes for 1811 yards 13 TD's 16 TO's, as of the end of week 12. The official stats list him as having 301 attempts with 27 sacks for 328 total drop backs. I'm guessing that, since they don't count spikes against incompletions that the spikes factor in somehow, and maybe the botched snap figures in as an attempt...The official stats put him at 2011 yards passing with 158 yards lost due to sacks, and 8 yards rushing...no matter how you add/subtract you don't come up with 1811. The TD/TO numbers are consistent... 138978[/snapback] No need for the barbs, Gross, I was asking for help deciphering. Man, the animosity around here... Anyhow - I think the reason for the stat discrepancy you cite has to do with them treating QB rushes as "passes" - that's how I read their formula, but I could be wrong.
Dan Gross Posted November 30, 2004 Posted November 30, 2004 No need for the barbs, Gross, I was asking for help deciphering. Man, the animosity around here... Anyhow - I think the reason for the stat discrepancy you cite has to do with them treating QB rushes as "passes" - that's how I read their formula, but I could be wrong. 139097[/snapback] Well, you told me to flame you! The number of attempts they list is less than his "drop back" count (attempted passes + sacks). I believe they don't count spiked passes in their stats, but I believe he has had more than two. As I stated he has 8 yards on 15 rushing attempts...still just trying to figure out how they do the math to come up with the 200 yard difference in yardage. I'm not really a fan of "all year" stats anyway because they give no credit for progress. For example, this team has come away with points every time they've hit the red zone since the Baltimore game (at which point they had scored points only 2/3 of the time, and yes their D and ST were giving them good field position back then as well). That, to me, is progress. Then there's the number of sacks. That's a better example, because if you look at the official stats for the year, the Bills are at the "lower-middle" of the pack, but if you look game by game, and separate the first Q of the season from the rest, you'll notice that 18 of the 28 sacks came in the first four games, and the current "running sack rate" is up there with some of the best teams in the league.
34-78-83 Posted November 30, 2004 Posted November 30, 2004 Nerds across the land can try and try and try some more til they're blue in the face but there is , nor ever will be , a true and completely fair way to rate ANY individual perfomance of ANY position on ANY NFL team based on statistics. It is a team sport through and through, and too often there are hidden factors that contribute to both the success and failures of any given player. The only true way to e v a l a player is on game film, by a qualified viewer, which none of us truly are. With that said, I still find some of these statistical breakdowns interesting and somewhat useful at times.
AKC Posted November 30, 2004 Posted November 30, 2004 We need to head right over to the Falcons board to explain to their fan base why Michael Vick is such an obvious recipe for disaster based upon this statistical analysis. There may still be time to save their season by getting Matt Schaub the balance of the team's snaps in '04.
Dan Gross Posted November 30, 2004 Posted November 30, 2004 With that said, I still find some of these statistical breakdowns interesting and somewhat useful at times. 139450[/snapback] Especially when they seem to support whatever point you're trying to make...
Guest Guest_Coach_Tuesday Posted November 30, 2004 Posted November 30, 2004 We need to head right over to the Falcons board to explain to their fan base why Michael Vick is such an obvious recipe for disaster based upon this statistical analysis. There may still be time to save their season by getting Matt Schaub the balance of the team's snaps in '04. 139469[/snapback] Actually, interestingly, the guys at that site point out that one of the "anomolies" they've discovered is that mobile qbs tend to get sacked more often than "statues," because they're running around trying to make plays (then again, even those mobile qbs - Culpepper, Vick - don't get sacked as often as Bledsoe/Warner/Collins per their stats). Regardless - I disagree with those of you who think the stats aren't useful. The problem is just that they're not specific enough, which these guys admit. What some of you are trying to say, which is legit, is that you can't measure from those stats whether the values they assign are due to the QB, the protection, the receivers not getting open, or some combination thereof. They don't measure, for example, a QB's effectiveness off of play action, or throwing to his right versus across his body, etc. The reason they don't measure that stuff is because, as of now, the NFL scorers don't keep those stats. Eventually I think you'll see a greater appreciation in the NFL for the kind of "sabremetrics" that have become commonplace in MLB - when that happens, you'll see more specific, and therefore more useful, measurements.
Guest Guest_Coach_Tuesday Posted November 30, 2004 Posted November 30, 2004 SORRY. Check this out - they link to TBD - apparently they were following our discussion yesterday: http://www.footballoutsiders.com/ramblings.php?p=2053&cat=3 Speaking of new people checking out the site, I like to do is go through the referrer logs to see who’s talking about us on various message boards. I’m used to seeing us linked on various Patriots, Packers, and Redskins boards, but there are some new ones this week. So I hope somebody from Two Bills Drive is reading because I’m going to answer some of your questions about our stats here even though you didn’t even ask directly! Buffalo’s offensive rating did in fact climb significantly after the Seattle game, from -15.3% to -8.6%. Someone asked why we have Bledsoe listed with 1811 yards instead of 2011 yards and 326 pass attempts instead of 301 pass attempts. We count sacks and aborted snaps in the passing numbers, but we do not count spikes to stop the clock. Quarterback rushing is not the reason for the discrepancy, either; you’ll notice a separate table for rushing numbers. I have Bledsoe listed with 29 sacks for 170 yards, as opposed to the NFL numbers of 27 sacks for 158 yards. I’m not sure why. Sometimes the league makes a stat change weeks later without telling anyone. Other times the totals just don’t match what you get when you add up play-by-play (last year, for example, a Jamal Lewis fumble just disappeared from his totals). I also list Bledsoe with 42 yards lost due to three aborted snaps. Note that when the aborted snap fumble is charged to the center, that doesn’t show up in the quarterback stats on our site even though it does still count against that team’s passing offense rating. When DPAR (Defense-adjusted Points Above Replacement) is comparing a player to replacement level, it is comparing that player to a generic replacement level across the league for that position, not to the specific second-stringer on his team. In other words, I’m saying Drew Bledsoe has been worth approximately 9.8 points more than some generic backup, not Shane Matthews specifically. For you non-Bills fans, yes, Shane Matthews is still in the league. Incredible, I know.
Dan Gross Posted November 30, 2004 Posted November 30, 2004 (edited) Looks like the offense moved up two spots this week... Edit: and gee, I had been refreshing the "raw listing" to see if it changed. Thanks for the pointer to the article on the update, Coach. Hope you football outsider guys are not insulted that I'm just too damn lazy to send an e-mail! Edited November 30, 2004 by Dan Gross
34-78-83 Posted November 30, 2004 Posted November 30, 2004 Especially when they seem to support whatever point you're trying to make... 139501[/snapback] I'm not sure if that was sarcasm or not but you nailed the key word. "Support", NOT "prove"....
JinVA Posted November 30, 2004 Posted November 30, 2004 Nerds across the land can try and try and try some more til they're blue in the face but there is , nor ever will be , a true and completely fair way to rate ANY individual perfomance of ANY position on ANY NFL team based on statistics. It is a team sport through and through, and too often there are hidden factors that contribute to both the success and failures of any given player. The only true way to e v a l a player is on game film, by a qualified viewer, which none of us truly are. With that said, I still find some of these statistical breakdowns interesting and somewhat useful at times. 139450[/snapback] Not a big John Clayton fan I see.
34-78-83 Posted November 30, 2004 Posted November 30, 2004 They don't measure, for example, a QB's effectiveness off of play action, or throwing to his right versus across his body, etc. The reason they don't measure that stuff is because, as of now, the NFL scorers don't keep those stats. Eventually I think you'll see a greater appreciation in the NFL for the kind of "sabremetrics" that have become commonplace in MLB - when that happens, you'll see more specific, and therefore more useful, measurements. 139516[/snapback] That's true, but there are many many more important and general things that cannot be statistically measured as well and never will, like: Was the pass completion a correct read progression or was it forced in.? Did the halfback make those three yards on his own, or did he stumble when he should have gained 3 more? Did the WR run a good route or did he misread the safety? Did the quarterbook pull his teammates together in that two minute drive and instill confidence in them?
Guest Guest Posted November 30, 2004 Posted November 30, 2004 That's true, but there are many many more important and general things that cannot be statistically measured as well and never will, like: Was the pass completion a correct read progression or was it forced in.? Did the halfback make those three yards on his own, or did he stumble when he should have gained 3 more? Did the WR run a good route or did he misread the safety? Did the quarterbook pull his teammates together in that two minute drive and instill confidence in them? 139940[/snapback] Except for the last category, I think the stats in essence DO cover the categories you list - it's all about the law of averages. You're comparing a specific offense's (and QB's) performance at a specific down-and-distance (and field spot) against the NFL average offense or replacement player. Doesn't really matter if it was the correct read if it produced the appropriate "value" for that down-and-distance (and, it should be tautological - isn't the correct read more often than not the one that gives you the first down?). Same goes for the WR's route. Also, it doesn't matter if the halfback stumbled on one play - unless he stumbles more than the average halfback, it all comes out in the wash - he's being compared to the average halfback, which means the average number of stumble-forwards over a season.
34-78-83 Posted November 30, 2004 Posted November 30, 2004 Except for the last category, I think the stats in essence DO cover the categories you list - it's all about the law of averages. You're comparing a specific offense's (and QB's) performance at a specific down-and-distance (and field spot) against the NFL average offense or replacement player. Doesn't really matter if it was the correct read if it produced the appropriate "value" for that down-and-distance (and, it should be tautological - isn't the correct read more often than not the one that gives you the first down?). Same goes for the WR's route. Also, it doesn't matter if the halfback stumbled on one play - unless he stumbles more than the average halfback, it all comes out in the wash - he's being compared to the average halfback, which means the average number of stumble-forwards over a season. 139962[/snapback] Well I suppose over the course of a season that arguement would hold water, not so much on a game by game account. With that said, a coach's eye is still different and I'm sure there is a player or ten out there that many perceive as a "great" player both visually and statistically who really is benefiting from strong play at another position, or vice versa.
Recommended Posts