Page 1 of 2 12 LastLast
Results 1 to 12 of 23
  1. #1
    Join Date
    Aug 2011
    Posts
    11,089
    Blog Entries
    7

    ESPN's QBR with respect to Flacco

    Obviously ESPN has been trying to push their QBR stat the past few years.

    For whatever reason, this stat does not seem to favor Flacco's game. I know it is a proprietary formula, and all of the components may not be known. But looking at it a little more closely, I think that it has to severely punish QB sacks and/or QB fumbles, which is why Flacco's QBR is generally lower.

    For example, the 2012 NE v Balt playoff game, which was sort of Joe's playoff coming out party before his big year this year, shows a stupid result via QBR. After that game, nearly everyone said he outplayed Brady. It was clearly evident on the field.

    Here are the stats from that game:

    Flacco:
    22/36, 306 yards (8.5ypc), 2 TD, 1 INT, 3 sacks for -24 yards
    QBR 43.0

    Brady:
    22/36, 239 yards (6.6ypc), 0 TD, 2 INT, 1 sack for -5 yards
    QBR: 71.5

    There is no way Brady played better, and there is no way, when comparing those stats, Brady should have a ~30 point higher QBR. Hell they have identical completions/attempts (22/36). One QB scored 2 TDs. The other scored none. On QB had 2 INTS, the other had 1. The only big difference is sacks and sack yardage.

    Now, let's look at how those same stats fair in the traditional QB rating stat:
    Flacco: 95.4
    Brady: 57.5

    This is by far a much more accurate representation of QB play in that game. It reflects reality.

    I can't respect the ESPN QBR stat in view of this outcome (and I am sure there are many other examples of how QBR falters compared to traditional QB rating).

    It seems pretty clear that the ESPN QBR places way too much emphasis on sacks, to the point where the resultant rating will frequently not mirror reality. I know they are trying to pitch it as being a more accurate representation of QB play, but when the stat produces results that are so far removed from reality, it is hard to give it any respect. Especially when sacks are frequently a product of OL play. To weight a QB formula so heavy with respect to sacks seems silly. And produces silly results.





  2. #2
    Join Date
    Aug 2011
    Posts
    11,089
    Blog Entries
    7

    Re: ESPN's QBR with respect to Flacco

    I would like Flacco to fumble less. But I know part of the reason for his sack totals are OL play, in addition to the routes called by Cameron that forced Joe to wait and wait for a chance to throw the ball. Sure a lot of his sacks are on him specifically, but a lot of them are also due to the playcalling and OL.

    The game in HOU last year was a perfect example of Cam screwing Joe with horrible play calls. Joe could barely get set before HOU was on him.

    The fact that QBR seemingly weights sacks so high means it will penalize certain QBs and playcalling (e.g. deep/sideline routes more prone to sacks due to the time they take to develop) and reward dink and dunk offenses like NE, etc. that allow for quick release. A stat shouldn't be so heavily biased to a particular style of play.





  3. #3

    Re: ESPN's QBR with respect to Flacco

    I may be mistaken, but I think that the formula somehow quantifies "clutchness." I don't know exactly how it works either, but I think if Evans held onto the pass and the Ravens won, Flacco's QBR would have been significantly higher, maybe even higher than Brady's, because it would have been "clutch."

    Having said that, I may not be right at all, but that's my impression based on what I read about it when it came out.





  4. #4

    Re: ESPN's QBR with respect to Flacco

    I don't think the fumbles and sacks are the issue. The issue is that it is a subjective accumulation of grades on every single play. It is not a "formula" that uses the game totals/stats as inputs.

    So for example, I suspect on Joe's 42-yard completion to Torrey Smith off of the play-action fake, Joe received 0 credit because the ball was underthrown (perhaps even received negative credit or was penalized, though I doubt it). Literally zero credit. And for Lee Evans twisting turning sideline catch, Joe probably received less credit than the play result indicated because Evans' got more credit than the WR normally does on a 'standard' completion of that length.

    So it really is a matter of a subjective and probably biased (I would say biased in favor of the 'name' QBs more than biased against any other specific QB) evaluator deciding how much of any given play's result is a result of the QB vs. the WR (or line, etc.). So a perfectly thrown ball gets more QB credit than a slightly off ball. A checkdown, that any NFL QB can make that goes for serious yardage likely is ignored by the ESPN grader (given no credit). A boneheaded INT is penalized more than an INT that is a result of the WR having the ball bounce off his hands. Sanchez's butt-fumble is penalized more than a jailbreak blindside sack-fumble, etc.

    It is a play-by-play evaluation. Overall game stats are not a factor, except insofar as they are "earned" by the QB, vs. given to the QB by circumstance or teammates' efforts, etc.





  5. #5
    Join Date
    Aug 2011
    Posts
    11,089
    Blog Entries
    7

    Re: ESPN's QBR with respect to Flacco

    Quote Originally Posted by redmike34 View Post
    I may be mistaken, but I think that the formula somehow quantifies "clutchness." I don't know exactly how it works either, but I think if Evans held onto the pass and the Ravens won, Flacco's QBR would have been significantly higher, maybe even higher than Brady's, because it would have been "clutch."

    Having said that, I may not be right at all, but that's my impression based on what I read about it when it came out.
    I think this is accurate.

    I can get behind rewarding "clutchness". But not if it produces results that clearly are out of line with what actually happened.





  6. #6

    Re: ESPN's QBR with respect to Flacco

    Quote Originally Posted by redmike34 View Post
    I may be mistaken, but I think that the formula somehow quantifies "clutchness." I don't know exactly how it works either, but I think if Evans held onto the pass and the Ravens won, Flacco's QBR would have been significantly higher, maybe even higher than Brady's, because it would have been "clutch."

    Having said that, I may not be right at all, but that's my impression based on what I read about it when it came out.
    I think the "clutchness" factor is better described as "meaningfulness."

    And I am not sure the QB gets penalized for a team loss or even a WR drop (so I doubt that Joe's QBR ends up much higher even if Evans had caught that ball).

    But stats accumulated in a game that is out of hand or already decided (like meaningless TDs against prevent defense in a loss) are definitely given way less (little to no) weight. And in tight games at the end, those stats or accomplishments are given heavier than average weight. At least from what I have read about the metric.





  7. #7
    Join Date
    Aug 2011
    Posts
    11,089
    Blog Entries
    7

    Re: ESPN's QBR with respect to Flacco

    Quote Originally Posted by Haloti92 View Post
    I don't think the fumbles and sacks are the issue. The issue is that it is a subjective accumulation of grades on every single play. It is not a "formula" that uses the game totals/stats as inputs.

    So for example, I suspect on Joe's 42-yard completion to Torrey Smith off of the play-action fake, Joe received 0 credit because the ball was underthrown (perhaps even received negative credit or was penalized, though I doubt it). Literally zero credit. And for Lee Evans twisting turning sideline catch, Joe probably received less credit than the play result indicated because Evans' got more credit than the WR normally does on a 'standard' completion of that length.

    So it really is a matter of a subjective and probably biased (I would say biased in favor of the 'name' QBs more than biased against any other specific QB) evaluator deciding how much of any given play's result is a result of the QB vs. the WR (or line, etc.). So a perfectly thrown ball gets more QB credit than a slightly off ball. A checkdown, that any NFL QB can make that goes for serious yardage likely is ignored by the ESPN grader (given no credit). A boneheaded INT is penalized more than an INT that is a result of the WR having the ball bounce off his hands. Sanchez's butt-fumble is penalized more than a jailbreak blindside sack-fumble, etc.

    It is a play-by-play evaluation. Overall game stats are not a factor, except insofar as they are "earned" by the QB, vs. given to the QB by circumstance or teammates' efforts, etc.
    Thanks.

    Given the accuracy component, my comment above regarding QBR rewarding dink and dunk offenses over deep ball offenses seems valid.

    It is a lot easier to be "accurate" on a 6 yard pass to Wes Welker than a 25 yard pass to Torrey Smith. Tom Brady comes out looking like a QBR god and Flacco a scrub. The game tells a different story, a story more accurately reflected by traditional QBR.

    If your stat is mean to be an accurate reflection, and, at least in this example clearly doesn't reflect reality, perhaps it is time to do some tweaks.

    Anyone else have any silly QBR outcomes? I remember seeing one from someone here where Tebow had some crazy QBR rating over some other QB that clearly had a better game.





  8. #8
    Join Date
    Sep 2011
    Posts
    5,069
    Blog Entries
    1

    Re: ESPN's QBR with respect to Flacco

    As it's ESPN, I'm pretty sure having the name "Manning" or "Brady" adds 20 points.

    In no small part due to Cameron's version of the Coryell, Flacco was never going to run up dream numbers with respect to completion percentage or sacks. The system intends to push the ball down the field into a higher-risk pass plays. Cameron's version increased the risk. That's why Flacco's pass percentage hovered around 50% in years 4 and 5, while the team was among the league leaders in plays over 20 yards.

    I suspect we're going to see a better pass percentage out of Flacco with a full year of Caldwell calling the plays. He's willing, unlike Cameron, to work Flacco into a rhythm with shorter pass plays and more friendly pass routes for the QB, especially early. Even while they were getting thumped by the Broncos during the regular season game, I was much, much happier with the offensive game plan. The offense wasn't in tune with it yet, but the playcalling was a lot better.





  9. #9

    Re: ESPN's QBR with respect to Flacco

    The system definitely punishes those that run up the score early and coast to a smooth finish late. 4th quarter counts as much if not more then the first 3 combined.





  10. #10
    Join Date
    Aug 2007
    Location
    Houston, TX Y'all
    Posts
    34,414

    Re: ESPN's QBR with respect to Flacco

    Their formula is not a big secret.

    It's on their web site.





  11. #11

    Re: ESPN's QBR with respect to Flacco

    Please never mention QBR again. It's a joke.





  12. #12

    Re: ESPN's QBR with respect to Flacco

    Quote Originally Posted by alienrace View Post
    Please never mention QBR again. It's a joke.
    What he said. The example above is all the proof I need.





Bookmarks

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •  
Link To Mobile Site
var infolinks_pid = 3297965; var infolinks_wsid = 0; //—->