1. This site uses cookies. By continuing to use this site, you are agreeing to our use of cookies. Learn More.

Stats! Qbr 2014

Discussion in 'Miami Dolphins Forum' started by Galant, Mar 31, 2015.

  1. Galant

    Galant Love - Unity - Sacrifice - Eternity

    19,127
    11,058
    113
    Apr 22, 2014
    I agree as far as a common sense approach to 'pressure' is taken. That is, hits and sacks are pressure - pressure that just hit the QB in the gut.

    But if someone wants to tell me that an OL which couldn't handle the defence and let the QB get hit is better than an OL which struggled but kept the defence from hitting the QB, I'm going to have to disagree completely. I think that's a case of losing the forest for the trees.
     
  2. cbrad

    cbrad .

    10,659
    12,657
    113
    Dec 21, 2014
    I was responding to your general claim about stats>eyes in that post. Regarding this specific discussion, it looks like people on both sides are using stats, but no stats are conclusive. And the kind of analysis in that video doesn't tell you how much blame there should be on the OL vs. QB etc..
     
  3. dWreck

    dWreck formerly dcaf

    5,200
    2,975
    113
    Oct 23, 2011
    Sebring, FL
    Ryan Tannehill was the #3 sacked QB in the 2014 NFL per ESPN as well as #1 in 2013. Not sure where you're coming up with 10.

    2014:

    [TABLE="class: tablehead"]
    [TR="class: oddrow player-28-16724"]
    [TD="align: left"][TABLE="class: tablehead"]
    [TR="class: oddrow player-28-16724"]
    [TD="align: left"]1[/TD]
    [TD="align: left"]Blake Bortles, QB[/TD]
    [TD="align: left"]JAX[/TD]
    [TD]280[/TD]
    [TD]475[/TD]
    [TD]58.9[/TD]
    [TD]2,908[/TD]
    [TD]6.12[/TD]
    [TD]63[/TD]
    [TD]11[/TD]
    [TD]17[/TD]
    [TD="class: sortcell"]55[/TD]
    [TD]69.5[/TD]
    [TD]208[/TD]
    [/TR]
    [TR="class: evenrow player-28-14001"]
    [TD="align: left"]2[/TD]
    [TD="align: left"]Colin Kaepernick, QB[/TD]
    [TD="align: left"]SF[/TD]
    [TD]289[/TD]
    [TD]478[/TD]
    [TD]60.5[/TD]
    [TD]3,369[/TD]
    [TD]7.05[/TD]
    [TD]80[/TD]
    [TD]19[/TD]
    [TD]10[/TD]
    [TD="class: sortcell"]52[/TD]
    [TD]86.4[/TD]
    [TD]211[/TD]
    [/TR]
    [TR="class: oddrow player-28-14876"]
    [TD="align: left"]3[/TD]
    [TD="align: left"]Ryan Tannehill, QB[/TD]
    [TD="align: left"]MIA[/TD]
    [TD]392[/TD]
    [TD]590[/TD]
    [TD]66.4[/TD]
    [TD]4,045[/TD]
    [TD]6.86[/TD]
    [TD]50[/TD]
    [TD]27[/TD]
    [TD]12[/TD]
    [TD="class: sortcell"]46
    [/TD]
    [TD][/TD]
    [/TR]
    [/TABLE]
    [/TD]
    [TD="align: left"][/TD]
    [TD][/TD]
    [TD][/TD]
    [TD][/TD]
    [TD][/TD]
    [TD][/TD]
    [TD][/TD]
    [TD][/TD]
    [TD][/TD]
    [TD="class: sortcell"][/TD]
    [TD][/TD]
    [TD][/TD]
    [/TR]
    [TR="class: evenrow player-28-14001"]
    [TD="align: left"][/TD]
    [TD="align: left"][/TD]
    [TD="align: left"][/TD]
    [TD][/TD]
    [TD][/TD]
    [TD][/TD]
    [/TR]
    [TR="class: oddrow player-28-14876"]
    [TD="align: left"][/TD]
    [TD="align: left"][/TD]
    [TD="align: left"][/TD]
    [TD="class: sortcell"][/TD]
    [TD][/TD]
    [/TR]
    [/TABLE]

    2013:

    [TABLE="class: tablehead"]
    [TR="class: oddrow player-28-14876"]
    [TD="align: left"]1[/TD]
    [TD="align: left"]Ryan Tannehill, QB[/TD]
    [TD="align: left"]MIA[/TD]
    [TD]355[/TD]
    [TD]588[/TD]
    [TD]60.4[/TD]
    [TD]3,913[/TD]
    [TD]6.66[/TD]
    [TD]67[/TD]
    [TD]24[/TD]
    [TD]17[/TD]
    [TD="class: sortcell"]58[/TD]
    [TD]81.7[/TD]
    [TD]245[/TD]
    [/TR]
    [TR="class: evenrow player-28-11252"]
    [TD="align: left"]2[/TD]
    [TD="align: left"]Joe Flacco, QB[/TD]
    [TD="align: left"]BAL[/TD]
    [TD]362[/TD]
    [TD]614[/TD]
    [TD]59.0[/TD]
    [TD]3,912[/TD]
    [TD]6.37[/TD]
    [TD]74[/TD]
    [TD]19[/TD]
    [TD]22[/TD]
    [TD="class: sortcell"]48[/TD]
    [TD]73.1[/TD]
    [TD]245[/TD]
    [/TR]
    [TR="class: oddrow player-28-11237"]
    [TD="align: left"]3[/TD]
    [TD="align: left"]Matt Ryan, QB[/TD]
    [TD="align: left"]ATL[/TD]
    [TD]439[/TD]
    [TD]651[/TD]
    [TD]67.4[/TD]
    [TD]4,515[/TD]
    [TD]6.94[/TD]
    [TD]81[/TD]
    [TD]26[/TD]
    [TD]17[/TD]
    [TD="class: sortcell"]44[/TD]
    [TD]89.6[/TD]
    [TD]282[/TD]
    [/TR]
    [/TABLE]

    Which now makes him:

    #128 on the all time sack list for QBs since 1969
    #11 for most sacks in a single season all time.

    Meh, not that much worse than average. :rolleyes:


    Btw, NFL.com has the "dolphins" at #10 as a TEAM. How many sacks a TEAM gets OVERALL and how many sacks a single QB takes are not the same thing. He is THE MOST sacked qb in the last 2 years. that is FACT. Sad you are stretching it that way.
     
  4. Tannephins

    Tannephins Banned

    1,818
    572
    0
    Dec 23, 2014
    Even if we were to go with the measure you prefer (sacks plus QB hits), it's still the case that the correlation between QB rating and the percentage of pass dropbacks in which sacks and/or QB hits occurred -- in the table you posted later in the thread -- is -0.37.

    So, there is still 86.3% of the variance in QB rating that isn't associated with sacks plus QB hits.

    It's also the case that Tannehill's percentage of sacks plus QB hits (22.64%) wasn't even significantly higher than the league norm of 18.77%.
     
  5. Tannephins

    Tannephins Banned

    1,818
    572
    0
    Dec 23, 2014
    What the video shows is that when people typically attribute a sack to the offensive line, there may be more to the story that isn't visible on TV.

    And when the "subjective analysis" you're talking about as being in "competition" with statistical analysis doesn't incorporate what would be needed to rule out those "visual confounds," the stats are going to prevail the vast majority of the time in my opinion.
     
  6. Tannephins

    Tannephins Banned

    1,818
    572
    0
    Dec 23, 2014
    Click on this link, scroll down to "Passing Offense," and click on the "Sk" column header:

    http://www.pro-football-reference.com/years/2014/
     
  7. dWreck

    dWreck formerly dcaf

    5,200
    2,975
    113
    Oct 23, 2011
    Sebring, FL
    Teams overall sacks =/= single QB sacks. That's cool that Jacksonville somehow racked up over 70 as a team, but it doesn't matter. He is number the 1 sacked QB over the last 2 years. Fact.
     
  8. Tannephins

    Tannephins Banned

    1,818
    572
    0
    Dec 23, 2014
    So I suppose then that if you want to distinguish team sacks along the lines of the status of their QBs (starters or backups), then by definition you're implying that sacks are attributable to QBs to at least some degree?

    What does that do to your "fact"?
     
  9. cbrad

    cbrad .

    10,659
    12,657
    113
    Dec 21, 2014
    Yeah, but what I'm saying is that in this thread, both sides are using stats, and it's not clear based on the stats alone what interpretation is correct.
     
  10. dWreck

    dWreck formerly dcaf

    5,200
    2,975
    113
    Oct 23, 2011
    Sebring, FL
    Ofcourse it does. Who is saying the QB doesn't have some degree of responsibility? However the O-line plays a MUCH bigger part in the sacks than the QB does, anyone who says the QB is more responsible for sacks than the o-line is, is foolish, or has never played a single snap in a real game.... Let alone as an O-line or a QB. I'm not saying the QB doesn't play any part at all, i'm saying that:

    Is EXTREMELY misleading. You used a TEAM stat and said a PLAYER wasn't sacked as much as the LEAGUE AVERAGE. That is WRONG and you know it. Regardless of the cause, he is THE most sacked QB over the last 2 years.

    Not #10,
    Not Average
    the MOST.
     
  11. Tannephins

    Tannephins Banned

    1,818
    572
    0
    Dec 23, 2014
    Both of them involve a statistic that 1) isn't strongly correlated with QB performance statistics, and 2) doesn't distinguish Tannehill from the league norm.

    Does it matter which one is correct?
     
  12. resnor

    resnor Derp Sherpa

    16,327
    9,874
    113
    Nov 25, 2007
    New Hampshire
    No, it simply means that Ryan Tannehill was sacked more than two QBs who split games for the same team. It really changes nothing. Ryan Tannehill was sacked the most, out of any QB, the last two years. If I recall correctly, didn't Jax split time between two QBs? Like, one started some games, before the other guy came in? It doesn't say anything, necessarily, about the QB.

    And no one is arguing that sacks are sometimes attributable to the QB. I caution, however, when considering the Dolphins, it wasn't like Tannehill just couldn't avoid pressure coming from the left, when he could have escaped to the right, or up the middle, but instead just allowed himself to get wrapped up. Teams were rushing 4 guys, and getting pressure off both edges AND up the middle. He literally had nowhere to go most of the time, and nowhere to throw, when there were 6 or 7 guys back in coverage. The type of pressure matters. It matters if you are for instance, the Dolphins, generating pressure mostly on the right side of the oline with Cam Wake, or you're a team that is able to get pressure off both edges and up the middle.
     
    dWreck likes this.
  13. Tannephins

    Tannephins Banned

    1,818
    572
    0
    Dec 23, 2014
    And what does that mean, when his performance improved so much from one of those years to the next?
     
  14. resnor

    resnor Derp Sherpa

    16,327
    9,874
    113
    Nov 25, 2007
    New Hampshire
    It. Doesn't. Matter. How. Far. From. The. League. Norm. It. Is.

    If a team wants to be good, then they are trying to be BETTER than the league norm. These statistics are far from the league best...that is what people are saying.
     
  15. resnor

    resnor Derp Sherpa

    16,327
    9,874
    113
    Nov 25, 2007
    New Hampshire
    It's meaningless in the context that you're talking, because you have no way of knowing what his improvement would have looked like behind a good oline.
     
  16. dWreck

    dWreck formerly dcaf

    5,200
    2,975
    113
    Oct 23, 2011
    Sebring, FL
    That he's a strong kid and still progressing very favorably, despite an absolute abortion of an o-line in '13 and a patchwork musical chair o-line in '14? You can't really say one way or the other.
     
    resnor likes this.
  17. cbrad

    cbrad .

    10,659
    12,657
    113
    Dec 21, 2014
    Well.. if the stats aren't telling you how much of the blame goes to OL vs. QB, then you have to use something other than stats to argue it, which is what the guys you're arguing with are doing. And while sacks on its own may not influence QB performance much, you yourself said that QB pressures do, so yes I guess it does matter.
     
  18. Tannephins

    Tannephins Banned

    1,818
    572
    0
    Dec 23, 2014
    Not when Tannehill hasn't been pressured any more often than the average QB in the past two years.
     
  19. resnor

    resnor Derp Sherpa

    16,327
    9,874
    113
    Nov 25, 2007
    New Hampshire
    What???
     
  20. cbrad

    cbrad .

    10,659
    12,657
    113
    Dec 21, 2014
    Are we back to that again? We're talking absolute influence, not relative. Is QB pressure highly influential on QB performance, yes or no? If yes, then the question becomes how much of the blame should be assigned to OL vs. QB or something else. None of the stats presented really answers that question, so you have to go by general knowledge of the game. And as others have pointed out before, the purpose of the OL is to keep pressure off the QB, so that's really the first place to look.
     
  21. Tannephins

    Tannephins Banned

    1,818
    572
    0
    Dec 23, 2014
    And the best we're going to do there is the percentage of pass dropbacks in which QBs are pressured. Or do you see a better method of extracting that being offered?
     
  22. Tannephins

    Tannephins Banned

    1,818
    572
    0
    Dec 23, 2014
    You're not incoroporating the number of pass dropbacks the QBs in the table took. I covered that in post #126.
     
  23. Fin D

    Fin D Sigh

    72,252
    43,684
    113
    Nov 27, 2007
    WADR, you made up correlations in that post that aren't really indicative of anything.

    Tannehill was hit or sacked to the tune of near 25% of his attemtps. 1 in 4 attempts.
     
    resnor likes this.
  24. cbrad

    cbrad .

    10,659
    12,657
    113
    Dec 21, 2014
    Yeah, as you might have noticed, I've stayed away from that particular discussion because I'm not sure what the best method is. This last comment of yours is far afield from the post about stats vs. eyeballs that I responded to.
     
  25. Tannephins

    Tannephins Banned

    1,818
    572
    0
    Dec 23, 2014
    I'm done folks. Thanks for the discussion. :up:
     
  26. Galant

    Galant Love - Unity - Sacrifice - Eternity

    19,127
    11,058
    113
    Apr 22, 2014
    1 - The variance stat is practically irrelevant. First because we're looking at the effect of the OL on the QB, so we assume all else being equal. Second because it's impossible to know how OL play indirectly affects everything else. Although actions are divided up in order to measure statistics they aren't divided in reality. They're connected, and I'd say the offensive line is one area where this is most true because the work of the OL affects every single offensive play even though that affect will rarely show up on the stat sheet. That needs to be remembered. Additionally, the variance stat muddies the water, so to speak, because it's entirely relative between different teams and over only one year. It gives us no absolute value against which to measure anything and also ignores the fact that the dynamic between players on any given team and system can and will vary greatly between teams.

    2 - 22.6% VS. 18.8% tells us very little. As Fin D pointed out, we're interested in growth within a single team and not imagining how T17 would fare on any other team. This is where statistics have to be used carefully. We're not dealing with coin tosses and dice rolls. We're dealing with human actions, as part of a team, and within an arena of physical contest. That means that there will never be a magic number to explain or predict anything and everything. They can simply be used as markers and indicators of different factors which, when used in context and comparison, can give us some information. But the football field isn't a laboratory nor is it even 'level' between teams. All that considered, saying that the difference between 18% and 22% is very small is impossible without a field of reference. 4% could be a tiny difference or a colossal one depending upon the context. I don't see how you can cite 18.77% as a norm without reference to several years' data. Additionally, I see that Fin D pointed out that 22% is almost a quarter of plays. All of a sudden that sounds rather more significant.

    Honestly, I'm just not quite sure what the point is you're trying to make, unless you're trying to say that the OL is utterly irrelevant to QB play, but there's no way that that makes any sense in the real world. If it were true there wouldn't be any need for an OL and you don't believe that because you have said that pressures are significant. Which of course leads me to wonder how you can feel that pressures and hurries aren't significant but hits and sacks aren't therefore more significant.
     
  27. Galant

    Galant Love - Unity - Sacrifice - Eternity

    19,127
    11,058
    113
    Apr 22, 2014
    Ah. Okay. Thanks to you too.
     
    Tannephins likes this.
  28. cbrad

    cbrad .

    10,659
    12,657
    113
    Dec 21, 2014
    He's getting that from his stats because while a hit/sack influences QB performance far more than a pressure/hurry on a play-by-play basis, hits/sacks occur far less often than pressures/hurries so the "percent of variance explained" can be less.
     
  29. emocomputerjock

    emocomputerjock Senior Member

    5,649
    1,853
    113
    Nov 23, 2007
    DC
    The issue with trying to measure OL influence on the passing game is that there's no meaningful way to extract OL performance from passing stats. Until someone does that, it's not possible to try to correlate QB performance with OL performance.
     
  30. resnor

    resnor Derp Sherpa

    16,327
    9,874
    113
    Nov 25, 2007
    New Hampshire
    Mediocre, or worse, play from others in general
     
  31. cbrad

    cbrad .

    10,659
    12,657
    113
    Dec 21, 2014
    No more excuses this year though.
     
  32. resnor

    resnor Derp Sherpa

    16,327
    9,874
    113
    Nov 25, 2007
    New Hampshire
    Well, again, though, I'll find it hard to blame Tannehill if they miss the playoffs because of other parts of the team sucking. I want to see Tannehill play well, though. I don't think there would be too many excuses for him to play poorly. Making the playoffs is a team achievement, though.
     
    77FinFan likes this.
  33. cbrad

    cbrad .

    10,659
    12,657
    113
    Dec 21, 2014
    Yeah, but at some point the excuses have to stop. Luck would've taken us to the playoffs 2 of the last 3 years I'm pretty sure. I guess I could make an exception if our top 5 players other than him all get season ending injuries, but other than something dramatic like that, no more excuses for me.
     
  34. resnor

    resnor Derp Sherpa

    16,327
    9,874
    113
    Nov 25, 2007
    New Hampshire
    So your opinion is that Luck would have magically kept our defense from giving 30+ so many times?
     
  35. cbrad

    cbrad .

    10,659
    12,657
    113
    Dec 21, 2014
    No, my opinion is that he would have helped our offense score more points, thus turning to wins some losses we had on our record.
     
    resnor and djphinfan like this.
  36. danmarino

    danmarino Tua is H1M! Club Member

    15,354
    20,976
    113
    Sep 4, 2014

    Playoffs are not an indication of QB play alone. A 92 QBR is good enough to win championships.
     
    resnor likes this.
  37. djphinfan

    djphinfan Season Ticket Holder Club Member

    111,648
    67,540
    113
    Dec 20, 2007
    when ryan starts to understand that he can do more things then try to complete a pass on third down then and only then will his game go to another level..right now he has no clue..right now he's a 92 rated passer, emphasis on the word ''Passer''
     
  38. Galant

    Galant Love - Unity - Sacrifice - Eternity

    19,127
    11,058
    113
    Apr 22, 2014
    How much of that do you think is Ryan and how much the coaches?
     
  39. Galant

    Galant Love - Unity - Sacrifice - Eternity

    19,127
    11,058
    113
    Apr 22, 2014
    Sort of. There'll always be the guess-work involved. However, if measures for OL performance can be found and applied and measures for QB performance can be found and applied, and if correlations are consistently found to occur between the two then it might well suggest a real-world correlation. We can certainly try; if for no other reason than it seems pure common sense that OL performance will have some effect on passing performance. I suspect, though, that it will have to be done per team/QB rather than across teams simply because different styles, tactics and abilities will all factor into that relationship. Probably no shortcuts on that one.

    I think part of the problem people have with stats is over expectations. There's an expectation that if only we can get enough 'data' we can break it all down by numbers and be able to predict everything perfectly with numbers. That's never going to the case though. Statistical data has to be applied using observation, reason and wisdom. Put simply, stats should be a tool to help one understand, not a truth which defines understanding.

    It makes sense to me that, as I said before, hits and sacks are a good starting place for OL performance and that an improving OL would go through a period where hits and sacks go down but hurries and pressures go up, then, ideally, they would start to settle down too. If that's the case, and if OL line affects a given QB's passing, then we'd expect a correlation between that progress in the OL and the QB passer rating. That's the theory anyway. Only way to find out is to watch and see, and refine that understanding as things progress. It'll never be easy though, what with the NFL being such a changeable environment. New players, new opponents, new tactics. Statistics are always more useful over a longer period, and in the NFL it's a franchise dream to have that sort of consistency. Hopefully we'll see that happens for the Dolphins as they find the right system, coaches, and team.
     
  40. Tannephins

    Tannephins Banned

    1,818
    572
    0
    Dec 23, 2014
    I'm re-entering the discussion to address only this and not Ryan Tannehill and the Dolphins specifically:

    And there may indeed be one, but its strength may pale in comparison to that of the correlation between QB performance and other variables.

    It's entirely possible, for example, that either: 1) the variation in offensive line play throughout the league is so small that it can't possibly cause the much larger variation we see in QB play, and/or 2) that QB play is so much more strongly related to other variables that the variation in line play represents a very small contribution to QB play by comparison.

    Either one (or both, combined) of these scenarios would render the correlation between line play and QB play smaller than would perhaps be expected via "common sense." In that case uncommon sense would be correct.

    Sometimes what's thought to represent "common sense" is wrong, and the consensus is incorrect. Using the defense that "it's common sense" therefore isn't enough.
     

Share This Page