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

The pats keep Fin the Fins!

Discussion in 'Miami Dolphins Forum' started by Dorfdad, Jan 20, 2019.

Thread Status:
Not open for further replies.
  1. Fin D

    Fin D Sigh Club Member

    71,258
    41,515
    113
    Nov 27, 2007
    This feels, very "chicken or the egg" to me. A team's offense doesn't perform in a vacuum. They have to face a defense. There are factors on offense that will dictate how good that offense is, but there are factors on that defense that will also dictate how that offense runs and vice versa.

    I can just as easily and accurately take your data and say the team whose defense performed worse allowed the other teams offense to score more.

    (FTR, i don't subscribe to either philosophy of offense or defense winning championships.)
     
  2. Miamiforlife

    Miamiforlife Member

    58
    21
    8
    Jan 3, 2019
    I still would and could not root for new england.
     
  3. cbrad

    cbrad . Club Member

    6,720
    7,301
    113
    Dec 21, 2014
    No question there's an inter-dependency between offensive performance and defensive performance. Whatever any and all those inter-dependencies are, there's an average ability of offenses to score and an average ability of defenses to prevent scoring, both in the regular season and in the playoffs. The question is therefore whether, given any latent (no direct ability to measure) ability of offenses and defenses, does the ability to score (or prevent scoring) in the playoffs increase or decrease relative to the regular season?

    That's what those z-scores are measuring. It just turns out that offenses of the winning team become relatively speaking more potent in the playoffs and in the SB than defenses. Why is another story. I don't know, and that's where it would be interesting to see how the inter-dependencies change. But it's still statistically speaking true that ability to score and ability to prevent scoring increase in the playoffs and the SB for the winning team, but that the ability to score increases more.

    And note that you can't really argue for the SB where the discrepancy is largest that somehow the teams are more unmatched in strength than in regular season games. If anything they're more matched in strength (by measures in the regular season) than a randomly picked regular season game, yet you see the result.
     
  4. Fin D

    Fin D Sigh Club Member

    71,258
    41,515
    113
    Nov 27, 2007
    But see, when someone says defense (or offense) wins championships, they aren't saying the bold.

    They are saying the team with the better defense will win. Or to put it another way, the team whose defense is better at stopping the other team's offense will win. They say that, because most understand, by the nature of the game, the offense always has the advantage, for one simple reason.....they aren't the reactive unit. The better a defense can control a given offense, the more likely they are to win. The teams with defenses that can do that, tends to be in playoffs.

    Also, you shouldn't get caught up in comparing strengths, as the statically better unit may actually have a particular weakness that the other team's unit can exploit even though that unit may be overall worse.
     
  5. Dorfdad

    Dorfdad Well-Known Member

    3,015
    1,496
    113
    Dec 9, 2007
    Every play should be reviewable. So sick of the stupid only this or that you get three challenges per game let them be used for anything they think was unfair!!!
     
    Pauly likes this.
  6. cbrad

    cbrad . Club Member

    6,720
    7,301
    113
    Dec 21, 2014
    No, in a whole bunch of debates on this issue I was responding to people who were explicitly saying the most important measure of offense (or defense) is points scored/allowed. In other words they already put themselves into a position where their preferred measure of offensive ability or defensive ability was precisely what I used. All I did was show the statistics.
     
    Pauly likes this.
  7. Fin D

    Fin D Sigh Club Member

    71,258
    41,515
    113
    Nov 27, 2007
    WADR, I don't see what in this post is refuting what I said. Maybe I'm not following or something.
     
  8. cbrad

    cbrad . Club Member

    6,720
    7,301
    113
    Dec 21, 2014
    OK let's try to clarify. What you said.. which is maybe the way you're understanding it, is actually not testable because it requires the ability to measure offensive or defensive performance independent of any and all inter-dependencies between the two. As evidence of that, note that you could equally argue whichever unit was better at stopping the other was "more important". Well.. you can't test that directly.

    So if that kind of argument was being made I would have just pointed out it's not testable.

    But that's not the origin of these arguments about "defense wins championships". First, over time (years really) it's been beaten out (lol) of posters that one should use points scored/allowed and not yards gained/lost as the preferred measure of how good an offense or defense is. Then I pointed out both in Club and here that in the regular season the best fitting plane between the z-scores of offensive points scored and allowed to win% shows that offense is about 5% more important in the regular season for winning (increasing points scored by one point increases win% about 5% more than decreasing points allowed by one point).

    It's after that that the claim was "offense helps win in the regular season, but defense wins championships" or something like that. In other words, in these debates the measure used was not in question. The question was what happens IN the playoffs, not how playoff teams for example did in the regular season. And I didn't have those stats for a long time because that requires game records, not just season records. Once I got that data however, I showed "defense wins championships" is a myth, given of course the assumption the preferred measure is points scored/allowed. At least in Club, after those posts the issue hasn't resurfaced. Here it has a few times.

    Context is more clear now?
     
    Pauly likes this.
  9. danmarino

    danmarino Season Ticket Holder Club Member

    8,952
    9,559
    113
    Sep 4, 2014

    I still think that when measuring the effectiveness of a defense or offense the best method is points for/points against.

    Awhile back I did a comparison of each team in the (at the time) 7 Super Bowls that the Pats had been in since 2001.

    I showed that the Pats offenses (I think) in all but 1 Super Bowl underperformed in regards to their average season ppg. I also showed that their defenses overwhelmingly held the opponent to the same or less than their average points allowed for the season. Basically, the Pats offenses wasn't nearly as effective in their Super Bowls, but their defenses were the same or better.

    Because of this I feel that defense is more important. Now, I didn't go through every Super Bowl so maybe it's just the Pats that have this disparity, but I would assume I'd find the same types of thing in most SB's.
     
    Surfs Up 99 likes this.
  10. cbrad

    cbrad . Club Member

    6,720
    7,301
    113
    Dec 21, 2014
    Yeah, so if there's no issue with the measure used, then you just need to look at all SB winners, and all playoff game winners:
    [​IMG]
    [​IMG]

    Those don't include this year's playoffs yet, but you can see that on average offense is more important than defense in winning a playoff game, and especially in winning the SB.
     
    Surfs Up 99 likes this.
  11. danmarino

    danmarino Season Ticket Holder Club Member

    8,952
    9,559
    113
    Sep 4, 2014


    Anyway you can see how each respective offense and defense did compared to their season averages in regards to PPG and PA?
     
  12. cbrad

    cbrad . Club Member

    6,720
    7,301
    113
    Dec 21, 2014
    No that I haven't done. It's a good question. I'll have to get back to you on that in the near future (have to write a new program first).
     
    Surfs Up 99 and danmarino like this.
  13. M1NDCRlME

    M1NDCRlME Fear The Spear

    507
    283
    63
    Oct 26, 2009
    Orlando
    As much as I hate the outcome of the AFC game, it was a joy to listen to Romo doing color commentary. He was calling the plays before they were happening and doing a great job of explaining. You could tell he really put the time in to study game film on both teams. He might make a great coordinator if he ever wanted to get into the coaching side.
     
    Tin Indian, Pauly, Dorfdad and 2 others like this.
  14. Fin D

    Fin D Sigh Club Member

    71,258
    41,515
    113
    Nov 27, 2007
    Yes, I believe the idea that defense (or offense) wins championships isn't really testable because of how intertwined offense and defenses are.

    I guess what I'm not understanding is how a person defines defense performance (points or yards) changes my argument that it's not really testable.
     
  15. cbrad

    cbrad . Club Member

    6,720
    7,301
    113
    Dec 21, 2014
    It's testable as soon as you agree on a measure. So whether you choose points or yards it's testable. It's not testable if all you say is one unit was "better" at stopping the other because that refers to some kind of interdependency.
     
    Pauly likes this.
  16. Fin D

    Fin D Sigh Club Member

    71,258
    41,515
    113
    Nov 27, 2007
    But its not accurately testable because the ability fo the offense to score points or rack up yards is still dependent on the defense allowing that to happen, and vice versa.

    No yard or point is gained without a push and pull between offense and defense.
     
  17. Rock Sexton

    Rock Sexton Anti-Homer

    2,680
    1,883
    113
    Mar 14, 2015
    Strahan and Osi were DE's. Donald and Suh are DT's. It's an entirely different kind of pass rush as they disrupt O-lines straight up the middle. QB's literally can't step up in the pocket - so the ones that lack lateral mobility get eaten alive. We're talking about a Rams pass rush that led the league in sacks without a viable end rusher for the vast majority of the season.
     
  18. cbrad

    cbrad . Club Member

    6,720
    7,301
    113
    Dec 21, 2014
    Nothing in statistics is perfectly "accurate" because ALL statistics throws away information. Of course you wouldn't have much of the technological advances modern society provides if we decided not to use a stat because it wasn't perfectly accurate.

    The better question is how would you measure it better, not whether there's some technical reason why it's not perfectly accurate.
     
    Pauly and Surfs Up 99 like this.
  19. Fin D

    Fin D Sigh Club Member

    71,258
    41,515
    113
    Nov 27, 2007
    This to me isn't about being perfectly accurate or not. It's about being accurate at all.

    This to me is about people saying X is true or false as a certainty, when its not even close to a certainty.

    I have always argued that there's no real hard and fast rules/benchmarks/etc. in football. I have also always argued that neither stats nor film tell the whole story of what happened on the field.

    If you wanted me to accept your stance more, then come up with a way to weight the impact of a given defense on a given offense (and vice versa) and explain the reasoning. Then maybe I can at least see that there's an attempt to account for the interplay.
     
  20. cbrad

    cbrad . Club Member

    6,720
    7,301
    113
    Dec 21, 2014
    Points scored and points allowed are some of the BEST stats out there for offense and defense because they're more highly correlated with win% than practically any other offense-only or defense-only stat. The only other ones comparable are variations on points scored/allowed like points per drive etc.. You also have no "team vs. individual" issues as you might with passer rating.

    As far as accounting for the effect of one on the other, I've also done that by looking at points scored as a function of points allowed and vice versa. You don't see the best fitting line to either of those graphs "fit" better than the other one which suggests that the effect of one on the other is comparable (i.e., treating one as the independent variable and the other as dependent isn't affecting quality of fit). The magnitude of the influence is captured by the correlation across NFL history which is -0.3293, or the square of that showing that about 10.85% of points scored is due to points allowed and vice versa.

    Really, among offense only or defense only stats you won't find better ones.
     
    Pauly and Surfs Up 99 like this.
  21. Fin D

    Fin D Sigh Club Member

    71,258
    41,515
    113
    Nov 27, 2007
    Ok. Let me try it from this angle....

    You win a football game by scoring more points than the other team. So obviously, point differential correlates with winning. I mean WADR, that's like saying water is wet.

    What happens a lot here, is people are trying to figure out why X happened. Stats mostly tells us either that X happened or the frequency with which X happens.
     
  22. cbrad

    cbrad . Club Member

    6,720
    7,301
    113
    Dec 21, 2014
    Yeah no question, but to measure how good an offense or defense is points scored/allowed is arguably the best, primarily because of its high correlation to win%. I already acknowledged earlier it doesn't tell us why.
     
    Last edited: Jan 21, 2019
  23. djphinfan

    djphinfan Season Ticket Holder Club Member

    92,641
    45,094
    113
    Dec 20, 2007
    I finally thought there was discontent as well this past offseason, not any eroding of Brady’s skills talk but locker room discontent and resentment from Brady..I’ve learned my lesson..
     
    danmarino likes this.
  24. cbrad

    cbrad . Club Member

    6,720
    7,301
    113
    Dec 21, 2014
    Looks like NFL might decide to allow reviews of PI after that blatant non-call. It's just a rumor but still..

    NFL rumors: League to discuss making pass interference reviewable "at length":
    https://bleacherreport.com/articles...om&utm_medium=referral&utm_campaign=editorial
    About time.. the wrong team is in the SB.
     
    Surfs Up 99 likes this.
  25. MonstBlitz

    MonstBlitz Nobody's Fart Catcher

    21,094
    9,836
    113
    Jan 14, 2008
    Alexandria, VA
    This weeks games prove just how unwatchable the NFL has become. Skipping the Superbowl. We all know how that story ends. "Tommy Terrific" gets the rigged story book ending to his cheating career and walks off into the sunset and we all get to hear how great the Patriots are/were for the rest of our lives. I can't do the NFL anymore.
     
  26. Dorfdad

    Dorfdad Well-Known Member

    3,015
    1,496
    113
    Dec 9, 2007
    Why can’t the coaches review any 3 plays per half?? They get three!!
     
    cbrad likes this.
  27. texanphinatic

    texanphinatic Senior Member

    9,263
    2,878
    113
    Nov 26, 2007
    Austin TX
    I don't even get that. Like, why limit them at all? Why even bother to let them have it? Just have a replay official do it all.

    ****ing seriously. Having the 4 best teams, all with exciting players everywhere, and the NFL couldn't stop shooting themselves in the dick for one goddam game. Blatantly bad calls still happen all the ****ing time despite the time atrocity that replay has become. It hasn't really solved anything, and unless they can tamp it down, are fixing to make it worse.
     
    RevRick, danmarino and cbrad like this.
  28. djphinfan

    djphinfan Season Ticket Holder Club Member

    92,641
    45,094
    113
    Dec 20, 2007
    Exactly, just have a full time review officiall that’s there to look at and correct the mistakes in the moment, they call down to the head official to pause and correct..
     
    danmarino, Brasfin and cbrad like this.
  29. Pauly

    Pauly Season Ticket Holder

    2,924
    2,712
    113
    Nov 29, 2007
    I did some research on this last off-season. Albeit in a limited and fairly preliminary way.

    Defensive performance seems to have zero correlation to offensive performance,
    Good offensive performance seems to have zero correlation to defensive performance.

    The only correlation I could find was that teams with very poor offenses tended to have poorer than average defenses.

    I was only looking at teams from the salary cap era.
     
  30. 2socks

    2socks Rebuilding Since 1973

    8,172
    2,108
    113
    Nov 27, 2008
    Atlanta
    Not sure what makes anyone.think Flores is the right pick. This is his first yr as DC. He has been on a short leash. He has coached special teams 08 -12 safeties 16, linebackers 17, DC 18.
    I will leave it at that - another stop gap of mediocrity because Tannenbaum and Grier can control him.
    Me - I would go get Wade Phillips, Doug Marone or Greg William's - maybe all three
     
  31. The_Dark_Knight

    The_Dark_Knight Defender of the Truth

    8,181
    6,253
    113
    Nov 24, 2007
    Melbourne, FL
    I'm not sure gauging a defense by points allowed is the best measure. For a while, our defense was one of those "bend but don't break" defenses and we were losing games often. We were giving up a TON of yardage and giving opposing offenses more opportuities to score, even if it was just field goals.

    Defensively, I'm almost at the point of gauging yards allowed as the more accurate measure of a defense's strength. The fewer yards allowed, the more likelihood of opposing offenses going 3 and out.
     
  32. Brasfin

    Brasfin Well-Known Member

    2,526
    1,716
    113
    Apr 27, 2013
    Brazil
    Exactly, they have to change up the reviews somehow.

    If they leave it the way it is and add the ability to challenge penalties games would drag on for nearly another hour, we'd have 4+ hour games on a regular basis.
     
  33. Brasfin

    Brasfin Well-Known Member

    2,526
    1,716
    113
    Apr 27, 2013
    Brazil
    That's the thing, though, the Pats didn't exactly "pour resources" into their defense. By all accounts this Patriots defense is very much average, talent-wise.

    It's their system that keeps churning out great results year in and year out both offensively and defensively, regardless of the players, and THAT'S what he have to hope Flores brings to Miami.
     
    danmarino, Nappy Roots and Tin Indian like this.
  34. Fin D

    Fin D Sigh Club Member

    71,258
    41,515
    113
    Nov 27, 2007
    On the surface, logically, that seems literally impossible.
     
  35. Pauly

    Pauly Season Ticket Holder

    2,924
    2,712
    113
    Nov 29, 2007
    Why?
     
  36. Dorfdad

    Dorfdad Well-Known Member

    3,015
    1,496
    113
    Dec 9, 2007
    I think they are looking st pedigree and hoping that someone well rounded can grow into the position. I mean if you look historically Belicheck has been the best coach in the NFL so in their mind so many years under his tutelage should translate to a well rounded coach.

    The Head coach while important isn’t the same as it used to be. Today you need a good surrounding cast of coaches.

    Let’s just hope the players by in
     
  37. Fin D

    Fin D Sigh Club Member

    71,258
    41,515
    113
    Nov 27, 2007
    How can a defense have no bearing on the offense?
     
  38. cbrad

    cbrad . Club Member

    6,720
    7,301
    113
    Dec 21, 2014
    It depends on the year you look at. Across NFL history the average correlation is -0.3293, so about 10.85% of variation in offensive performance can on average be explained by the defense and vice versa. However, this varies tremendously by year.

    For example, in 2018 that correlation was -0.0937 or essentially nothing while in 2017 it was tremendous at -0.6161. Here's a graph of the correlation between points scored and points allowed in the SB era:
    [​IMG]
     
  39. cbrad

    cbrad . Club Member

    6,720
    7,301
    113
    Dec 21, 2014
    Ignoring stats for a moment, the reason points is more important is because whoever scores more points than the other team wins the game, so whatever you do you want it to reflect in points scored or points allowed.

    Statistically you look at correlations to win%. For points allowed it's on average at -0.7342 throughout NFL history while for yards allowed it's -0.4910. The square of those numbers is what you want to compare because that tells you how much of variation in win% can be explained by either stat. For points allowed that's 53.9% while for yards allowed it's 24.11%.

    And if you're wondering how points allowed can explain more than 50% of win% when points scored is even more important, that's because some (~11%) of points allowed is due to points scored. Either way, w.r.t. what matters (win%) points allowed is about twice as important as yards allowed, but ~25% is still a lot so looking at total yards is not bad, it's just dwarfed by looking at points.
     
    danmarino likes this.
  40. Surfs Up 99

    Surfs Up 99 Well-Known Member

    1,288
    1,196
    113
    May 5, 2016

    Please excuse my stupidity, but can you explain what this means? Is it saying that the more points an offense scores, the fewer points their defense allows? Sorry! I just can't get the gray matter functioning today. :-(

    I think I understand what Pauly is saying where teams with very poor offenses also have very poor defenses. Heck, didn't we see that this year? Not defending Burke, but when your defense is on the field a ton because your offense is a league leader in 3 and outs, guys are going to get tired physically and mentally.
     
    danmarino likes this.
Thread Status:
Not open for further replies.

Share This Page