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

QB or Chase Young?

Discussion in 'Miami Dolphins Forum' started by Galant, Oct 28, 2019.

  1. Pauly

    Pauly Season Ticket Holder

    3,696
    3,743
    113
    Nov 29, 2007
    I think that points scored by the leading team in “garbage time” are the ones that should be discounted, not the trailing teams’.
    There has been research done that shows that being able to score points while trailing in “garbage time” is something that is indicative of being able to score points in other situations.
    On the other hand trailing teams will take riskier options leading to more turnovers, leading to more point scoring opportunities from playing on shortened fields.

    Sustaining long drives in and of itself creates more turnover opportunities, as each play run has a roughly 1% of being a turnover. If your plan is to put together series of long dozen+ play drives, then you are going to give your opponent more opportunities to score TDs.
     
  2. cbrad

    cbrad .

    10,659
    12,657
    113
    Dec 21, 2014
    Which doesn't in any way prove "defense wins championships". Notice you didn't even mention or consider absolutely everything else that happened in that game, including all the offensive plays that proved crucial to the win.

    Humans simply are not good at figuring out how important different kinds of information are to the final outcome as evidenced by their inability to accurately estimate in-game win probability. And if you can't estimate how influential any particular play is on the final outcome, how is it possible for you to know how to weight all the different kinds of information (e.g., points allowed, turnovers at crucial moments, etc...) to arrive at a good measure of the defense? It's just not possible (or way too improbable). And indeed no one has done it.

    The only way to come up with a superior measure of the defense is to use a machine to test a whole range of possible set of weights on the different component measures and see which one actually works best. If it was important enough I'd do it, but getting 84% of the variability in win% by looking at points alone (and remember whatever measure works for defense works on the opposing offense, which is why you can look at correlations between point differential and win%) means you've already approximated to a good degree the final result.

    In other words, to 84% probability we already know that "defense wins championships" is on average false in the NFL. Hard to beat.
     
    The Guy likes this.
  3. AGuyNamedAlex

    AGuyNamedAlex Well-Known Member

    3,582
    2,579
    113
    Sep 12, 2015
    On the other hand if you're going 3 and out that doesnt help the defense much either.
     
  4. Unlucky 13

    Unlucky 13 Team Raheem Club Member

    51,930
    63,009
    113
    Apr 24, 2012
    Troy, Virginia
    It might have been Travis. He's a WSU guy, and liked him a lot.
     
  5. Unlucky 13

    Unlucky 13 Team Raheem Club Member

    51,930
    63,009
    113
    Apr 24, 2012
    Troy, Virginia
    I will certainly give you that the Dolphins inability to score points, or even gain yardage, in garbage time has been a frustration of mine for years. In so many other games around the league, you'll see a team with a big lead let up and let the opposition gain yardage in smaller chunks while they eat up the clock. But the Dolphins usually can't even do that, and are just as inept in the 4th quarter losing by 20+ as they are in the first. And that's been going on for 15 years, through various coaches.
     
    resnor likes this.
  6. The Guy

    The Guy Well-Known Member

    6,598
    3,323
    113
    Oct 1, 2018
    I would be interested to see whether that was the case during the one season in which they had very good quarterback play, when Chad Pennington finished second in the league in the MVP voting.
     
  7. Unlucky 13

    Unlucky 13 Team Raheem Club Member

    51,930
    63,009
    113
    Apr 24, 2012
    Troy, Virginia
    LOL, we'll have to disagree about 2008. I still feel that was a six-win quality team that caught every break possible against an easy schedule.
     
    resnor likes this.
  8. Fin D

    Fin D Sigh

    72,252
    43,684
    113
    Nov 27, 2007
    Its funny to me how people remember that season.

    They clearly forget that we had too dust off the Wildcat BECAUSE Penny couldn't generate any chunk yardage on offense.
     
    mlb1399, resnor, Unlucky 13 and 2 others like this.
  9. KeyFin

    KeyFin Well-Known Member

    10,488
    12,821
    113
    Nov 1, 2009
    Yeah, but the unveiling of the Wildcat against New England is one of my top Dolphins memories ever! Ronnie Brown was an absolute stud that afternoon and the Pats had no clue what the heck was happening.
     
    Hooligan, resnor and Tin Indian like this.
  10. Pauly

    Pauly Season Ticket Holder

    3,696
    3,743
    113
    Nov 29, 2007
    I suspect that a lot of the “missing” 16% of explaining W-L is in time management.

    Football isn’t won by the team that scores more. It’s won by the team that scores more after 60 minutes. At certain points in the game, like the 2 minute drill or a leading team running out the clock in the 4th quarter, it is very obvious to see how important time management is. BTW I’m not talking about time of possession, I’m talking about efficiency and making the clock slower or faster depending on what is more advantageous for you.
     
  11. Fin D

    Fin D Sigh

    72,252
    43,684
    113
    Nov 27, 2007
    By all means, enjoy it,

    I just wonder why everyone forgets it had happen in the first place.
     
    resnor, Unlucky 13 and KeyFin like this.
  12. cbrad

    cbrad .

    10,659
    12,657
    113
    Dec 21, 2014
    This is a good point. Pennington started playing well only after the Wildcat was unveiled. And you can see that he was much less effective the year before as well as the year afterwards when the Wildcat was no longer as unexpected.

    So Pennington ending up #2 in passer rating that year (which is really an accomplishment.. Marino ended up #2 or better only TWICE in his career) definitely has a lot to do with the Wildcat. On the other hand, Pennington did end up #1 in passer rating in 2002 so it's not completely unexpected for Pennington to do that well. He was definitely an above average QB that in some years was below average (his career z-score is 0.8643 which is good enough to build a team around if you want a SB win about once every 15 years).

    btw.. Marino ending up #2 or better in passer rating only twice (and #1 only once) just puts into perspective how impressive it was for Steve Young to end up #1 a total of 6 times!! (1991-1994 and 1996-1997). Absolutely unbelievable and a good reason why he's statistically the most efficient QB in the SB era.
     
    resnor likes this.
  13. The Guy

    The Guy Well-Known Member

    6,598
    3,323
    113
    Oct 1, 2018
    Regardless of the reasons for it, the team still had very good quarterback play in 2008, and so the question remains of whether the offense functioned differently that year, and whether that was caused by its better quarterback play.

    Whether Chad Pennington is an all-time great is beside the point.
     
  14. AGuyNamedAlex

    AGuyNamedAlex Well-Known Member

    3,582
    2,579
    113
    Sep 12, 2015
    The mind of an all time great inside the body of an average physical player would be my evaluation of Pennington.
     
    resnor and KeyFin like this.
  15. Tin Indian

    Tin Indian Rockin' The Bottom End Club Member

    7,929
    4,404
    113
    Feb 10, 2010
    Palm Bay Florida
    Thats true in a big way. They had nothing happening on offense and on a flight home the coaching staff started talking about it. It was sprung on the unsuspecting Pats in Foxborough the next weekend and they had no clue what was happening. Never seen a Belichek coached team so out of sorts. One of my favorite memories was the fans leaving the stadium early in the 4th quarter. That was awesome!

    But it all happened because they didn't have anything to loose. The offense wasn't working.
     
  16. Tin Indian

    Tin Indian Rockin' The Bottom End Club Member

    7,929
    4,404
    113
    Feb 10, 2010
    Palm Bay Florida
    Pennington is not even close to being an all time great. He was an ALL time smart qb but he didn't have a great arm to begin with and then he kept injuring his throwing shoulder. He completed so many passes on his brains alone.

    Another of my favorite things that year was knocking the Jets out if the playoffs with their discarded QB at the helm. Man that was sweet!
     
  17. Striking

    Striking Junior Member

    1,775
    646
    113
    Apr 21, 2008
    Aurora, Colorado
    I would have said Young before the, Rosen is, no he's not, whiplash decision-making.
     
  18. Patster1969

    Patster1969 Season Ticket Holder Club Member

    679
    788
    93
    Nov 8, 2017
    Definitely helped Young throwing to the greatest receiver in history plus the other weapons & the running game - not doubting that he was a great player but he did very little with the 'talent' he had at Tampa. Dan didn't always have the great receivers or ever had a running game, so the defense knew he was going to be throwing the ball - he also didn't have Young's running ability, so defenses had to account for that and couldn't just drop 7 into coverage every time
     
    resnor likes this.
  19. The_Dark_Knight

    The_Dark_Knight Defender of the Truth

    11,817
    10,321
    113
    Nov 24, 2007
    Rockledge, FL
    Sorry to disagree with you about defenses winning championships...

    http://www.nfl.com/superbowlchamps/defense

    And if that is not enough...



    As if the date of that article...

    42 of the winning Super Bowl teams had a top 10 defense.

    29 of those teams had a top 5 defense

    25 had a top 3 defense

    And 14 had the number one overall defense.

    Offensively in comparison, only 38 of those teams had a top 10 offense

    27 of those treats were rated in the top 5 offensively

    18 were rated top 3

    And 10 were rated number one.

    Defense does indeed win championships. Numbers don’t lie
     
    KeyFin, Hooligan, Fin-O and 2 others like this.
  20. djphinfan

    djphinfan Season Ticket Holder Club Member

    111,651
    67,546
    113
    Dec 20, 2007
    Travis from locked on dolphins..
     
  21. Galant

    Galant Love - Unity - Sacrifice - Eternity

    19,127
    11,058
    113
    Apr 22, 2014
    @Aqua4Ever04
     
  22. Two Tacos

    Two Tacos Season Ticket Holder Club Member

    11,121
    5,828
    113
    Nov 24, 2007
    Why not both? We have the ammo...
     
  23. cbrad

    cbrad .

    10,659
    12,657
    113
    Dec 21, 2014
    See.. at least this type of post I can respect even if the statistical analysis is wrong because you went out and actually decided to get data to try to argue a point. Why is it wrong? Because ranks aren't measures (in other words.. those numbers do "lie"). This is a crucial distinction that had to be made in a large number of fields, from medical research to educational testing, before the data could be trusted to truly represent disease severity or a student's ability based on testing.

    The reason ranks aren't measures is because the difference between two neighboring ranks isn't equal, which is a necessary condition for anything to be called a "measure". For example, the difference between rank #1 and rank #2 isn't a priori the same as the difference between any other consecutive ranks, such as rank #2 and rank #3. A true measuring instrument such as a ruler does not have this flaw: the actual physical distance between two neighboring tick marks (e.g., 1 and 2 vs. 6 and 7) is always identical.

    That's why you need to transform raw scores (or raw stats) into z-scores like I did before. z-scores put everything in standard deviation units and that equates scales across distributions. There may still be an issue of how skewed the distributions are, and it's important to show that the skew doesn't change that much (it doesn't for NFL points scored/allowed data), but whether it does or not z-scores give you an actual measurement scale while ranks don't.

    An example of where this is useful: the 1971 Cowboys and 1972 Dolphins both had the #1 rank offensively but those were not equally impressive offenses. The 1971 Cowboys were by FAR more impressive. The z-scores are 2.5578 for the Cowboys and 1.5551 for the Dolphins. And you can see that if you look at the raw data (even before it has been transformed into z-scores):
    https://www.pro-football-reference.com/years/1971/index.htm
    https://www.pro-football-reference.com/years/1972/index.htm

    Note that the Cowboys scored 406 points while the 2nd best team that year scored 344. In comparison the Dolphins scored 385 points while the 2nd best teams scored 367. The Cowboys were WAY better than everyone else while the Dolphins were marginally better. That distinction is completely lost when you use "ranks" but is preserved with z-scores.

    So the reason the graphs I posted in post #71 are the stats you want to go by is because it's the proper way to do the statistical analysis:
    https://www.thephins.com/threads/qb-or-chase-young.94713/page-2#post-3216872

    So once again.. to about 84% probability "defense wins championships" is a myth in the NFL.

    btw.. in order for the remaining 16% variance unexplained to lead to a change in this conclusion, something very very unlikely needs to happen: there needs to be a systematic bias using points scored/allowed AGAINST the defense. Why would that ever be the case?
     
    Pauly and The Guy like this.
  24. djphinfan

    djphinfan Season Ticket Holder Club Member

    111,651
    67,546
    113
    Dec 20, 2007
    Possibility
     
  25. The_Dark_Knight

    The_Dark_Knight Defender of the Truth

    11,817
    10,321
    113
    Nov 24, 2007
    Rockledge, FL
    Brad, no disrespect but I feel you over think things like this, with your transcribing into “z scores”. I tend to go with Sir Arthur Conan Doyle’s explanation...

    “When you eliminate the impossible, whatever remains however unlikely, must be the truth”

    There is one virtual common constant in Super Bowl champions and that’s a top tiered defense.
     
  26. AGuyNamedAlex

    AGuyNamedAlex Well-Known Member

    3,582
    2,579
    113
    Sep 12, 2015
    A teams overall ranking isnt always indicative of their play as the season wears on and closes.

    I agree that defense doesnt win championships on it's own. However, at some point in the playoffs you are going to have to play defense and make a key stop.

    You've already said that the second most influential statistic/measure of the game is pass defense. It seems contradictory to come out and say defense doesnt really matter.

    I will say that all time great defenses do tend to win championships. The Bears, Ravens, Buccaneers and Seahawks all hoisted a trophy.

    T
    You can say the Seahawks have Wilson but they dont win that title without their defense playing lights out all season.

    Unless I'm misunderstanding you.
     
  27. Sceeto

    Sceeto Well-Known Member

    13,501
    6,246
    113
    Oct 13, 2008
    New York
    F---!!!!!!!!!!!!!!!!!!!!!
     
  28. cbrad

    cbrad .

    10,659
    12,657
    113
    Dec 21, 2014
    z-scores = number of standard deviations above or below the mean is not only by far the most commonly used measure of how far away some data point is from average but also probably the absolute minimum you'd need to compute to even begin a discussion with a statistician about how far away some data point is from average. Not sure you could get away with less. The data you linked to simply doesn't even qualify for discussion (at least in science) because all it takes is for someone to point out ranks aren't measures and that ends the discussion.

    So this isn't me getting cute. This is the absolute minimum level of analysis that would even be acceptable.

    There's no logic to this line of reasoning. You're saying that because the average SB winner has a defense that's well above average (a better way of putting it) that therefore it's impossible for the average SB winner to have an offense that's even higher above average?

    Not only is it possible, that's what actually occurs.

    Yes you're misunderstanding. First, go back to post #71:
    https://www.thephins.com/threads/qb-or-chase-young.94713/page-2#post-3216872

    You see those two graphs at the top for the SB winner, one showing z-scores for offense and the other showing z-scores for defense? Each time the z-score is above zero that means that the unit (defense or offense) was above average. The average z-score on offense for the SB winner is 0.9389 which is top 18th percentile while the average z-score for the SB on defense is 0.4072 which is top 35th percentile. Note that both are well above average.

    So first thing to note is that your average SB winner has a well above average offense AND a well above average defense. However, on average at least the offense is better (regardless of what the "rank" is). That's all I'm pointing out.

    Now.. for any specific SB you can see in those graphs that sometimes the offense z-score is way higher while sometimes the defense z-score is way higher. So you will see many cases where it was the offense that was more important and also many cases where the defense was more important. It's just that on average the offense is more important.

    Finally, I pointed out that the 2nd most important statistic ON DEFENSE was passer rating allowed. That's ignoring stats on offense or those that combine both offense and defense like turnover differential. The correlation between passer rating allowed and win% across NFL history is -0.593 while it's +0.633 for offensive passer rating so whatever offensive passer rating measures is more important for winning than whatever defensive passer rating measures. So none of these z-scores for SB winners are inconsistent with that finding.
     
    AGuyNamedAlex, Pauly and The Guy like this.
  29. Unlucky 13

    Unlucky 13 Team Raheem Club Member

    51,930
    63,009
    113
    Apr 24, 2012
    Troy, Virginia
    The beauty of football - the reason I love football so much more than the other sports, is that no matter what one team does - no matter what twenty other teams do, there's always another way. Two excellent teams can be as different as night and day, and neither one is right and neither one is wrong.

    Thats why whenever someone suggests that a team "must" do this or that to be successful, it always rubs me the wrong way.
     
    Pauly likes this.
  30. cbrad

    cbrad .

    10,659
    12,657
    113
    Dec 21, 2014
    I think people just mean we "should" do X because (for example) most SB winners did X rather than we "must" do it. I mean, there's no denying that the 2000 Ravens won the SB with a historically great defense and a QB like Trent Dilfer whose stats were around average while going 7-1 in the regular season and 4-0 in the postseason lol. That actually happened so it's not like it's impossible to build a team like that and win the SB. But there's also no question it's historically more likely to win a SB with an above average QB!

    So maybe there's nothing to worry about here once it's understood we're talking about odds of success.

    As to the first paragraph, is there a sport where it's not true that champions have won using different styles and/or strategies? I think all the major team sports are similar to football in that there isn't just one style that has been successful. And some sports see far more variation in styles/strategies among champions than football IMO. Try MMA for example.
     
  31. Unlucky 13

    Unlucky 13 Team Raheem Club Member

    51,930
    63,009
    113
    Apr 24, 2012
    Troy, Virginia
    While I love baseball, MLB is becoming a league where nearly every team does things the exact same way, both on and off the field. Mechanical, by the book, what the computer says in every situation. Its made it a lot less fun to watch. I really don't want the NFL to go down that road.
     
  32. cbrad

    cbrad .

    10,659
    12,657
    113
    Dec 21, 2014
    I see.. yeah obviously analytics is much more difficult to develop and successfully implement in the NFL than in MLB because of all the interaction effects, but eventually I think analytics will prove to be successful in the NFL and alter it in a similar way. And you're right that might make it less interesting to watch.. or possibly more interesting if you get involved in the analytics!
     
  33. AGuyNamedAlex

    AGuyNamedAlex Well-Known Member

    3,582
    2,579
    113
    Sep 12, 2015
    Excuse my ignorance here, I'm not saying you are wrong in this scenario I'd just like something explained a bit after reading your reply.

    Z-Scores seem necessary for comparing teams that played in different seasons I give you that 100%

    Isnt there inherent bias towards offense in the data? Let me use an example.

    A team, let's say the Chargers, is ranked #1 on defense allowing 10 PPG. The same team is also scoring 30 PPG tanked #1 in the league.

    Team B, let's say the Lions, are ranked #2 on defense giving up 13 PPG and scoring 20 also ranked #2.

    The Z-Score would indicate that the offense of the Chargers was more impressive compared to its peers than the defense and conclude that the offense was more important correct?

    That said, you can argue that the Chargers offense could score let's say 21 PPG with similar results.

    In other words some of the points the offense scores are irrelevant but still count for them. Every point the defense allows is counted as a negative even if it has no impact.

    Z-Scores seem heavily skewed towards offense to me. Can you explain how it accounts for that?
     
  34. AGuyNamedAlex

    AGuyNamedAlex Well-Known Member

    3,582
    2,579
    113
    Sep 12, 2015
    It doesnt logically make sense that the difference between passer rating allowed and passer rating would be that big.

    If you are giving up less passer rating you should be winning that battle and more games as well.

    It almost leads you to believe every great defense has a crap QB.

    Which actually might not be far off lol
     
  35. cbrad

    cbrad .

    10,659
    12,657
    113
    Dec 21, 2014
    If you reduce average PPG then win% will suffer unless of course you just happen to somehow score fewer points in precisely the games you won, but still enough to win. Given that football isn't a game a team can easily control the outcome of, I think it's untenable to suggest that reducing the average won't have an effect on win%.

    And w.r.t. unnecessary points scored, I'd actually say the stats are more likely to be biased TOWARDS defense because offenses usually don't run up the score when trying to run out the clock. It's another reason that using PPG in this "defense wins championships" argument is very unlikely to be biased against the defense.. if anything I bet it's slightly biased towards it.

    Finally, since average PPG across NFL history is 20.8 and is closer to 22-23 in recent years, the Chargers' offense (in your example) with 30 PPG would likely not have as high a z-score as its defense with 10 PPG which is farther away from the mean. But of course z-score depends on the entire distribution of PPG so we can't know a priori until we calculate it.

    Since that first statement of yours could be read in different ways, let me make it clear: the sign (negative or positive) on the correlation should be ignored when asking the question of how important is that stat for winning. Actually you want to square those numbers to really compare in the right units: variance explained. So (-0.593)^2 = 35% variance explained (of win%) by passer rating allowed and 0.633^2 = 40% variance explained by passer rating. So it's about a 5 percentage points difference in variance explained.

    Why is there a difference? Because in one case you keep the QB constant (on offense) while in the other case you are playing against all kinds of different QB's. So it actually makes sense that the metric that keeps the most influential player on the metric constant will show higher correlation to win%. Not really a surprise IMO.

    As far as how much having a good pass defense means you'll have a good passing offense (proxy for QB for now), the correlation between the two is -0.2 so yes there's some effect but (-0.2^2) = 4% of variance explained so 96% of what constitutes a good passing offense cannot be explained by your pass defense.
     
    The Guy likes this.
  36. xphinfanx

    xphinfanx Stay strong my friends.

    10,823
    2,214
    113
    Nov 1, 2009
    Chase Young
    then
    Herbert.
     
  37. Hoops

    Hoops Well-Known Member

    1,183
    1,484
    113
    Dec 11, 2016
    Minshew will be benched soon enough and deservedly so.
     
  38. AGuyNamedAlex

    AGuyNamedAlex Well-Known Member

    3,582
    2,579
    113
    Sep 12, 2015
    I dont think anyone forgets it happened. That team lacked talent all over the offense, I wouldnt say it was just Chad who couldnt generate chunk yards.

    All the WC really did was give other teams something new to prepare for. We still had strong QB play and I dont believe just plugging any QB in provides similar results.
     
  39. Fin D

    Fin D Sigh

    72,252
    43,684
    113
    Nov 27, 2007
    I disagree.

    Penny has reached some mythical status with this fanbase for some reason. The point is, we HAD to go gimmick BECAUSE Penny's QB play wasn't good enough. He was smart and managed the game great, but his arm was straight up deficient.
     
    resnor likes this.
  40. ExplosionsInDaSky

    ExplosionsInDaSky Well-Known Member

    3,163
    2,325
    113
    Sep 13, 2011
    We also had the luxory of an easy schedule that year. We also had freak occurances like the game where we beat the Jets behind Tedd Ginn returning 2 TD's. We got extremely lucky at times that year and were still barely able to win our division. Pennington definitely did an admiral job, but we were the epitome of a team that was overrated and it showed once we played Baltimore in the Playoffs. This year...That team is the Buffalo Bills.
     

Share This Page