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

Titans to start Ryan Tannehill

Discussion in 'Miami Dolphins Forum' started by bbqpitlover, Oct 16, 2019.

  1. The Guy

    The Guy Well-Known Member

    1,598
    831
    113
    Oct 1, 2018
    cbrad I definitely defer to you with regard to the correctness of any statistical approach. However my issue in this instance is that, if we're comparing his current eight-game stretch to all others in his career and determining the significance of the degree to which this one deviates from his typical one, then wouldn't the two eight-game stretches I pointed out in my post above also deviate significantly (statistically) from his typical eight-game stretch?

    If so, then what meaning do we make of that? Certainly we can't attribute that to "Dolphins versus Titans," because he was on the Dolphins for two of those and the Titans for the third, and all three deviate significantly statistically from his norm.
     
  2. cbrad

    cbrad . Club Member

    7,381
    8,650
    113
    Dec 21, 2014
    Well.. yes and no. Not sure which aspect you're referring to so let me give a two part answer.

    1. Yes it's "all other things being equal" if you're talking about whether the passer rating formula itself is capturing (true) QB ability in the same way for every 8-game stretch. Nothing else we can do except assume that hasn't changed, so from that point of view yes it's assuming "all other things being equal".

    2. No it's not assuming "all other things are equal" because it's literally testing to see if, statistically speaking, all other things were equal. That is, it's testing to see whether you could explain the performance in Tennessee by random variation alone.

    Hope that clears things up a bit.
     
    Irishman likes this.
  3. cbrad

    cbrad . Club Member

    7,381
    8,650
    113
    Dec 21, 2014
    They don't. The 2019-adjusted mean of the means of all 8-game stretches is 91.87 while the standard deviation is 8.82. So statistical significance is found below 75.2 and above 110.5, and the lowest rating there is 76.2 and the highest from Miami is 104.95.

    The nature of the test changes when you look at sampling distributions (distributions of the means of samples). The t-test on the other hand is looking at the distribution of differences between pairs of games, one from one set and one from the other. Very different concept, but in general results from different tests should agree, and they do here because one is indeterminate due to sample size.
     
    Irishman and The Guy like this.
  4. AGuyNamedAlex

    AGuyNamedAlex Well-Known Member

    1,674
    1,446
    113
    Sep 12, 2015
    Honestly I dont see how you can compare two stretches with different supporting casts at all.

    The casts influence those #'s to a degree and dont tell us how much better Tannehill himself is playing.

    Unless you just believe a QB puts up the same #'s in every situation.
     
    resnor and PhinFan1968 like this.
  5. cbrad

    cbrad . Club Member

    7,381
    8,650
    113
    Dec 21, 2014
    That was the scenario I was referring to in answer #1. You have to assume that the degree to which passer rating is capturing (or influenced by) true QB ability remains the same + random variation of course. Like I said, nothing else we can do but to assume that.

    Now if that assumption is too much to bear then it's best not to use any stat for any comparison across 2 or more games. And that's fine if you take that approach, but you have to be consistent by never using stats for comparison across games.
     
    Irishman and AGuyNamedAlex like this.
  6. AGuyNamedAlex

    AGuyNamedAlex Well-Known Member

    1,674
    1,446
    113
    Sep 12, 2015
    That's fair enough I suppose, I was just curious how youd take that into account in the comparison.
     
  7. resnor

    resnor Derp Sherpa

    13,261
    7,469
    113
    Nov 25, 2007
    New Hampshire
    But if you're assuming something true to begin with, then you can't dogmatically stand behind those stats as actual proof of anything. The best you can say is that you are ASSUMING these stats to be proof.
     
  8. cbrad

    cbrad . Club Member

    7,381
    8,650
    113
    Dec 21, 2014
    Depends on the claim. If let's say sample size is not an issue and you show 2 sets of ratings are not significantly different from each other, that literally does prove (up to the threshold used to claim statistical significance, which is typically set at 5%) that one set of ratings is consistent with whatever produced the other set of ratings + random variation.

    All you're pointing out is that we wouldn't know which sources of random variation were most important, but that might be irrelevant for a particular discussion, for example if you're just interested in whether there's evidence the conditions were (significantly) different.

    The case we're dealing with right now is a bit trickier because you have statistical significance and people naturally start to care about WHY. That part the stats can't answer. But it's not true that you can't dogmatically stand behind an argument just because you can't tease apart the individual contributions to "random variation". It all depends on the claim made. For some claims you can, for others you can't.
     
    The Guy likes this.
  9. The Guy

    The Guy Well-Known Member

    1,598
    831
    113
    Oct 1, 2018
    And that's why we need more time to figure out what's going on. We could have a scenario here where the causes of Tannehill's improved performance are 1) his motivation to keep his starting job, 2) his motivation to get a bigger contract next year than he would have, and 3) elevated play, on a temporary basis, by the surrounding players, by virtue of the fact that the QB they reportedly wanted to be the starter out of camp was finally named the starter.

    None of those variables necessarily represents any improvement in the surroundings Tannehill experienced with the Dolphins, and none of them is likely to be sustained.

    On the other hand the causes of Tannehill's improved performance could be 1) better surrounding players, 2) better coaches, and 3) a better fit between the offensive scheme and Tannehill's particular abilities. Obviously those variables can far more easily be sustained.

    Only how Tannehill is playing about this time next year can tell us which of the above explanations (or others like them) is more likely correct.
     
  10. cbrad

    cbrad . Club Member

    7,381
    8,650
    113
    Dec 21, 2014
    Yeah I agree that more data will help determine which explanations are more likely to be correct, but I also think we'll never really figure this out. I mean what we're seeing is elite level play, not just a higher level of play per se, and I don't know how you "explain" that.

    It might also be worth contrasting two things: statistics "explaining" what happened vs. statistics changing the likelihood of different hypotheses being correct.

    I would say statistics "explains" what happened if you can adjust a stat for some confounding variable (e.g., adjust passer rating for defense) and show that it fits the data. That certainly doesn't "prove" that the particular confounding variable you adjusted for was the the reason for what was observed, but it does show that adjusting for the confounding variable was sufficient to explain the data and that the effect of all other variables cancel out – it would also be the most parsimonious explanation.

    I doubt we'll ever be able to do that for Tannehill's performance in Tennessee because we don't know how to quantify so many factors in the explanations people propose. I think for Tannehill's performance in Tennessee you'll always have a large number of plausible hypotheses no matter how much extra data you obtain and there will be no closure (we'll see).

    So I'm also interested in what he does in the future but not really to find an explanation for what he's done so far. I just want to see how he plays in must-win games and whether he can play at a high level for multiple years.
     
  11. AGuyNamedAlex

    AGuyNamedAlex Well-Known Member

    1,674
    1,446
    113
    Sep 12, 2015
    Not to detract from your other insightful posts, but I think this is one of your best posts here or at least one of my favorite.

    Well presented opinion written in a way people who arent big "numbers guys" or who just arent good with numbers period can understand and probably agree with at least on some level.
     
    Pauly, FinFaninBuffalo and cbrad like this.
  12. FinFaninBuffalo

    FinFaninBuffalo Well-Known Member

    462
    945
    93
    Dec 13, 2007
    Well said. Especially the 4th paragraph. There are just too many factors for the numbers to tell the story by themselves. That is why you have to see what is happening on the field.

    I see a guy playing at an elite level in terms of what "he" is doing. That includes getting the team out of bad plays and into good plays (acknowledged by his coach), making ridiculously accurate throws all over the field even in tight coverage, making plays with his legs, killing the blitz, and playing well under pressure (KC game). I can see every type of throw being made, bombs, back shoulder touch throws, darts between defenders, and laser beams into the tiniest of opening. I can see that the stats are not due to receivers bailing him out, or dump offs that get yards after the catch.

    I also don't see a lot of meanings stats. Against the Raiders, he only threw 3 passes in the 4th quarter and he still had 391 yards passing.
     
    resnor, cbrad and PhinFan1968 like this.
  13. FinFaninBuffalo

    FinFaninBuffalo Well-Known Member

    462
    945
    93
    Dec 13, 2007
    upload_2019-12-14_18-22-22.png
     
    Tin Indian, Pauly, smahtaz and 2 others like this.
  14. FinFaninBuffalo

    FinFaninBuffalo Well-Known Member

    462
    945
    93
    Dec 13, 2007
    upload_2019-12-14_18-31-22.png
     
  15. FinFaninBuffalo

    FinFaninBuffalo Well-Known Member

    462
    945
    93
    Dec 13, 2007
  16. FinFaninBuffalo

    FinFaninBuffalo Well-Known Member

    462
    945
    93
    Dec 13, 2007
    Quotes from an article that back up what I am seeing:


    https://www.washingtonpost.com/spor...s-by-playing-like-one-nfls-best-quarterbacks/


    upload_2019-12-14_18-46-8.png

    upload_2019-12-14_18-44-52.png
     

    Attached Files:

    Tin Indian, resnor and PhinFan1968 like this.
  17. FinFaninBuffalo

    FinFaninBuffalo Well-Known Member

    462
    945
    93
    Dec 13, 2007
    upload_2019-12-14_18-54-34.png
     
    resnor and PhinFan1968 like this.
  18. FinFaninBuffalo

    FinFaninBuffalo Well-Known Member

    462
    945
    93
    Dec 13, 2007
    upload_2019-12-14_18-56-15.png
     
    Tin Indian, Irishman, smahtaz and 2 others like this.
  19. FinFaninBuffalo

    FinFaninBuffalo Well-Known Member

    462
    945
    93
    Dec 13, 2007
  20. PhinFan1968

    PhinFan1968 To 2020, and BEYOND! Club Member

    Just staaaaahp already...we all know he was the problem in Miami....

    :yahoo::yahoo::yahoo::yahoo::yahoo::yahoo:

    Seriously though...keep 'em comin'!
     
    Irishman and resnor like this.
  21. PhinFan1968

    PhinFan1968 To 2020, and BEYOND! Club Member

    Key (something he was routinely dogged out about in Miami) --

    "This isn’t the first time that comments like these have surfaced either. John Glennon’s piece at The Athletic following the Panthers game included several quotes from teammates that talked about Tannehill holding players — particularly receivers — accountable and demanding that they be in the right spots at the right time if they want the ball.

    It’s pretty clear that the former 8th overall pick has staked his claim to a leadership role in this offense both on and off the field since being given the starting job."
     
    resnor likes this.
  22. Fin D

    Fin D Sigh Club Member

    72,134
    43,569
    113
    Nov 27, 2007
    Honestly, I feel like at this point, for people who have been here for awhile and really paid attention to the Thill supporters and their arguments, to still be acting like they don't know where this came from or how this could be happening, it just smacks of sour grapes.
     
  23. resnor

    resnor Derp Sherpa

    13,261
    7,469
    113
    Nov 25, 2007
    New Hampshire
    Tannehill is dominating the league in pretty much every thing that people on here have said for years he was bad at.
     
    Tin Indian, Irishman, smahtaz and 2 others like this.
  24. PhinFan1968

    PhinFan1968 To 2020, and BEYOND! Club Member

    Except sacks and ball security...those are still alive and well. Still...I'll take it.
     
    Tin Indian and Pauly like this.
  25. Fin D

    Fin D Sigh Club Member

    72,134
    43,569
    113
    Nov 27, 2007
    Its more than that, he's succeeding at the level we said he could if specific things about his surroundings were changed, which now are.
     
    Tin Indian, resnor and PhinFan1968 like this.
  26. Hoops

    Hoops Well-Known Member

    1,166
    1,443
    113
    Dec 11, 2016

    Ball security stuff was over blown. Still is.
     
    resnor and PhinFan1968 like this.
  27. PhinFan1968

    PhinFan1968 To 2020, and BEYOND! Club Member

    Very interested to hear your take on that, specifically. Seems he gets sacked a lot still (even though he had none against Oakland), and has fumbled a good bit.
     
  28. FinFaninBuffalo

    FinFaninBuffalo Well-Known Member

    462
    945
    93
    Dec 13, 2007
    I posted these on Finheaven for you. Probably justified here as well.

     
  29. Hoops

    Hoops Well-Known Member

    1,166
    1,443
    113
    Dec 11, 2016

    He’s gonna take some sacks other QBs don’t. Some of it is because he’s a pocket progression qb so he’s gonna stand in there and wait on some things longer than one read types and the other part of it is he’s always been a tick late to feel pressure and get out of dodge. So by the time he does feel it and he tries to bolt out it’s too late at times.

    so I think it’s a combo of two things. I can tell ya there are a lot of one read QBs in the nfl these days. If the primary isn’t there they are taking off with it. But guess what they are leaving plays on the field in the passing game cause they never get to their 3rd read in the progression. The take off with it mask it somewhat til the pocket deficiencies show up. Those things usually show up when you are chasing the scoreboard/sticks. If you never really are you can hide them. For the short term at least.

    That’s why I think so many QBs now have flash in the pan type results. Up one year down the next. Because the non pocket progression QBs are so dependent upon everything else around them. Be it coaching, scheme, o line play, defense, and talent.

    The nfls got a ton of non pocket progression/pro style QBs who are flat overrated or unable to sustain play year to year.

    Just look at the last few years even.

    the pre snap pocket progression qb is a dinosaur with all these college concepts invading the nfl but it’s also the one that has withstood the test of time.
     
    Last edited: Dec 14, 2019
    Pauly, Fin D, resnor and 1 other person like this.
  30. FinFaninBuffalo

    FinFaninBuffalo Well-Known Member

    462
    945
    93
    Dec 13, 2007
    This is what most people miss!!!!!
     
    PhinFan1968 likes this.
  31. PhinFan1968

    PhinFan1968 To 2020, and BEYOND! Club Member

    I much prefer the pocket guy...always have. I mean, Lamar is a freak, but he can't keep that up forever...they never do.
     
  32. Hoops

    Hoops Well-Known Member

    1,166
    1,443
    113
    Dec 11, 2016

    Yeah there’s some rg3 type stuff going on there in terms of the threat of his acceleration and legs doing some driving. And he’s taking some contact that over time scrapes shelf life. But he doesn’t take as much direct hard contact as rg3 did to date and he’s more instinctual a player.

    He’s a mucb more consistent passer right now than I think anyone could have seen him being on college tape. So the combo of that his instincts an offense that’s catered to his game and all those heavy tight end packages they run everything out of makes him right now almost impossible to stop.

    you stop the run and make him beat you from the pocket with his arm and if he does you flat tip your cap. But forcing him to beat you from the pocket is much easier said than done.

    that’s my current read on it at least.
     
    Last edited: Dec 14, 2019
    resnor and PhinFan1968 like this.
  33. Hoops

    Hoops Well-Known Member

    1,166
    1,443
    113
    Dec 11, 2016

    Well for example with tua I’m gonna run a rpo and west coast concepts based offense.

    and I think that is just about as sustainable as the pre snap progression read. I think the rpo is here to stay and it’s usage in the nfl is heavy.

    athough if I can get you behind the sticks I can limit the rpos impact and force your qb to do more reads wise from the pocket.

    rpo on 3rd and 8 plus isn’t gonna move the chains consistently in other words
     
    PhinFan1968 likes this.
  34. The Guy

    The Guy Well-Known Member

    1,598
    831
    113
    Oct 1, 2018
    And that was an issue with regard to not only sacks, but the plays that weren't made after evading pressure and throwing downfield.

    Think of all the plays Tannehill could've made if he'd had the ability to sense and evade pressure, extend the play, and hit receivers downfield. That's an awful lot of missing plays, and a big part of the puzzle that was missing in his game.

    Picture the days of Marino and his ability to step up in the pocket to evade pressure and make the throw downfield, and how many plays he made in that manner. A tremendous percentage of those opportunities were missed by Tannehill.

    I suspect that if there is an explanation for his current performance that involves the offensive scheme, it's that the Titans are minimizing that weakness in his game.
     
  35. Hoops

    Hoops Well-Known Member

    1,166
    1,443
    113
    Dec 11, 2016
    Someone already said on this board tannehill has a few flaws to his game none of which are fatal.

    And that person was spot on. Really only flaw I’ve found to be legit is feel for pressure in the pocket. And I don’t know how mucb of that is a result of muscle memory from having guards/tackles in his lap before his back foot could even hit for so long in Miami and how much of it is what he was always gonna be for instance.

    Everything else I’ve found to be fabrication frankly. In terms of what he’s asked to do and his execution. The timing stuff in 2018 in terms of ball out of hand is proving to be game rust more than anything else brought on by games and time missed due to injury.

    In terms of offense ask with Ryan he can run anything you’d ask of him. There are no limitations there and he can set/ verify protections and account for blitzers etc.

    there’s times also where he could be more instinctual but that’s also showing to be overstated with his off script play in 2019.

    ball goes where the coverage says it should go and some have coined that as robotic.
     
    Last edited: Dec 14, 2019
  36. The Guy

    The Guy Well-Known Member

    1,598
    831
    113
    Oct 1, 2018
    Right, and my point is that it isn't just sacks that translates to. It's also missed passing plays, and so when we're talking about passer rating for example, we have to consider how much more that statistic would've been elevated toward the level of the best QBs in the league if he'd had the ability to sense and evade pressure and extend plays.
     
  37. Hoops

    Hoops Well-Known Member

    1,166
    1,443
    113
    Dec 11, 2016
    If there are for example 6 sacks in a game and you want to tell me on average 2 of those will be on the qb with Ryan I’m fine with that.

    regardless it’s not a fatal flaw. Never has been. Not when you can do everything else at such a high level.
     
    PhinFan1968, resnor and Fin D like this.
  38. The Guy

    The Guy Well-Known Member

    1,598
    831
    113
    Oct 1, 2018
    But in many cases it doesn't matter which part of the team is to blame for the sacks. The issue is whether the best QBs in the league, at a far greater rate, are evading pressure, extending plays, and making plays downfield in that manner, regardless of who is at fault for the pressure on them, because they don't share Tannehill's deficiency in sensing and evading pressure.

    Again the point is to look not only at sacks, which are the observable result of such plays, but also at what didn't happen, which is evading pressure, extending plays, and hitting passes downfield.
     
  39. Hoops

    Hoops Well-Known Member

    1,166
    1,443
    113
    Dec 11, 2016
    Russell Wilson takes plenty of sacks that are on him. Completely on him. As does deshaun watson. there are others.

    and they aren’t progression read guys either.
     
    Irishman, PhinFan1968 and resnor like this.
  40. The Guy

    The Guy Well-Known Member

    1,598
    831
    113
    Oct 1, 2018
    Wilson could be the best in the league at hitting passes downfield after evading pressure and extending plays. If the rest of his game is no better than Tannehill's, then that alone could explain the difference between Wilson's second-highest career passer rating of all time of 101.2, and Tannehill's career rating of 89.1.

    If that's true, that's an awfully big difference in performance on the basis of a single trait.
     

Share This Page