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

Making a Case for Ryan Tannehill

Discussion in 'Miami Dolphins Forum' started by KeyFin, Dec 27, 2018.

  1. djphinfan

    djphinfan Season Ticket Holder Club Member

    94,728
    47,008
    113
    Dec 20, 2007
    There’s just too much bad tape out there at the end of the season..Taylor had just come off a pretty decent season, playoffs, and I don’t think anyone was sure what the bills were gonna do.

    I think everyone knows what we’re gonna do..we have no leverage.
     
  2. pumpdogs

    pumpdogs Well-Known Member

    3,779
    1,383
    113
    Sep 22, 2009
    delaware
    The Bills made the playoffs because of the Defence last year.Taylor had a QB rating of only 89.2 and only threw 14 tds all year.
    If you remember the offence was so ineffective with him they benched for a game for Nate Peterman.Peterman was so bad(5 ints in the first half against chargers)he got his job back by default.They still traded him so don't be shocked if we can get something for him.
     
    mbsinmisc likes this.
  3. Surfs Up 99

    Surfs Up 99 Team Flores

    1,495
    1,431
    113
    May 5, 2016
    Good points! Being careful what we wish for has bit us in the backside before. I think it is important to find a QB the whole team believes in because it can make a huge difference. Don't the Browns look like a totally different team with Mayfield? They are playing hard and with purpose. I am concerned that if we don't find a guy the players have confidence in then they might start mailing it in. I can't remember the specifics, but wasn't there issues with players not having confidence in Tannehill? Not doubt Grier has a big challenge this off-season.
     
    KeyFin likes this.
  4. The Guy

    The Guy Season Ticket Holder Club Member

    475
    435
    63
    Oct 1, 2018

    The question of whether he should be replaced, and the question of whether he can be replaced are two separate issues.

    Whether he can be replaced is based on who else is available.

    Whether he should be replaced is based on whether the team can win a Super Bowl with him.

    The problem with the statistical analysis above is that it totally discounts situational football. Tannehill plays at a slightly above-average level as measured by statistics that don't incorporate situational variables such as the strength of opposing defenses or scoreboard situations.

    With regard to statistics that do incorporate those variables, he's routinely in the bottom quarter of the league. This year he was 32nd in the league, for example. In his best season with regard to traditional statistics (such as passer rating), he was nonetheless in the bottom third of the league with regard to statistics that incorporate situational variables.

    In other words, Tannehill tends to make his living against poorer defenses, in non-clutch situations.

    The above has strong relevance to playoff football, where, not coincidentally, we have never seen Ryan Tannehill play!

    Here's a thought experiment: replay the past seven years as though Tannehill and the Dolphins made the playoffs several times, only to have Tannehill falter badly in the first round and contribute heavily to the team's early exit.

    Based on statistics that incorporate situational variables (again, primarily strength of opposing defenses and in-game situations), that's precisely how he would've been predicted to have played.

    So, consider that experience we haven't been privy to -- Tannehill's expected performance in the playoffs -- is perhaps the single best reason to consider that he should be replaced.

    The team simply cannot win a Super Bowl with him, in my opinion.
     
    Last edited: Jan 5, 2019
    resnor and Bumrush like this.
  5. Nappy Roots

    Nappy Roots Well-Known Member

    10,323
    4,246
    113
    Dec 3, 2007
    Bradenton,FL
    I know its not perfect, but ESPNs QBR take situations into account. ESPNs QBR rankings from last season...

    30. Blake Bortles
    31. Lamar Jackson
    32. Ryan Tannehill
    33. Josh Rosen

    For me, that stat shows reality far more than traditional QB Rating. Neither are perfect however.
     
    Bumrush and The Guy like this.
  6. The Guy

    The Guy Season Ticket Holder Club Member

    475
    435
    63
    Oct 1, 2018

    No statistic is perfect, but again, if a quarterback plays routinely poorly against better defenses and in clutch situations -- over a seven-year sample of play -- how can he possibly be viewed as a player who can contribute to a Super Bowl win, when the road to the Super Bowl will inevitably consist of such variables?
     
    Nappy Roots likes this.
  7. Dolphin Dundee

    Dolphin Dundee Well-Known Member

    418
    282
    63
    Sep 27, 2015
    Interesting that with Tannehill return to Miami next year undecided have any current Dolphins been lobbying for him to stay?
     
    Bumrush likes this.
  8. Nappy Roots

    Nappy Roots Well-Known Member

    10,323
    4,246
    113
    Dec 3, 2007
    Bradenton,FL
    I agree. I have had this view for a few years now...
     
    Bumrush and The Guy like this.
  9. cbrad

    cbrad . Club Member

    6,969
    7,744
    113
    Dec 21, 2014
    ESPN's QBR also has Aaron Rodgers ranked #20 this year. So he's also pretty bad at situational football and clutch situations right?

    It's easy for anyone to proclaim they "take game situation into account". The devil lies in the details and ESPN and Football Outsiders never publish those details. So I'd be careful about thinking they're better representing reality since we have no idea precisely how they apportion credit among players.

    Also, the correlation between the two stats averages around 0.8 so they're pretty similar to each other no matter how you look at it. And most people can't distinguish between lists that have that high correlation (meaning you have to take random variation into account), so I think all you're doing is looking at individual cases (e.g., Tannehill or like I pointed out with Rodgers) which is no way to evaluate a proposed metric.

    Either way, whether you prefer QBR or DVOA over traditional passer rating or not, don't call those "statistics". That's an insult to the whole field of statistics because a statistic has to be derived from data. If all they have is a complicated formula that incorporates subjective assumptions about how to apportion credit among players, then that's a "model" not a "statistic". Even traditional passer rating has this problem, but they say (can't verify this myself) that the weights on the different components of passer rating were inferred from data in the 1970's and if that's true then it's a statistic.
     
    The Guy likes this.
  10. The Guy

    The Guy Season Ticket Holder Club Member

    475
    435
    63
    Oct 1, 2018

    But consider that there is some accurate measure of situational quarterback play that can exist, and if it also correlates at 0.8 to traditional quarterback statistics, that leaves 36% of the variance unexplained by the latter. What if that 36% becomes highly relevant during playoff football, and Tannehill again has shown very poorly with regard to those types of statistics?

    In other words, if these rough approximations are good enough for government work so to speak, then we have to consider that they indicate Tannehill could be highly unlikely to win a Super Bowl.
     
    Bumrush likes this.
  11. cbrad

    cbrad . Club Member

    6,969
    7,744
    113
    Dec 21, 2014
    Yeah but we have no idea if QBR is even a decent approximation to an accurate measure of situational QB play. We don't even know the range of variance explained (or unexplained) any of these stats or models have to whatever is the truth. So specific numbers here are meaningless. As stated, it would be MUCH easier to evaluate how well QBR captures situational QB play if they published their method.

    More importantly, I don't think you want to use any stat to summarize how Tannehill plays in clutch situations. In fact, you'll probably just run into unnecessary "unreliability" issues because of small sample size. That being said, I think it's clear Tannehill is bad in clutch situations. For example, just pick specific win-or-go-home games (e.g. last 2 in 2013 and arguably last 4 this year, etc..) and look at how he played. His stats in those games generally support the impression most people had watching those games, but using a stat to summarize his play would be problematic for sample size reasons (for both passer rating and QBR).
     
    The Guy likes this.
  12. KeyFin

    KeyFin Well-Known Member

    5,714
    7,484
    113
    Nov 1, 2009
    Well, almost every game he won this year (and in 2016) came down to a last-minute drive in a one-score game. One could argue that he's pretty darn good in the clutch.

    As I've said before, I personally rate Ryan Tannehill on his December performances...I always have and it's why I wanted him replaced 3, 5 and even 7 years ago. And while 2016 was great, it really stinks that we didn't get to see how he performed in those closing games. Maybe my opinion would be completely different if he finished with the same 2-1 record as Moore and then played average in the WC game. It just didn't happen though and I don't think he gets that benefit of the doubt when he's never closed a season well.

    As The Guy said, it is completely different "if" we should replace RT and "can we" replace RT. I think the first point is a foregone conclusion...we should part ways AS SOON AS A BETTER PROSPECT IS ON THE ROSTER. I do not agree in just firing him though and expecting miracles to discover the new Dan Marino.

    I honestly think we'd be better off as a franchise with another season of Tannehill if we can't find someone this off-season or in the draft. That next name doesn't have to be better than RT on day one, but he should definitely show enough potential to be at that level sometime in 2019. But if such a person doesn't exist, then I'm regretfully still on the Tannehill bandwagon for the time being.
     
    Pauly, resnor and The Guy like this.
  13. cbrad

    cbrad . Club Member

    6,969
    7,744
    113
    Dec 21, 2014
    Uhh.. no you can't just cherry pick cases where we ended up winning! Much better than looking at individual plays or drives is how he played in crucial games, you know games where we (almost mathematically) had to win to have a chance to make the playoffs. And Tannehill is pretty pathetic there for the most part.

    I think you've agreed with this before, but Tannehill is simply a QB you cannot rely on to win games for you. He plays well when everything else is going fine, but not when they don't. And usually the going gets a lot tougher on the road and in December. So how can a guy that can't be relied on to win a game could be good in the clutch??

    No one is expecting miracles. Most QB's don't become great. The issue is that the longer you keep saying Tannehill's replacement is likely to be worse turns into one more year you have a mediocre QB. Change for change sake is important here because we KNOW Tannehill isn't good enough (for any decent play on the odds). You have to keep trying to find that franchise QB until you do.
     
    texanphinatic and The Guy like this.
  14. KeyFin

    KeyFin Well-Known Member

    5,714
    7,484
    113
    Nov 1, 2009
    I'm definitely not cherry picking. This season alone, either we won a close game or we got blown out....there really wasn't a middle road except for Indy (where I'd argue that Gase blew it more than anyone). To me, "clutch" means there's a minute left on the clock and you're going to live or die by your QB's arm....we've been good in those types of situations this season (even under Osweiler).

    I wouldn't consider "clutch" asking a QB to fight back in a 44-17 blowout though...that's obviously more problems than just the QB. Hopefully that clarifies my thinking there.

    I'm making a quick poll in another thread basically to show you, yes...people here do expect miracles.
     
    resnor likes this.
  15. cbrad

    cbrad . Club Member

    6,969
    7,744
    113
    Dec 21, 2014
    Every person can have a different definition of "clutch" so instead of using that word I think it's better to focus on "must win" situations. And most such plays occur in (almost mathematically) must-win games where Tannehill is pretty bad. As far as our blowouts, it was really "blowouts on the road" which brings up another issue: unless you're the #1 seed you're going to have to play well on the road and clearly Tannehill isn't the QB you want in those cases.
     
    KeyFin and The Guy like this.
  16. The Guy

    The Guy Season Ticket Holder Club Member

    475
    435
    63
    Oct 1, 2018

    I would define "clutch" as situations in which the caliber of the player's performance (good or bad) changes the team's likelihood of achieving its goals a great deal. And that can range all the way from one play to an entire game (or even several games), based on the situation.
     
    resnor likes this.
  17. adamprez2003

    adamprez2003 Senior Member

    37,540
    14,763
    113
    Nov 27, 2007
    new york ciity
    Let's be clear about how the season ended. We faced the number one, two and three defenses against the pass without an offensive line or receivers who could get quick seperation. It had nothing to do with being clutch. It had nothing to do with being december. It had only to do with horrible matchups.
     
    resnor and KeyFin like this.
  18. The Guy

    The Guy Season Ticket Holder Club Member

    475
    435
    63
    Oct 1, 2018

    And that's in essence the story of Tannehill's career. He requires favorable matchups to play well.
     
    KeyFin, cbrad and Dolphin Dundee like this.
  19. cbrad

    cbrad . Club Member

    6,969
    7,744
    113
    Dec 21, 2014
    Interesting.. there's actually some scientific literature on this and one concept that's common to how they use the word is "performance under pressure". That is.. they try to either measure pressure directly or indirectly and then measure how well an athlete did. Needless to say there's no agreement on how to infer how much "pressure" there is on an athlete.
     
    The Guy likes this.
  20. adamprez2003

    adamprez2003 Senior Member

    37,540
    14,763
    113
    Nov 27, 2007
    new york ciity
    Everyone does. Football is a game of matchups.
     
    resnor likes this.
  21. The Guy

    The Guy Season Ticket Holder Club Member

    475
    435
    63
    Oct 1, 2018

    It may not bear on how much pressure the player is actually subjectively experiencing, but you could argue that pressure is a function of the degree to which a team’s likelihood of achieving its goals will change based on the player’s performance.
     
    cbrad likes this.
  22. The Guy

    The Guy Season Ticket Holder Club Member

    475
    435
    63
    Oct 1, 2018

    In an absolute sense that’s true, but there is also variation in how well players play as a function of the strength of the opposition, and Tannehill plays among the worst QBs in the league when the opposition is strong. So he requires favorable matchups more so than do other QBs.
     
  23. cbrad

    cbrad . Club Member

    6,969
    7,744
    113
    Dec 21, 2014
    Precisely how I'd approach it.

    There have been several huge debates about clutch on this board, mostly years ago, and I suggested you could try to operationally define clutch as "performance given a function of win probability and the mathematical chances of winning the SB". The function of win probability would just be 1-WP (so if win probability is 0.75 that becomes 0.25) and then multiply that by the mathematical chances of winning a SB. That way you get both in-game and across-game probabilities in there.
     
    Last edited: Jan 5, 2019
    The Guy likes this.
  24. adamprez2003

    adamprez2003 Senior Member

    37,540
    14,763
    113
    Nov 27, 2007
    new york ciity
    I wouldn't quite phrase it like that but I understand your meaning and agree. Tannehill can play well when the matchups are even or favorable. When the matchups are unfavorable, especially oline mismatches he becomes useless and where another QB might have a respectable loss, Tannehill will have a horrendous loss. Both will lose because you dont beat mismatches but one will look better doing it
     
    The Guy and resnor like this.
  25. djphinfan

    djphinfan Season Ticket Holder Club Member

    94,728
    47,008
    113
    Dec 20, 2007
    Your simply not accounting for the impact what his legs did for the team..
     
  26. djphinfan

    djphinfan Season Ticket Holder Club Member

    94,728
    47,008
    113
    Dec 20, 2007
    Backup QBs can play well..
     

Share This Page