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

Ryan Tannehill

Discussion in 'Other NFL' started by bbqpitlover, Oct 16, 2019.

Ryan Tannehill is...

  1. A terrible QB

    0 vote(s)
    0.0%
  2. A below average QB

    4 vote(s)
    5.7%
  3. An average QB

    7 vote(s)
    10.0%
  4. An above average QB

    39 vote(s)
    55.7%
  5. An elite QB

    16 vote(s)
    22.9%
  6. The GOAT.

    4 vote(s)
    5.7%
  1. cbrad

    cbrad .

    10,659
    12,657
    113
    Dec 21, 2014
    Not ignoring that, and if he keeps this up sure I'll change my opinion. But you can't just ignore all the data from Miami. That's the bulk of the data on Tannehill and how he responds to pressure.. unless you think he's like a completely different QB, but you did say you didn't think that.
     
  2. FinFaninBuffalo

    FinFaninBuffalo Well-Known Member

    2,474
    2,954
    113
    Dec 13, 2007
    Going to die on this hill now?..... LOL.

    Tannehill is going to burn you twice.
     
  3. cbrad

    cbrad .

    10,659
    12,657
    113
    Dec 21, 2014
    Dude.. I don't mind being wrong. I just go with what the stats say. You're right though that Tannehill has bucked the predictions from statistics several times, and in important ways. Not changing my approach though because it works most of the time (stats tell you what happens most of the time).

    Anyway we'll see. I don't really care one way or the other but I do enjoy the discussion.
     
  4. FinFaninBuffalo

    FinFaninBuffalo Well-Known Member

    2,474
    2,954
    113
    Dec 13, 2007
    Or unless the pressure is substantially different. That, I do believe. And it is supported by data. He has been very good since at least 2016 against the blitz. Where he struggled in Miami is quick pressure from 3 or 4 rushers, just like Tom Brady.
     
    resnor likes this.
  5. FinFaninBuffalo

    FinFaninBuffalo Well-Known Member

    2,474
    2,954
    113
    Dec 13, 2007
    Sorry to say this, but you rely on stats too much. You need to look at more film. There is no reason to rely exclusively on stats to understand what has already happened and has been recorded. That is the part that you stats guys get wrong.
     
  6. cbrad

    cbrad .

    10,659
    12,657
    113
    Dec 21, 2014
    For large sample size I think the stats based approach is superior. Obviously with Tannehill's move to Tennessee it didn't work, but more often than not it works. For small sample size you know I look at film. The discussion on Tua alone should tell you that. Comments I've made on timing with Wallace and the 4- or 2-minute offense were also mostly film based.

    So it's not exclusive. But in my experience when you have large sample size it's better to defer to stats. This same kind of debate occurred in many other fields (e.g., in medicine which I'm more familiar with), and stats won out eventually.
     
  7. FinFaninBuffalo

    FinFaninBuffalo Well-Known Member

    2,474
    2,954
    113
    Dec 13, 2007
    The point is that even with large sample sizes, stats alone isn't always the best approach. More use of film of Tannehill in Miami would have given you an inkling. I used film for the entire time in Miami to understand what the numbers meant. I argued with people for years that the film was showing that much more of the blame should be placed on the OL. They argued that the team was changing the OL year after year and it MUST be Tannehill, despite what was on film.

    I agree that the gap between Tannehill in Miami and Tennessee is highly unusual. You interpreted it as a "QB change" and a change of that size is improbable. I interpret it as it is very unusual for a team to be unable to fix their issues for so long. The stats would tell you that it must be Tannehill, IMO the film said otherwise. That interpretation is further supported by two things, the performance of the players and coaches that Miami dumped in the Tannehill era and the play of Tannehill in Tennessee.
     
    resnor likes this.
  8. cbrad

    cbrad .

    10,659
    12,657
    113
    Dec 21, 2014
    Has to be a method that can be applied. That is, you should be able to say do X, Y and Z and everyone who does that should: 1) come to the same conclusions, and 2) be better at predicting events compared to not doing X, Y and Z.

    I see no way to incorporate film watching in that regard because even pros disagree too much watching the same film. Again, only with large sample size of course because with small sample size it's the stats that are unreliable.

    The solution in other fields has generally been to develop better statistics, either by measuring more things (i.e., get better data) and/or by developing more capable statistical methods. In football what's missing is the better data.
     
  9. FinFaninBuffalo

    FinFaninBuffalo Well-Known Member

    2,474
    2,954
    113
    Dec 13, 2007
    How many games is it going to take to give Tannehill the credit he deserves?

     
  10. FinFaninBuffalo

    FinFaninBuffalo Well-Known Member

    2,474
    2,954
    113
    Dec 13, 2007
    Well until you can develop stats in football that don't "need to be a lot better", I'll use film to augment the admittedly poor stats.
     
    Last edited: Dec 21, 2020
    resnor likes this.
  11. FinFaninBuffalo

    FinFaninBuffalo Well-Known Member

    2,474
    2,954
    113
    Dec 13, 2007
    I do find it bizarre that the Tannehill situation hasn't made you made you question Tannehill's stats in Miami and everything you thought they meant. IMO, that is a huge blindspot, regardless of your position on the use of stats.
     
    resnor likes this.
  12. cbrad

    cbrad .

    10,659
    12,657
    113
    Dec 21, 2014
    Again I prioritize methodology over ad hoc approaches that might work in one case but won't in many others.
     
  13. FinFaninBuffalo

    FinFaninBuffalo Well-Known Member

    2,474
    2,954
    113
    Dec 13, 2007
    Okay, so you'll be wrong in those cases and then double down. Cool.
     
    Cashvillesent likes this.
  14. FinFaninBuffalo

    FinFaninBuffalo Well-Known Member

    2,474
    2,954
    113
    Dec 13, 2007
     
  15. cbrad

    cbrad .

    10,659
    12,657
    113
    Dec 21, 2014
    Yeah. Better to test the method, see where it doesn't work, and develop a better one than keep adding ad hoc assumptions to something that isn't a method. Long term this approach is more successful. Short term not necessarily.
     
  16. The_Dark_Knight

    The_Dark_Knight Defender of the Truth

    11,817
    10,321
    113
    Nov 24, 2007
    Rockledge, FL
    Every quarterback needs a good offensive line to play effectively. To argue otherwise is moronic!!!
     
  17. FinFaninBuffalo

    FinFaninBuffalo Well-Known Member

    2,474
    2,954
    113
    Dec 13, 2007
    But we are talking about ONE CASE ... Tannehill. You're wrong. Keep your approach otherwise.
     
  18. cbrad

    cbrad .

    10,659
    12,657
    113
    Dec 21, 2014
    No. I need to see which stats would have best predicted whatever we are about to see. That's how you improve on statistical methodology. Until then the only changes in interpretation will be those due to Tannehill actually changing the probabilities of different things occurring. Of course, with large enough sample size an "anomaly" can turn into a "pattern" and then the interpretation will naturally change, but in some (many?) cases that will probably come a bit too late for your liking.

    Either way I'm taking the long view on things. For me this isn't even about Tannehill except that he's providing a good case study on when a (normally) reliable prediction can go wrong. That's useful information for improving statistical analysis.
     
  19. FinFaninBuffalo

    FinFaninBuffalo Well-Known Member

    2,474
    2,954
    113
    Dec 13, 2007
    LOL.... we are discussing what Tannehill had already done, not predicting what he will do. If you want to know the current weather or the weather from from yesterday, you don't need stats. You need to stick your head out the window or check the tape......
     
    Cashvillesent likes this.
  20. cbrad

    cbrad .

    10,659
    12,657
    113
    Dec 21, 2014
    It's not that simple. There are many possible hypotheses, each making different predictions, and each having some likelihood of being correct given the data. As it currently stands many of those are still viable. Can't just assume one is correct before the data show statistical significance is achieved.

    For one key hypothesis, whether Tannehill is at minimum above average and that his Tennessee performance isn't just consistent with random variation, I provided the week by week analysis of that until it became clear there was no need to keep testing that hypothesis. But for many other hypotheses, we're not at the stage where it's clear which hypothesis best fits the data. Not rushing this dude.

    Also, when I say "predictions" I mean "derive the data from assumptions", even if the data is already there. Just so you know.
     
  21. FinFaninBuffalo

    FinFaninBuffalo Well-Known Member

    2,474
    2,954
    113
    Dec 13, 2007
    So you’ve concluded that Tannehill isn’t the QB you thought he was. Why doesn’t that extend to how you thought he deals with pressure? You should be rethinking everything you thought you knew about Tannehill.
     
    resnor and Cashvillesent like this.
  22. FinFaninBuffalo

    FinFaninBuffalo Well-Known Member

    2,474
    2,954
    113
    Dec 13, 2007
    Tannehill is DIFFERENT....
     
    resnor likes this.
  23. cbrad

    cbrad .

    10,659
    12,657
    113
    Dec 21, 2014
    Using pressure stats alone I can't make such an inference. Is there a way to condition pressure stats on other stats that would improve predictive power? Maybe. Haven't looked into that. That's still something that probably needs larger sample size in Tennessee though (intuitively) and is better left till another year or two passes before looking into.

    Clearly, if you just use statistics there's an entire distribution of passer rating decrements as a function of pressure, so there are a lot of QB's that would fall into the category of needing a good OL. Tannehill isn't "different" from many others.
     
  24. Cashvillesent

    Cashvillesent A female Tannehill fan

    770
    641
    93
    Dec 8, 2019
    Anyways Im just glad we no longer have Mariota, Locker, Vince Young linning up under center. Titans finally found a QB that can throw the ball and make the recievers better. I mean he is on pace on making both Davis and Brown 1,000 recievers on a first run offense. Thats crazy.
     
    resnor likes this.
  25. FinFaninBuffalo

    FinFaninBuffalo Well-Known Member

    2,474
    2,954
    113
    Dec 13, 2007
    You can't use pressure stats alone from 24 games of Tannehill in Tennessee to make an inference but you can use pressure stats alone from 13 games (that's all you shown) during his time in Miami to make an inference..... Got it....
     
  26. resnor

    resnor Derp Sherpa

    16,329
    9,874
    113
    Nov 25, 2007
    New Hampshire
    The proper approach would be, imo, to say that since the pressure stats are woefully inaccurate, then we can't come to any actual conclusions from them.

    The point of the stats is to be able to compare QBs to each other across the league. Unfortunately, the current statistics are not robust enough to allow a true comparison.
     
  27. cbrad

    cbrad .

    10,659
    12,657
    113
    Dec 21, 2014
    I don't have the links handy anymore, but the pressure stats argument was pretty well trodden for many years while Tannehill was in Miami. It was at least from 4 years there.

    Even if I had all the stats handy right now we still wouldn't achieve statistical significance (most likely) because separating "pressure" vs. "no pressure" reduces sample size and you saw how much sample size was necessary with passer rating without such separation to get statistical significance.
     
  28. resnor

    resnor Derp Sherpa

    16,329
    9,874
    113
    Nov 25, 2007
    New Hampshire
    They problem is cbrad, that we debated this nonstop when he was in Miami. Myself and a few others beat the drum in regards to the pressure stats being inadequate because they don't take into account blitz vs no blitz, etc. All pressure is not equal, and all QBs aren't in the same situations with the same variables affecting them. Therefore, it's really inaccurate to pretend that you can draw actual conclusions from stats that are missing incredibly important information.
     
    Cashvillesent likes this.
  29. cbrad

    cbrad .

    10,659
    12,657
    113
    Dec 21, 2014
    Like I said, the confidence intervals are large (i.e., the evidence is not strong), but it's still evidence. You don't just throw information like this out. So I have no issue with someone saying it's "inconclusive" (that's a reasonable position) but to say it's not evidence is flat out wrong because it changes the likelihood of different hypotheses being true.
     
  30. Fin-O

    Fin-O Initiated Club Member

    11,375
    11,392
    113
    Sep 28, 2015
    Tannehill getting snubbed for the pro bowl once again is laughable.

    I have no doubts he’s the next alternate in line but it really shouldn’t even come down to that.
     
    resnor and Cashvillesent like this.
  31. Cashvillesent

    Cashvillesent A female Tannehill fan

    770
    641
    93
    Dec 8, 2019
    Thats why i never take probowls serious. Its a popular contest honestly....
     
    Fin-O and Tin Indian like this.
  32. Tin Indian

    Tin Indian Rockin' The Bottom End Club Member

    7,929
    4,404
    113
    Feb 10, 2010
    Palm Bay Florida
    I can't believe I am posting i this again.

    They don't even have to be "good" just capable would be enough. Throughout Thills time here he had a terrible offensive line, outside of a few games with the so called "Unicorns". How many times did we watch a guy blow up the middle between guard and center and blast Thill as soon as he finished his drop? That's not on Tannehill. How often was the pocket non existant? The problem was the oline. Always was and those of us that tried to point it out were told that it doesn't matter. Well I think the evidence shows clearly that it does matter and it's obvious. The real data is not lumping all of Tannehills games through his career together and latching onto that number, The real number is the comparison between his numbers in Miami and in Tennessee. It's obvious, Tannehill's troubles in Miami was the offensive line. He didn't all of a sudden become a different guy or have the light come on since he got the Tennessee job. He has always been what we see now.

    We were right and all of you that say other wise were wrong. Just have the stones to admit it and get over yourselves.
     
  33. FinFaninBuffalo

    FinFaninBuffalo Well-Known Member

    2,474
    2,954
    113
    Dec 13, 2007
    cbrad wants us to believe that Tannehill was somehow transformed from the end of the 2018 season (where he admittedly played as poor as I had ever seen him play) to his first start in Tennessee. This somehow happened while going to a new team, learning a new system, getting familiar with new teammates with backup reps in practice..... just magical.....

    or he went to a team with a good enough supporting cast and coaching.
     
    resnor, Tin Indian and Cashvillesent like this.
  34. cbrad

    cbrad .

    10,659
    12,657
    113
    Dec 21, 2014
    Why do you have to resort to deliberate misrepresentations? That entire statistical hypothesis testing I did was to see if Tannehill was actually above average but being held back in Miami (even though performance was clearly average). There are now 2 years worth of posts of mine explaining that, yet you have to resort to deliberate misrepresentations to make an argument.

    What I did say however is that the jump is so large that it’s hard to explain this with change of surrounding cast alone (using stats), and that you can see evidence he’s improved as a player, for example in the 4- and 2-minute offenses. Hardly a controversial position except for people hypersensitive about Tannehill.
     
    Last edited: Dec 22, 2020
  35. FinFaninBuffalo

    FinFaninBuffalo Well-Known Member

    2,474
    2,954
    113
    Dec 13, 2007
    You're the one saying he isn't the same QB.... WTF am I missing here?

    upload_2020-12-22_13-25-10.png
     
    Cashvillesent likes this.
  36. cbrad

    cbrad .

    10,659
    12,657
    113
    Dec 21, 2014
    You’re missing or deliberately ignoring almost everything I’ve said. Even this latest debate started with me talking about his much improved surrounding cast. Clearly that’s primary difference. That does not imply Tannehill himself could not improve.

    Again, nothing controversial.
     
  37. FinFaninBuffalo

    FinFaninBuffalo Well-Known Member

    2,474
    2,954
    113
    Dec 13, 2007
    It is NOT hard to explain. If a QB that doesn't change teams can swing by 30 points from one season to the next, why can't a QB that goes from terrible circumstances to very good circumstances experience the same swing?

    Tom Brady had a passer rating of 87.9 in 2006, a rating of 117.2 in 2007, a rating of 87.3 in 2013 and back up to 112 in 2016.

    That is multiple 30 point swings. Are you suggesting that Brady "wasn't the same QB" in the down years? WTF? It is ABSOLUTELY the supporting cast, opposing teams, etc.
     
    Cashvillesent likes this.
  38. cbrad

    cbrad .

    10,659
    12,657
    113
    Dec 21, 2014
    That’s why you use statistical tests. Not going through the entire argument again. Anyway I think we’ve reached the point where you’re just ignoring the arguments I’ve made so it’s time to end this debate.
     
  39. Cashvillesent

    Cashvillesent A female Tannehill fan

    770
    641
    93
    Dec 8, 2019
    Or its a waste of time. Stats are pretty useless in football considering the fact that alot of things have to develop with different skill position for a play to be succesful. Deny that.
     
  40. FinFaninBuffalo

    FinFaninBuffalo Well-Known Member

    2,474
    2,954
    113
    Dec 13, 2007
    He was held back in Miami. End of story. No caveats. No asterisk. No additional unexplainable improvement over 6 months of inactivity. Nothing. You (and many others) stubbornly misread the situation in Miami. Tannehill isn’t suddenly faster, stronger, more accurate, more knowledgeable, or anything else. He is playing on a team that is better prepared, has a better supporting cast, and has better play calling. That’s it. Admit it and we can move on.....

    And, BTW, his situation in Tennessee isn't some magical, unreproducible, nirvana. It is good, maybe very good. And, in that situation, he has been excellent. His situation will fluctuate. He will have worse situations and his passer rating will suffer, just like every other QB. But, at this point, his high water mark is simply undeniable.
     
    Last edited: Dec 22, 2020

Share This Page