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. resnor

    resnor Derp Sherpa

    16,327
    9,874
    113
    Nov 25, 2007
    New Hampshire
    I'm not saying the whole league is going to deflate. But Tannehill showed last year that he could play at high level. But I'd also be very surprised if the 7 guys currently over a 110 rating stay that high. It's a young season.
     
  2. cbrad

    cbrad .

    10,659
    12,657
    113
    Dec 21, 2014
    If Tannehill ends this year with 0.55 z-score (which btw is right in the middle of the range I said is realistic: 0-1), you wouldn't want to focus on that alone and ignore 2019. Last year counts too when comparing "Tennessee" vs. "Miami". If that difference is statistically significant, then this year would be better interpreted as random variation around a higher mean.
     
  3. resnor

    resnor Derp Sherpa

    16,327
    9,874
    113
    Nov 25, 2007
    New Hampshire
    But we have argued that seasons like those showed Tannehill was better than he was given credit for. When they lost Tannehill to injury, the team was markedly worse.
     
    The_Dark_Knight likes this.
  4. Bumrush

    Bumrush Stable Genius Club Member

    29,473
    34,332
    113
    Nov 25, 2007
    You guys need to sell popcorn and lazy boy recliners.
     
    The Guy likes this.
  5. The Guy

    The Guy Well-Known Member

    6,598
    3,323
    113
    Oct 1, 2018
    Why wouldn't it be that 2019 was random variation around a lower mean?
     
  6. cbrad

    cbrad .

    10,659
    12,657
    113
    Dec 21, 2014
    Because it's statistically significant and wouldn't be expected to occur at least once in 8 or so seasons around a lower mean.
     
  7. The Guy

    The Guy Well-Known Member

    6,598
    3,323
    113
    Oct 1, 2018
    Even though it consists of only 11 games?
     
  8. cbrad

    cbrad .

    10,659
    12,657
    113
    Dec 21, 2014
    Sure, sample size is taken into account when calculating statistical significance.
     
  9. The Guy

    The Guy Well-Known Member

    6,598
    3,323
    113
    Oct 1, 2018
    I thought we had already established that among the QBs who have performed similarly over 11-game stretches, too big a percentage of them were career average?
     
  10. cbrad

    cbrad .

    10,659
    12,657
    113
    Dec 21, 2014
    Different questions. That's conditioning on the QB having a similar 11-game stretch. If your goal however is to say we should interpret 2019 as just random variation for an average QB, then you have to condition on the QB being average and directly calculate the probability of 2019 occurring assuming the QB is average.
     
  11. Puka-head

    Puka-head My2nd Fav team:___vs Jets Club Member

    8,605
    6,743
    113
    Nov 25, 2007
    Slightly left of center
    Could someone show me the stats on statistical arguments IRT

    1) how many times a different statistical argument has been brought up in this particular debate by each side and whether that is average, above average, or elite?
    2) What statistical proof is there that statistical arguments in general change any ones opinion about anything ever. If so what is the minimum number of statistical arguments it takes to change someones opinion, is there a maximum? What is the average number of statistical arguments it takes to change someones opinion? Does skill level and and equipment of the statistician have any influence on the successful changing of opinions?
    3) the correlation between the number of statistical arguments made, the number of people making those statistical arguments, the number of numbers in those statistical arguments and the number of times the word statistical is used?

    And could I have that in a pretty graph form with lots of colors and little bitty words so my non statistical self can understand it all?
     
    resnor likes this.
  12. The Guy

    The Guy Well-Known Member

    6,598
    3,323
    113
    Oct 1, 2018
    Which is the one that makes Tannehill's 2019 something different from Nick Foles's 2013 for example?
     
  13. resnor

    resnor Derp Sherpa

    16,327
    9,874
    113
    Nov 25, 2007
    New Hampshire
    :theman:
     
  14. cbrad

    cbrad .

    10,659
    12,657
    113
    Dec 21, 2014
    Actually, statistical arguments have changed people's opinions quite a bit. Just in the last few pages I think you can see The Guy acknowledging certain stats change the interpretation of some of the stats he was posting.

    The minimum number required is 1 btw.
     
    The Guy likes this.
  15. cbrad

    cbrad .

    10,659
    12,657
    113
    Dec 21, 2014
    That's the one where you condition on the 11-game stretch because you're asking IF a QB had such stats what could you conclude about how the rest of his career went.
     
    The Guy likes this.
  16. JPPT1974

    JPPT1974 2022 Mother's Day and May Flowers!

    410
    84
    28
    Apr 15, 2012
    He found the right kind of team to suit his services there Ryan did. Good for him!
     
  17. The_Dark_Knight

    The_Dark_Knight Defender of the Truth

    11,815
    10,319
    113
    Nov 24, 2007
    Rockledge, FL
    Matt Moore was a fine back-up but did anyone truly think he was going to play as well or better than Tannehill? Cutler?
     
  18. AGuyNamedAlex

    AGuyNamedAlex Well-Known Member

    3,582
    2,579
    113
    Sep 12, 2015
    I thought he would be better than Cutler for the simple fact Cutler is a bad leader.
     
  19. The_Dark_Knight

    The_Dark_Knight Defender of the Truth

    11,815
    10,319
    113
    Nov 24, 2007
    Rockledge, FL
    Statistically speaking, it’s statistically unlikely that statistics statistically alter the opinion of statisticians who statistically rely on statistics to statistically opine on their statistical views of the game...

    Statistically speaking of course
     
    Puka-head likes this.
  20. The_Dark_Knight

    The_Dark_Knight Defender of the Truth

    11,815
    10,319
    113
    Nov 24, 2007
    Rockledge, FL
    I would agree that Moore would have been better than Cutler, but did anyone ever truly believe Moore would be as good or better than Tannehill? I don’t and didn’t think so.
     
    AGuyNamedAlex likes this.
  21. AGuyNamedAlex

    AGuyNamedAlex Well-Known Member

    3,582
    2,579
    113
    Sep 12, 2015
    I didnt think he would be better overall, but I will admit I thought his reckless style of play fit our chaotic team in the short term.

    That isnt so much a huge compliment to Moore or negative on Tannehill though as a reflection of our team at the time, which was pretty bad.
     
    The_Dark_Knight likes this.
  22. The_Dark_Knight

    The_Dark_Knight Defender of the Truth

    11,815
    10,319
    113
    Nov 24, 2007
    Rockledge, FL
    Seems to me the current New York Jets are actually the Miami Dolphins 2.0
     
    resnor likes this.
  23. Galant

    Galant Love - Unity - Sacrifice - Eternity

    19,127
    11,058
    113
    Apr 22, 2014
     
    The_Dark_Knight likes this.
  24. The Guy

    The Guy Well-Known Member

    6,598
    3,323
    113
    Oct 1, 2018
    There are very few backup QBs who anybody thinks are going to play better than the starters they reside behind.

    However, the 2016 season permits the only direct comparison of Moore and Tannehill's performance within the same surroundings.

    In the four games in which Moore played that season, his passer rating was 105.6, whereas Tannehill's previously that season was 93.5.

    So Moore certainly didn't play any worse than Tannehill, under conditions that permitted a direct comparison between them. @cbrad can tell us whether that difference was statistically significant in favor of Moore if he's willing.
     
  25. The_Dark_Knight

    The_Dark_Knight Defender of the Truth

    11,815
    10,319
    113
    Nov 24, 2007
    Rockledge, FL
    LMAO...

    Here we have someone staying that a 4 game stretch permits a direct comparison to Tannehill’s 12 game stretch.

    So when it suits, sample size matters/doesn’t matter depending on your personal bias.

    Too much!!!!
     
  26. The Guy

    The Guy Well-Known Member

    6,598
    3,323
    113
    Oct 1, 2018
    If you re-read the post, you'll see that the only conclusion I made regarding their samples of play that year was that Moore didn't play any worse than Tannehill under highly similar conditions, which is in direct contrast to your original point.

    What's actually "too much" is how often you throw something out there with a firm conclusion about it, when only a modicum of research would make your conclusion far more tentative. It's amazing how little time and effort it took to make your point above about Tannehill/Moore appear far more questionable.

    You did the same thing yesterday when you said you'd take the Titans this year over anybody but the Chiefs. The Titans, a team that's beaten three teams with a combined record of 1-8, by a margin of 2 points per game.

    Do a little research.
     
  27. resnor

    resnor Derp Sherpa

    16,327
    9,874
    113
    Nov 25, 2007
    New Hampshire
    Honestly, Moore is much like Fitz, would have a couple awesome games, then crash and burn. So using 4 games to compare to Tannehill's 12 seems like not a fair comparison.
     
  28. cbrad

    cbrad .

    10,659
    12,657
    113
    Dec 21, 2014
    Yeah there's a reason I keep saying wait until at least 150 attempts to even ask that question. Moore had only 87 attempts that season, and it's not really 4 games it's 3 games because that first game he had 5 passing attempts. Regardless, just to show you how difficult it is to achieve statistical significance with small sample size, if you compare Tannehill's 13 games in 2016 to Moore's 4 games, the probability those stats "come from the same QB" is 68.36% while if it's to the 3 games Moore started (leaving out that 5 attempt game) it's 58.92%. Both of which are WAY beyond the 5% threshold for significance.

    Here's a little exercise. How many passer rating points does one need to add to each one of Matt Moore's 3 games starting for his stretch to become statistically significant? Answer: 23!

    In other words, had Matt Moore had something closer to an average rating of 129 for those 87 attempts (technically: 82 attempts for the games he started) instead of a 105.6 rating THEN you can say that the particular stretch he had is statistically unlikely to have come from Tannehill that year. Just goes to show you why you don't want to compare ratings with small sample size. The difference needed for statistical significance is HUGE.

    Oh, and just to complete the thought, it's precisely because it's so hard to distinguish two sets of ratings with small sample size that one shouldn't compare ratings in such cases. Only in extremely rare cases (e.g., a bottom feeder over several games vs. sustained elite play over several games) might you even contemplate asking that question. As a rule of thumb, wait until a minimum of 150 passing attempts before comparing. And that's just a minimum. To detect smaller differences, like between "average" and "above average", you'll generally need a lot more (depends on variance).
     
    Last edited: Oct 2, 2020
    The Guy likes this.
  29. cbrad

    cbrad .

    10,659
    12,657
    113
    Dec 21, 2014
    lol. Don't tell that to @KeyFin

    Here's a trivia question:

    "Matt Moore is to __________ as Tannehill is to The Dark Knight"

    Answer: KeyFin!!!
     
    KeyFin likes this.
  30. KeyFin

    KeyFin Well-Known Member

    10,488
    12,821
    113
    Nov 1, 2009
    Matt Moore had a higher QB rating and a better TD/INT ratio than Tannehill every season he played! cBrad blew it off because it didn't meet the 150 attempts threshhold, but to me a 99+ rating is a 99+ rating until it goes up or down.

    In his prime MM > RT!
     
  31. AGuyNamedAlex

    AGuyNamedAlex Well-Known Member

    3,582
    2,579
    113
    Sep 12, 2015
    Why is it so difficult to comprehend that the same offenses and systems will produce different results for different QB?

    This assumption that two players playing in the same system means you can directly compare them is asinine.

    Michael Vick would have more success with the worst offensive line imaginable in football than Dan Marino would. Yeah, I said it.

    That in no way implies that as circumstances get better around the QB that their results will elevate at an equal rate.
     
    resnor likes this.
  32. cbrad

    cbrad .

    10,659
    12,657
    113
    Dec 21, 2014
    Gotta put that threshold somewhere. In 2015 Matt Moore had a whopping 118.7 rating!!! Thing is.. that rating is based on precisely 1 passing attempt lol. So I do think sample size matters?

    In any case, Moore didn't do that well every season he played in Miami, even if you were to ignore sample size.
    https://www.pro-football-reference.com/players/M/MoorMa01.htm

    btw.. I should point out every one of us can do better than Moore did in 2013. You don't have to even have played football. Moore was 2 of 6 for 53 yards with no TD's and 2 INT's giving him a 27.1 rating.

    If any one of us was QB and just went 0 for 6 for 0 yards, 0 TD's and 0 INT's that's a 39.58 rating, which would be higher than Moore lol (sorry had to do that Key!).
     
    KeyFin likes this.
  33. The Guy

    The Guy Well-Known Member

    6,598
    3,323
    113
    Oct 1, 2018
    Well and there you have the basis for the proposition that Tannehill's 2019 was a function of the optimization of his surroundings for him, and why Marcus Mariota's performance within the same system may not mean much with regard to a comparison with Tannehill. But you've had people here throughout the thread talking about how Mariota's performance in the same system means something in relation to Tannehill. Counsel these folks, would you?
     
  34. resnor

    resnor Derp Sherpa

    16,327
    9,874
    113
    Nov 25, 2007
    New Hampshire
    Talking about Tannehill vs Mariota, at least for me, was not about the numbers that each QB put up. It was about the numbers that HENRY put up. You were talking long ago about how Tannehill was successful because of Henry, and we were pointing out, that Henry was not nearly this dominant prior to Tannehill.
     
  35. The Guy

    The Guy Well-Known Member

    6,598
    3,323
    113
    Oct 1, 2018
    I'd be interested to see an analysis of that by @cbrad if he's willing, taking into account the the 2018 season and the games so far this season.
     
  36. resnor

    resnor Derp Sherpa

    16,327
    9,874
    113
    Nov 25, 2007
    New Hampshire
    This season has Tannehill, so how would that apply to Henry without Tannehill?
     
  37. The Guy

    The Guy Well-Known Member

    6,598
    3,323
    113
    Oct 1, 2018
    It would be 2018 and 2019 with Mariota, compared to 2019 and 2020 with Tannehill.
     
    resnor likes this.
  38. KeyFin

    KeyFin Well-Known Member

    10,488
    12,821
    113
    Nov 1, 2009
    Darn it, fans get to have selective memories and I'm telling you that Matt Moore was the best QB ever! Stop trying to ruin my Matt Moore parade ceremony!
     
    resnor and cbrad like this.
  39. cbrad

    cbrad .

    10,659
    12,657
    113
    Dec 21, 2014
    Wait, you guys want to know how Henry with Mariota compares statistically to Henry with Tannehill? I'll use Y/C.

    Not sure how you want to count that Denver game in 2019, as both QB's had a decent number of passing attempts (18 for Mariota and 16 for Tannehill), but if you don't count that game for either QB, then the p-value (probability the stats are from the same RB) is 0.2929 while if you count it for both the p-value is 0.3804. So anywhere between 29%-38% probability Henry's Y/C stats with Tannehill and with Mariota are from the same RB.

    So while there was a huge increase in Y/C for Henry a few games after Tannehill started, that lasted for only several games and isn't sufficient to say the Mariota vs. Tannehill record is beyond what you could explain from random variation alone.
     
    The Guy and KeyFin like this.
  40. The Guy

    The Guy Well-Known Member

    6,598
    3,323
    113
    Oct 1, 2018
    That's what I figured, and of course 2020 isn't helping the case for Tannehill's effect on Henry any. His number of yards per carry so far this year is under 4.
     

Share This Page