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. AGuyNamedAlex

    AGuyNamedAlex Well-Known Member

    3,582
    2,579
    113
    Sep 12, 2015
    Okay, we disagree that there is a ceiling. You cant convince me there isnt and I'm 100% sure I cant convince you there is, so I'll leave it at that with my answer to your statement.

    There has to be a ceiling. We can either believe it can be 100% accurate in dictating which players will be successful, using whatever statistics can be used to determine future success in which case the ceiling is 100%

    Or

    On the other hand if you DONT believe it can be 100% accurate no matter how much we measure the ceiling is whatever # you come up to represent accuracy.

    The reason I say eyes dont have a ceiling isnt accuracy. Its accuracy vs others. Since we are imperfect, you dont need a 100% success rate. There isnt really a recipe for finding a hidden gem of a player. Accuracy will vary among humans, which means you only need to be better than your peers to outperform them.

    How is this different from using statistics? If every team has the same hard data available and knows how to use it, the only possible conclusion is no-one is at an advantage or disadvantage.

    I agree with having a department to do those things. I just wouldnt make it the end all be all of every decision.

    Just to note, I'm using accuracy to describe how well it will help predict the future success of a player. In case I'm not clear. The statistics themselves are already "accurate" but not in the way I'm trying to convey.
     
    resnor likes this.
  2. resnor

    resnor Derp Sherpa

    16,327
    9,874
    113
    Nov 25, 2007
    New Hampshire
    There is no way to quantify things like motivation, so how can you say there is no ceiling?
     
  3. The Guy

    The Guy Well-Known Member

    6,598
    3,323
    113
    Oct 1, 2018
    ...as opposed to the ceiling of observational analysis, where Billy Joe Jim Bob says he saw the quarterback play one way, and Billy Ray John Bob says he saw him play the opposite way. At that point we’re stuck.
     
    cbrad likes this.
  4. cbrad

    cbrad .

    10,659
    12,657
    113
    Dec 21, 2014
    What I said is fact, not opinion. Whether you're convinced of it is irrelevant. However, so there's no miscommunication, note that I'm talking about there being no ceiling in practice, not in principle. There can be a ceiling in principle, but in practice there isn't because you keep improving on what you had. And what matters here is what occurs in practice not in principle.

    EDIT: here's one way to intuit this. Add up the numbers:
    (1/2) + (1/4) + (1/8) + (1/16) +...
    The Nth term there is 1/(2^N). No matter what number N we're talking about the sum of the first N terms is less than 1 so 1 is a theoretical "ceiling" (mathematicians call that a limit in this case). However, you can always go higher than you did by adding the N+1st term.

    Same thing with advances in science or statistical analysis. Every year statistical analysis keeps improving. There's no way to EVER reach some theoretical limit of "100% accuracy" because that would mean you have a perfect representation of the physical system, which is impossible in practice. So in practice you always improve.
     
    Last edited: Dec 17, 2019
    The Guy likes this.
  5. resnor

    resnor Derp Sherpa

    16,327
    9,874
    113
    Nov 25, 2007
    New Hampshire
    You have a limit in what stats can be measured. There is no way to quantify someone's ability to function under duress. You can't quantify someone's motivation, or someone's effort. We can say someone does well under pressure, or someone does poorly under pressure, but we can't measure it. We know that not all people perform the same under pressure, but we have no way of measuring it any more than that. That creates a limit for stats. I think we're taking about the ceiling for what stats can measure...not varying degrees of accuracy within the stats we are collecting, which it sense it's what you are saying.
     
  6. cbrad

    cbrad .

    10,659
    12,657
    113
    Dec 21, 2014
    Alex was explicitly talking about accuracy, not what can be measured per se, as you can see by going back to his posts.

    Regarding your argument about limitations in measurement, I agree that we currently cannot measure or infer a person's motivational state from tape, but there's no proof that such an advance could never be made. Who knows, maybe machine learning advances to the point it can discern subtle gestures from tape and map them onto motivational states, etc.. Not saying it will happen, but consider how impossible DNA testing must have sounded like to someone 100 years ago.

    Either way, the ability to measure things, even when things are defined by a latent variable (an unobservable variable) has measurably increased over time (e.g., 60 years ago no one knew how to transform subjective ratings into a valid measure). So in practice even the ability to measure things constantly improves even if there are known theoretical limits to it (e.g., from quantum mechanics).
     
    The Guy likes this.
  7. The Guy

    The Guy Well-Known Member

    6,598
    3,323
    113
    Oct 1, 2018
    Something interesting here in my opinion.

    Ryan Tannehill's number of yards per pass attempt this season is 4.09 standard deviations higher than the average number of yards per pass attempt of the teams with the top 100 season passer ratings since 2004.

    It's even 4.07 standard deviations higher than the average number of yards per pass attempt of the teams with the top fifty season passer ratings since 2004, as well as 2.74 standard deviations higher than the average number of yards per pass attempt of the teams with the top ten season passer ratings since 2004.

    I'd be interested to hear cbrad's perspective on the likelihood that will continue at anywhere near that level.
     
  8. resnor

    resnor Derp Sherpa

    16,327
    9,874
    113
    Nov 25, 2007
    New Hampshire
    For a computer to measure that, we have to program it.

    You can't measure how much effort someone is putting out. There are limits to what stats can do, man.
     
    Sceeto likes this.
  9. cbrad

    cbrad .

    10,659
    12,657
    113
    Dec 21, 2014
    The concept of z-score makes sense if the distribution is roughly normal. When you only pick the top 100 teams or the top 50 teams that distribution generally becomes highly skewed (I didn't double check your work and look at the distributions you mentioned, just going by experience here). So z-score might not mean what you think it means in this case. That's the first point I'd make.

    The second point I'd make is that z-scores are useful in telling you the probability an average QB would do something given random variation. They don't tell you what an above average QB would do. So what if Tannehill in Tennessee turns out to be a well above average QB on a consistent basis? Then those z-score based probabilities for "average QB" no longer apply and a different calculus is necessary.

    So I don't think your question can be answered because there's statistical evidence the conditions changed in Tennessee: his z-score for Y/A this year is a whopping 3.055 (using team Y/A distributions as a benchmark) which is top 0.11th percentile and clearly statistically significant. One can't just assume he'll regress. We just don't know. So I don't think we can make a prediction based on the stats in this case.
     
    KeyFin and The Guy like this.
  10. PhinFan1968

    PhinFan1968 To 2020, and BEYOND! Club Member

    Not that it's a game-changer or all that meaningful, but Tannehill was voted alternate to Pro-Bowl.
     
    KeyFin likes this.
  11. The Guy

    The Guy Well-Known Member

    6,598
    3,323
    113
    Oct 1, 2018
    What I’m getting at here is regression toward the mean. When the mean number of yards per pass attempt of the top 10 season passer ratings since 2004 is 2.74 standard deviations lower than a player’s current season number, don’t we have to anticipate regression toward at least that mean, if not toward the mean of the overall distribution?

    In other words, if we turned the top decile into its own distribution, Tannehill would be way out in the tail of even that!
     
  12. cbrad

    cbrad .

    10,659
    12,657
    113
    Dec 21, 2014
    Anyone whose z-score for any stat in a particular year is 3.055 is performing WAY above that person's average regardless of the situation, so from that perspective sure there will be regression to the mean. I mean.. Steve Young has the highest career z-score passer rating in history at 1.8627 so 3.055 is in no way sustainable (regardless of stat). The problem is that you have no idea what Tannehill's (possibly new) mean level is since there's evidence the conditions have changed in Tennessee. So there's no way to calculate probabilities here.


    That's not the main issue I was trying to highlight in the previous post however. One shouldn't calculate z-scores for tail ends of a normal distribution because the concept of z-score doesn't mean the same thing anymore. Let me illustrate with an example.

    Start with a standard normal distribution with mean 0 and standard deviation 1. Now.. suppose we take the equivalent of your "top 10" from this distribution. Since 10/32 teams = 31.25%, that's basically the same as taking all data from around z-score = 0.5 and above. What is the mean of that "top 10" portion (z-score 0.5 and above)? It's about 1.14. No particular issue with that.

    But what's the standard deviation of that "top 10" portion? It's 0.51 which is MUCH smaller than the standard deviation of the original distribution, which is 1. Why? Because that tail end of the distribution is highly skewed.

    Here's the problem. If you have a point at 3.055 (Tannehill's Y/A z-score) on the x-axis, that's obviously 3.055 z-scores above the mean of the original distribution. But guess what? That's (3.055 - 1.14)/(0.51) = 3.75 standard deviations away from the mean of the "top 10"!

    In other words, just by choosing the tail end of the distribution, you artificially get much larger "z-scores" because you're calculating them from a skewed distribution when they're not supposed to be. Can't do that.

    So yes there will be regression to the mean. No question with a 3.055 z-score. But what is that mean? Who knows. And you don't want to even intuit how likely something is in this case by looking at z-scores from just the "top 10" (tail ends of normal distributions).
     
    Last edited: Dec 18, 2019
    Irishman, Pauly and The Guy like this.
  13. Sceeto

    Sceeto Well-Known Member

    13,500
    6,244
    113
    Oct 13, 2008
    New York
    I really wish there was a separate forum for this statistical analysis nonsense. No offense to you, cbrad. I told you how I feel about you. However, this has gotten out of hand. It's really quite sad. It's almost every freakin' thread now and it always gets to the same place; nowhere with the exception of some very basic common sense conclusions which could have just as easily been reached without having to jump through these endless loops. It's almost impossible now to come here and have any normal conversations about the team and football in general. This is extreme even for baseball. I know about stats and I know this has gotten out of hand. With that said as with a post I made like 40 pages back, I don't expect things to change. I hope. "...and hopes a good thing, Red, perhaps the best of things". ..a little balance? :tt1::up:
     
    Last edited by a moderator: Dec 18, 2019
    Finatik, Mcduffie81, resnor and 3 others like this.
  14. cbrad

    cbrad .

    10,659
    12,657
    113
    Dec 21, 2014
    I personally don't mind if there are threads where no one is allowed to post stats lol. I'd have no issue with that, and I can carry on a conversation without any stats. But once I see too many misleading stats I do feel compelled to say something.

    Anyway.. if you're wondering how to "restore balance" I think the best approach isn't to wait for someone else to do something but to post non-stat arguments yourself. Of course you have to careful and make sure those arguments aren't easily rebutted with stats LOL.

    The other possibility is to just ignore all the stat stuff and not let it bother you, though based on your post I'm guessing that's not possible.
     
    Irishman and The Guy like this.
  15. adamprez2003

    adamprez2003 Senior Member

    37,392
    14,745
    113
    Nov 27, 2007
    new york ciity
    Exactly. Perfectly stated. Every thread is derailed by this statistical nonsense. It really needs it's own thread
     
    Last edited by a moderator: Dec 18, 2019
    resnor and Sceeto like this.
  16. Puka-head

    Puka-head My2nd Fav team:___vs Jets Club Member

    8,605
    6,743
    113
    Nov 25, 2007
    Slightly left of center
    I'm a little clueless here, but I'll just venture my opinion as an OG ( I think I can claim that???) on the forum.

    If one of the posters in this thread is Shouright,(like I said, clueless) the mans done his time and I would welcome him back to the forum. Been some a holes over the years I wouldn't say that about.

    Just my 2 cents
     
  17. Sceeto

    Sceeto Well-Known Member

    13,500
    6,244
    113
    Oct 13, 2008
    New York
    Hahaha! Oh boy. It's the bad, huh?

    ...and maybe I would ignore it if only it wasn't every freakin' thread.
     
  18. smahtaz

    smahtaz Pimpin Ain't Easy

    I'm glad Tannehill's playing well. It gives me a team to root for while mine is taking a do-over.
     
    Mcduffie81, Fin-O, Irishman and 3 others like this.
  19. KeyFin

    KeyFin Well-Known Member

    10,488
    12,821
    113
    Nov 1, 2009
    What this all boils down to isn't statistics...it's a select few people telling others why Tannehill will inevitably fail. And as others have said countless times, if it happens then it happens. But in the meantime, it would be awesome if a select few can stop COMPLETELY IGNORING THE PRESENT and stop acting so childish because folks here are happy about our 7-year QB doing well.

    It is RIDICULOUS that there are over 400 posts in this thread alone telling us what RT did this year doesn't matter because it's a fluke. But even worse, we keep getting these outlying, cherry-picked statistics that have nothing to do with this season at all...that's why folks here have their pitchforks in hand over "statistics".

    The stats aren't the problem- the endless trolling is what has others so angry. And I think because you're our resident math guy, you've unfairly taken most of the heat for what others have incessantly posted over and over and over again.

    It's just Dirty...like a receiver named Landry.
     
  20. cbrad

    cbrad .

    10,659
    12,657
    113
    Dec 21, 2014
    At least half the threads here don't include posts with statistical analysis.

    For example, the "Florez blows up" thread, the "Deiter demoted" thread, the "Chase Young returning for his senior year" thread, the "Contract extension for Parker" thread, the "Who are we rooting for these last 3 games" thread.. all don't involve statistical analysis.

    So it isn't anywhere close to "every freakin' thread".

    Anyway, if you want more posts of the type you like, post more yourself. It's just a suggestion.
     
    Last edited: Dec 18, 2019
    Irishman, The Guy and Pauly like this.
  21. cbrad

    cbrad .

    10,659
    12,657
    113
    Dec 21, 2014
    Yeah I have no suggestions for remedying that except to block the poster(s) you don't like. Otherwise you'll just have to take it in stride lol.
     
    Fin-O, Irishman and The Guy like this.
  22. Hoops

    Hoops Well-Known Member

    1,183
    1,484
    113
    Dec 11, 2016
    Tannehills int wasn’t even remotely on tannehill. That’s a called in from the sideline tight red predetermined primary throw where as long as the qb places it where he should he’s done his job.

    Watson’s tight red int was completely on the qb.

    There’s the difference. And you aren’t gonna get that from z scores and statistical analysis.

    that said I didn’t think tannehill played a great game. There was some catchable balls that guys didn’t come down with but there was a bad misread where he read man and got zone to the tight end which romo made too big a deal about “getting away with it” given the db barely got a hand on it but still a bad read none the less.

    There was also one sack that I’d say situationally tannehill can’t take but there was another to end the game where he had zero chance as the right guard got destroyed. climb the pocket I can hear the Miami fans say if it was in Miami... yeah climb what freaking pocket.
     
  23. Sceeto

    Sceeto Well-Known Member

    13,500
    6,244
    113
    Oct 13, 2008
    New York
    Yep, or it's own forum.
     
    Last edited by a moderator: Dec 18, 2019
    adamprez2003 and resnor like this.
  24. Sceeto

    Sceeto Well-Known Member

    13,500
    6,244
    113
    Oct 13, 2008
    New York
    I'm glad you got the point.
     
  25. Hoops

    Hoops Well-Known Member

    1,183
    1,484
    113
    Dec 11, 2016
    I will just add one more thing. These are 3 damn good teams that the titans are playing to close the season. Texans twice and the saints. 2 and 4 start means there’s no real margin for error.

    so let’s keep that in mind too with regards to what happens these next few weeks. One thing I’m pretty certain of is if/when the titans lose it’s gonna be about more than just the qb play.

    that said that was a golden opportunity at home and the titans whiffed on it.
     
  26. cbrad

    cbrad .

    10,659
    12,657
    113
    Dec 21, 2014
    It's really the Texans game to end the season that matters for the Titans making the playoffs. If the Titans beat the Texans and the Steelers lose to either the Jets or Ravens, then the Titans claim the final WC spot. The Saints game is only important if the goal is to win the AFC South, but that means the Titans have to win out and Houston lose out. Very unlikely.

    Anyway, Titans have about a 50% chance of making the playoffs as of right now by most game simulators (which aren't perfectly accurate of course but they give you some idea of the chances).
     
    Irishman and Hoops like this.
  27. Hoops

    Hoops Well-Known Member

    1,183
    1,484
    113
    Dec 11, 2016

    What’s the guarantee the ravens will be playing for something still come week 17?
     
  28. cbrad

    cbrad .

    10,659
    12,657
    113
    Dec 21, 2014
    No guarantee. If the Ravens beat the Browns they clinch 1st seed. But then again that's a pretty solid team and you can't rest too many players. Steelers don't have a QB so not having Jackson might even things out.

    Regardless, what I was trying to point out is that the Texans game is the one that really matters because it's a conference game, which is important for WC tiebreakers. Beating the Saints and losing to the Texans with a Pittsburgh win leaves the Titans out of the playoffs.
     
    Hoops likes this.
  29. Hoops

    Hoops Well-Known Member

    1,183
    1,484
    113
    Dec 11, 2016

    I was hoping you wouldn’t say a win clinches the 1 seed. Just like I’m hoping you won’t say a win for the pats clinches the 2.

    any chance you see the Texans if they win Saturday sitting guys week 17 vs the titans? Mainly watson and Hopkins lol

    I’d imagine they’d be locked into their 4 seed regardless assuming kc wins vs Chicago.
     
  30. cbrad

    cbrad .

    10,659
    12,657
    113
    Dec 21, 2014
    Patriots do NOT clinch #2 seed with a win. If KC wins out and Patriots lose a game KC gets the #2 seed because of head-to-head record (KC beat the Pats 23-16).

    If Houston wins out and KC loses one Houston takes the #3 seed. Same if Houston wins one and KC loses out. They beat KC earlier 31-24.
     
    Irishman and Hoops like this.
  31. AGuyNamedAlex

    AGuyNamedAlex Well-Known Member

    3,582
    2,579
    113
    Sep 12, 2015
    I'm officially done with this thread.

    I'm all for any discussion involving reason, but when someone tells me I'm wrong, all while making irrelevant points and not addressing any of my own?

    You're denying a a hard wired truth of the universe in your reply and violating scientific reasoning/logic. I'll let you figure out how.

    Blocked. I'd encourage you all to do the same honestly. If someone is incapable of having a conversation I dont see why we continue to try and give the benefit of the doubt.
     
    Last edited: Dec 18, 2019
    Phins_to_Win and resnor like this.
  32. The Guy

    The Guy Well-Known Member

    6,598
    3,323
    113
    Oct 1, 2018
    I would like to make my position on the above clear.

    My position is not that Tannehill's performance is a fluke. It's that his performance tells us nothing yet about whether he has the ability to be one of the best quarterbacks in the league on a consistent basis. We need more time to make that determination.

    And that could go either way. He could continue what he's doing over the necessary time period and be deemed one of the league's best QBs, or he could not. I don't have a crystal ball in that regard. I'm asserting only that we need more time to make that determination.

    So, if it's me you're including in the group who, as you said in the post quoted above, is "telling others why Tannehill will inevitably fail," or "telling us what RT did this year doesn't matter because it's a fluke," I hope this post clarifies why that's incorrect. You won't find a single post in this thread in which I've said his performance is a fluke or that he will inevitably fail. So, if you have a pitchfork in hand about that, you've picked it up mistakenly, and perhaps you should examine your own sensitivities.


    Another thing I'd like to make clear here is that if there are posters whose positions about things give others feelings they would rather not have, as long as those posters are following the rules of the forum, it's the responsibility of the people with the feelings to own them and address them themselves. Don't suggest that others should take responsibility for your feelings and change their behavior so the world can be a better place for you.

    Many times the way this occurs is by way of suggesting that the person generating the feelings for you is "trolling." There is nobody here, myself included, who knows the true intent of anybody else and can therefore determine with certainty whether someone else is "trolling." Far more often than not, the word "trolling" is used to indicate that someone else's behavior is responsible for what you feel.

    What I'd suggest is that if you believe someone is trolling, whether they truly are or not, the way to manage that is not by openly accusing the person of trolling. Address that by 1) communicating with the person backchannel and politely indicating your concerns, 2) if that doesn't work, report the individual to a moderator, and 3) if that doesn't work, block him.

    There are appropriate remedies here for your problem. Note the italics on the word "your."

    Regardless of anyone's responses, this will be my last post in this thread about people's conduct in the forum. I'm happy to debate about the Miami Dolphins, but I won't be debating about how people on a message board should conduct themselves.
     
    djphinfan likes this.
  33. AGuyNamedAlex

    AGuyNamedAlex Well-Known Member

    3,582
    2,579
    113
    Sep 12, 2015
    I'm just personally not going to keep trying to unite both sides in the middle if one end is going to have a hissy fit and act personally assaulted every time we disagree. This doesnt go at anyone in particular, just the thread in general.

    I offered plenty of opportunities for everyone to take reasonable positions and come to an understanding in the middle. If no-one wants to do that, on either side, we are all wasting our breath in this thread saying the same things on repeat.
     
    djphinfan likes this.
  34. FinFaninBuffalo

    FinFaninBuffalo Well-Known Member

    2,474
    2,954
    113
    Dec 13, 2007
    upload_2019-12-18_9-45-15.png
     
    resnor likes this.
  35. FinFaninBuffalo

    FinFaninBuffalo Well-Known Member

    2,474
    2,954
    113
    Dec 13, 2007
    upload_2019-12-18_9-45-54.png
     
    Irishman and resnor like this.
  36. Hoops

    Hoops Well-Known Member

    1,183
    1,484
    113
    Dec 11, 2016
    Irishman likes this.
  37. The Guy

    The Guy Well-Known Member

    6,598
    3,323
    113
    Oct 1, 2018
  38. cbrad

    cbrad .

    10,659
    12,657
    113
    Dec 21, 2014
    Most obvious problem with that estimate is that they don't take into account salary cap inflation. The NFL salary cap has risen by at least $10 million in each of the last 6 years, so 3 seasons difference might lead to 15-20% difference in estimates (more realistic estimates would be higher for future signees).

    The other issue with that analysis is they're doing regression (fitting a straight line) on only 4 data points. That's not very reliable.

    I would have preferred taking a larger number of QB contracts (20+) and adjusting for both age and inflation and then fitting a line. That probably would give you a more realistic estimate.
     
    Pauly and The Guy like this.
  39. Hoops

    Hoops Well-Known Member

    1,183
    1,484
    113
    Dec 11, 2016
    Am I wrong or did Kirk cousins not get $87 mil over 3 years fully guaranteed?

    didn’t foles get similar? Why would tannehill be expected to play for less with cap increase and the real time market rate?

    are they thinking not hitting the market team discount or something?
     
    Pauly, resnor and Dol-Fan Dupree like this.
  40. Finatik

    Finatik Season Ticket Holder Staff Member Club Member

    4,323
    4,012
    113
    May 2, 2014
    SO Cal
    What he said.
     
    Last edited by a moderator: Dec 18, 2019
    adamprez2003 and Sceeto like this.

Share This Page