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
    They're not even close. LOL.
     
  2. The Guy

    The Guy Well-Known Member

    6,598
    3,323
    113
    Oct 1, 2018
    Here's another interesting game in which there was a big discrepancy between traditional passer rating and QBR, and QBR did the far better job of predicting the winner:

    https://www.espn.com/nfl/boxscore?gameId=401038850

    You can see there Tom Brady had a traditional passer rating of 77.1 and a QBR of 82.6, while Patrick Mahomes had a traditional passer rating of 117 and a QBR of 74.8. The Patriots won the game (the AFC Championship game) in overtime 37-31.

    So, obviously one game (or two) can't determine the validity of QBR, but these sorts of instances warrant greater exploration of whether QBR is more strongly related with winning than traditional passer rating.
     
  3. Cashvillesent

    Cashvillesent A female Tannehill fan

    770
    641
    93
    Dec 8, 2019
    Brady was a game managing QB for only his first two rings....

    Mahomes didnt impress much in his first SB appearance, lets see what he does in his second.
     
  4. cbrad

    cbrad .

    10,659
    12,657
    113
    Dec 21, 2014
    Unbelievable. You're still into black box methods AND you're still into cherry picking.

    Neither of which you should be doing if you want to argue statistics is useful. Not sure why you don't get that. Don't you understand that "black box" and "cherry picking" allows anyone to make ANY argument they want? I mean.. all someone arguing against you needs to do is to do the same thing you're doing and they can make the exact opposite argument.

    For example, why not cherry pick the Denver-Carolina SB (2015 season) where Peyton's QBR was WAY worse (8.6) than Newton's (31.6) but Denver won. Because it doesn't support your argument? I think passer rating captures that one better: both QB's were bad but one wasn't historically bad (passer ratings: Peyton = 56.6, Newton = 55.4).
    https://www.espn.com/nfl/boxscore?gameId=400820438

    This is a useless exercise and lacks credibility. AND it hurts any argument for using statistics! You should care about that.

    Also, just so it's clearly understood. That "or two" games you inserted in parenthesis is misleading. Your calculation of z-scores for QBR was wrong for the 2019 SB and it turns out that the discrepancy in z-scores was smaller for QBR for the 2019 SB. In other words your entire (cherry picked) SB example actually argued in favor of passer rating!

    I've said this before: don't use stats like passer rating for single games. Use them when sample size is large enough.
     
    Pauly, danmarino and FinFaninBuffalo like this.
  5. The Guy

    The Guy Well-Known Member

    6,598
    3,323
    113
    Oct 1, 2018
    All I'm saying here is that a single-game discrepancy between traditional passer rating and QBR warrants greater exploration of whether QBR is more strongly related to winning. If I had already concluded QBR is more strongly related to winning, I'd be cherry-picking. The Denver-Carolina game should of course be included in that exploration of correlation -- I'm certainly not saying we should exclude the games that don't indicate QBR is more strongly related to winning.

    My goal with this is to take a look at the relationship between QBR differential and win percentage, see how it compares to the relationship between traditional passer rating differential and win percentage, and see if that makes us any more "comfortable" with the parts of QBR we know aren't transparent. Maybe it would, and maybe it wouldn't. That to me seems like a decent enterprise that isn't inconsistent at all with science.
     
  6. cbrad

    cbrad .

    10,659
    12,657
    113
    Dec 21, 2014
    If your goal was to look at the correlation between QBR and win% then say so, don't just pick one game to say "hey look this was a game where some proprietary rating system did better". Because you can find at least one game for practically ANY rating system where it "did better". That IS cherry picking regardless of whether you concluded anything prior to that cherry picking or not.

    So what's the correlation of QBR to win%? Who knows, but this link says 0.68:
    https://www.footballperspective.com/is-espns-qbr-the-best-measure-of-quarterback-play/

    That's exactly the same as passer rating once it's adjusted for era. Keep in mind that the guy who wrote that article doesn't understand adjustments for era with passer rating so I hope he did the calculations correctly, but since QBR is in percentiles it's already "adjusted" so assuming the calculations are correct there's no difference. Also note that 0.68 for passer rating is for team passer rating and there's no comparable "team QBR" so the comparison isn't one-to-one.

    Regardless, the problem here is your willingness to consider methods that would allow anyone to make any argument. You have got to understand that "black box" methods allow anyone to claim anything. So there's no reason to care about QBR, yet you still do.
     
    danmarino and FinFaninBuffalo like this.
  7. The Guy

    The Guy Well-Known Member

    6,598
    3,323
    113
    Oct 1, 2018
    I thought I made that clear in my original post about this (the highlighted portion below)?
    The position I'm taking with this is no different from the one the author of the article above puts in his conclusion:
     
  8. cbrad

    cbrad .

    10,659
    12,657
    113
    Dec 21, 2014
    Like I said no matter how crappy the rating system you can find some game where that rating system did better. So pointing out one cherry picked example is NOT how you go about investigating things. It’s why sample size matters.

    And there is no trade-off one should consider when it comes to lack of transparency because, like I said, lack of transparency allows you to support any argument. Apparently this isn’t (yet) important for you.

    Not much else I can say but you’re harming the credibility of statistical analysis with your approach. Not going to go in circles on this but this is bad.
     
    FinFaninBuffalo likes this.
  9. FinFaninBuffalo

    FinFaninBuffalo Well-Known Member

    2,474
    2,954
    113
    Dec 13, 2007
    I tried to warn you......
     
    Irishman, resnor and cbrad like this.
  10. The Guy

    The Guy Well-Known Member

    6,598
    3,323
    113
    Oct 1, 2018
    Once again, I made that same point originally myself:
    Do I really seem to you like someone who believes that the way to properly investigate a comparison between QBR and traditional passer rating is to simply nominate a single game in which QBR seemed to outperform traditional passer rating and make a conclusion about the two measures solely on the basis of it?

    Please be honest. Is that how I seem to you? As many times as I've used correlation here and pointed out to people the concept of "the exception to the rule," you really figure I'm going to rest on one game as some kind of indication of the superiority of a measure?
     
  11. cbrad

    cbrad .

    10,659
    12,657
    113
    Dec 21, 2014
    Yeah let's be honest. Did you or did you not look at multiple games to see if you could find one game where QBR looked like it was the better predictor? If the answer to that question is "yes" then that's cherry picking because you ONLY showed us that one game!

    What you should have done is show us a whole bunch of games and come to some conclusion. But that's not what you did.

    And I'll say this again: "black box" is much worse than "cherry picking". So the willingness to consider "black box" absolutely needs to go.
     
    Irishman and FinFaninBuffalo like this.
  12. The Guy

    The Guy Well-Known Member

    6,598
    3,323
    113
    Oct 1, 2018
    That's not what I did because I'm all for doing the full-scale analysis with hundreds upon hundreds of games (when I have the time...), and I've done that kind of analysis here in the past, so I would think I'd receive the benefit of the doubt with regard to the position I was taking on this.

    I have to think if this was a message board of statisticians you wouldn't be so sensitive to this, because we'd both know what we're talking about and we'd know everybody else does too. I suspect you're trying to "protect" the people here who don't know any better, but we don't have to villainize me to do that. We can simply give them more or clearer information.
     
  13. cbrad

    cbrad .

    10,659
    12,657
    113
    Dec 21, 2014
    I don't believe you. You mean to tell me that the FIRST and ONLY game you looked at was the Patriots vs. Chiefs AFC championship game you used as an example? No way. Just admit it. You looked FOR a game to support the view that QBR was better than passer rating in a specific case. That's cherry picking dude.

    Just the opposite. If this was a message board for statisticians not only I but others would be livid. Right now I'm just annoyed.
     
  14. The Guy

    The Guy Well-Known Member

    6,598
    3,323
    113
    Oct 1, 2018
    Why does any of that matter when my "conclusion" on the basis of that game was this?
    OK, so perhaps I know little about the culture of a message board of statisticians.
     
  15. cbrad

    cbrad .

    10,659
    12,657
    113
    Dec 21, 2014
    Once again.. there's no reason at all to even post such an example because there's always such an example. So what you need to do is just look at a whole bunch of games and see if there's a trend. That would not have elicited any "cherry picking" issue. The "black box" issue, which is far worse of course, is still present.

    Just modify your approach in the future. And let's not give any credence to black box methods.
     
    Irishman, Pauly and FinFaninBuffalo like this.
  16. resnor

    resnor Derp Sherpa

    16,327
    9,874
    113
    Nov 25, 2007
    New Hampshire
    But then his argument disappears.
     
    Irishman, FinFaninBuffalo and cbrad like this.
  17. danmarino

    danmarino Tua is H1M! Club Member

    15,357
    20,976
    113
    Sep 4, 2014
    Did you bother to look at Brady during any other post season games where he played great defenses?

    Like last year's Super Bowl. They won, but he posted a 71 passer rating against the 20th ranked defense. The week before, against KC, he had a 77 against the 24th ranked defense. Against TN this season (12th ranked defense) Brady posted a 59.

    Against Houston in 2016, he posted a 68 against the 11th ranked defense.

    In 2016, against Denver, he posted a 56 passer rating.

    In 2013, against the Ravens, 62.

    Baltimore, 2012, 57.

    Baltimore, 2009, 49.

    Chargers, 2007, 66.

    Chargers and Colts, 2006, 52 and 61, respectively.
     
    Irishman and resnor like this.
  18. danmarino

    danmarino Tua is H1M! Club Member

    15,357
    20,976
    113
    Sep 4, 2014
    And Brady still threw 2 INT's in that game.
     
    Irishman and resnor like this.
  19. KeyFin

    KeyFin Well-Known Member

    10,488
    12,821
    113
    Nov 1, 2009
    There's an easier statistical formula to fix his argument. If you hit the "ignore" button, you're guaranteed not to hear his twisted cherry-picking analysis 100% of the time. And if that happens, this thread can actually go back to discussing football.
     
    Irishman, Dolphin Dundee and resnor like this.
  20. Cashvillesent

    Cashvillesent A female Tannehill fan

    770
    641
    93
    Dec 8, 2019
    Did you bother to look up on the majority of the SBs performance hes had?
     
  21. The Guy

    The Guy Well-Known Member

    6,598
    3,323
    113
    Oct 1, 2018
    Indeed there is always such an example, but in many cases we already know it's an exception to the rule, whereas in this case we don't!

    That's why it's perfectly fine to post that example as a cause for further exploration of the subject, as long as the caveat I gave (and reiterated several times) is included. If we already knew it was an exception to the rule, then posting the example would be misleading.
     
  22. resnor

    resnor Derp Sherpa

    16,327
    9,874
    113
    Nov 25, 2007
    New Hampshire
    That's cherry picking. Look at all his post season games, don't pull out one that supports your point.
     
    Irishman and danmarino like this.
  23. cbrad

    cbrad .

    10,659
    12,657
    113
    Dec 21, 2014
    We also know QBR is a black box method. So no reason to post any example of its supposed superiority.
     
    Irishman and resnor like this.
  24. Cashvillesent

    Cashvillesent A female Tannehill fan

    770
    641
    93
    Dec 8, 2019
    It isnt cherry picking. Alot of great QBs get judged on big games. Why do you think Manning and Favre arent looked as goats?
     
  25. The Guy

    The Guy Well-Known Member

    6,598
    3,323
    113
    Oct 1, 2018
  26. danmarino

    danmarino Tua is H1M! Club Member

    15,357
    20,976
    113
    Sep 4, 2014

    lol...I guess you don't realize that Mahomes' post season passer rating is 106 and Brady's post season passer rating is 89?

    Last year Mahomes had a 117 passer rating against the Pats while Brady had a 77. lol...
     
    Irishman and resnor like this.
  27. danmarino

    danmarino Tua is H1M! Club Member

    15,357
    20,976
    113
    Sep 4, 2014
    What you're doing is the definition of cherry picking.

    Judging QB's by "big games" is lazy... at best.

    Did you really just claim that Manning isn't looked at as a GOAT? Bwhahahahahah!!
     
    Irishman, Phin McCool and resnor like this.
  28. cbrad

    cbrad .

    10,659
    12,657
    113
    Dec 21, 2014
    Brian Burke joined ESPN and you think he's the type of person that would attack QBR? I've already shown errors in some of his articles so it's not like the guy knows statistical analysis that well (sometimes he does good work though).

    In any case, let it be clear that the view you and Burke espouse is precisely the opposite of what is considered good science. As long as people here don't confuse your view with "valid statistical analysis" we can move on.
     
    Irishman and danmarino like this.
  29. cbrad

    cbrad .

    10,659
    12,657
    113
    Dec 21, 2014
    When you adjust ratings to 2019 and take the weighted average (by passing attempts), Brady's postseason rating is 96.3 so a bit higher than unadjusted but still less than Mahomes, though of course Mahomes hasn't had anywhere near the same number of passing attempts.

    One more thing to consider: on average playoff passer rating is about 10 points lower than in the regular season. So while 96.3 doesn't look impressive relative to the 2019 league average of 90.8 it's VERY impressive compared to the adjusted playoff average of 81. And Mahomes' 105.5 about 25 passer rating points higher than playoff average!! (at least so far)

    Mahomes is absolutely impressive in the playoffs. Maybe the 1st 3 quarters in the SB colored some opinions, but that was against one of the best defenses. And the ability to still pull off a comeback win (3 times in the playoffs!) makes him clutch. Best QB in the NFL right now.
     
  30. resnor

    resnor Derp Sherpa

    16,327
    9,874
    113
    Nov 25, 2007
    New Hampshire
    So because some people judge things wrongly, means everyone should? Who doesn't view Making as the arguable GOAT? Favre was great.

    But, yes, it is cherry picking to take one game, and use it to represent what that QB does in big games, as if that's the level he always plays at. The media contributes to this problem by turning certain well liked players into heroes, attributing every success to the QB, while ignoring the rest of the team.
     
    Irishman likes this.
  31. resnor

    resnor Derp Sherpa

    16,327
    9,874
    113
    Nov 25, 2007
    New Hampshire
    You can't adjust all the ratings, because many of those ratings were earned in the same era.
     
  32. cbrad

    cbrad .

    10,659
    12,657
    113
    Dec 21, 2014
    It's still more accurate to adjust by each year's passer rating distributions (league average and standard deviation.. or as an approximation just the league average). If you let's say adjust 2018 ratings to 2019 ratings you only get a 2.5 rating point difference so it's not much. Either way it's better to use a principled approach than manually deciding where to place thresholds for adjustment.
     
    Irishman likes this.
  33. Cashvillesent

    Cashvillesent A female Tannehill fan

    770
    641
    93
    Dec 8, 2019
    No he isnt. He is known as the best regular season QB, and a playoff choker.
     
  34. cbrad

    cbrad .

    10,659
    12,657
    113
    Dec 21, 2014
    This is one case where the statistics say otherwise. The 2019 adjusted playoff rating for Brady is 96.3 while for Peyton Manning it's 95.4. Virtually identical, and I think you'll see that borne out when you go through them game by game.

    The guy who is REALLY impressive in the postseason is Joe Montana. Adjusted to 2019 his playoff rating is 114.5.
     
    Irishman and danmarino like this.
  35. The Guy

    The Guy Well-Known Member

    6,598
    3,323
    113
    Oct 1, 2018
    Well now hold on a minute -- just because you and I have a difference of opinion about QBR doesn't mean valid statistical analysis can't be applied to the question of whether it correlates more strongly than traditional passer rating with win percentage. Valid statistical analysis can certainly still be applied to that issue, regardless of what you think about QBR.
     
  36. resnor

    resnor Derp Sherpa

    16,327
    9,874
    113
    Nov 25, 2007
    New Hampshire
    The first argument for adjusting was due to rules changes. Fine. I can get on board with that. Adjusting every year to the current year is ridiculous. Think of baseball and batting .400. No one would be ok with adjusting stats around to really determine who batted .400. No, you just go look up the batting averages each year.
     
  37. cbrad

    cbrad .

    10,659
    12,657
    113
    Dec 21, 2014
    That's not what I'm saying. I'm saying that every valid statistical approach is transparent – we know precisely what the assumptions are and how things are calculated. There's no exception. So the fact you're even willing to consider QBR is what makes your approach "invalid" from a purely statistical point of view.
     
    Irishman likes this.
  38. cbrad

    cbrad .

    10,659
    12,657
    113
    Dec 21, 2014
    There's no inflation in batting averages over time. Batting averages over time are consistent with random variation around the same mean (the average from 1871-2019 is 0.263). That's why you can just look them up.

    Passer rating is different. In 1970 league average was close to 66, in 1980 it was close to 74, in 2000 it was close to 78 and in 2019 it's close to 90. That is NOT consistent with random variation around the same mean. And the rule changes and/or changes in application of those rules has been steady from 1978 onwards. It's a gradual increase, not sudden. So yes you need to adjust year by year.
     
    Irishman likes this.
  39. resnor

    resnor Derp Sherpa

    16,327
    9,874
    113
    Nov 25, 2007
    New Hampshire
    I disagree with much of what you say in regards to baseball, but I'll leave it at that.

    No, I do not agree that you should adjust ratings post probably 2007.
     
  40. cbrad

    cbrad .

    10,659
    12,657
    113
    Dec 21, 2014
    This is actually an example of the type of discussion that we could continue in that statistical methods thread.. if you want of course.. because discussions about when you don’t need to adjust ratings would be useful.

    I’d also suggest to The Guy that the QBR discussion be moved there.

    In any case, I did look at the batting averages from 1871 before making those statements.
     
    Irishman likes this.

Share This Page