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

Comparing Tannehill to Luck,

Discussion in 'Miami Dolphins Forum' started by Pauly, May 8, 2016.

  1. vt_dolfan

    vt_dolfan Season Ticket Holder Club Member

    Just wanted to point out... Tannehill basically played Luck to a draw while being shackled to Joe Philbin. Keep that in mind.


    Either way.. What matters is how he plays this season... And most have cautious optimism. If Gase is as good a coach as what we think... Tannehill should put these comparisons to rest.

    If he doesn't... We need to figure out how to get Patrick Mahomes.
     
  2. vt_dolfan

    vt_dolfan Season Ticket Holder Club Member

    Man... So awsome we have guys who know how to gather data... And then know how to use it. Thanks man.
     
  3. vt_dolfan

    vt_dolfan Season Ticket Holder Club Member

    I wonder if we could tell his avg down and distance during that time?

    When your down with less then 4 mins to go... We get in trouble when 17 had to convert when faced with more then 7 yards... Usually because our protection wouldn't hold up long enough to give him time.

    Now.. Not saying he wasn't at fault at all... It's just a factor to be considered
     
    resnor likes this.
  4. resnor

    resnor Derp Sherpa

    13,341
    7,555
    113
    Nov 25, 2007
    New Hampshire
    Yeah, everyone likes to discount the down and distance, as if all third downs are the same. I haven't seen actual data on it, but from watching games last season, we routinely were in third and 7 or more, especially in fourth quarters.
     
  5. adamprez2003

    adamprez2003 Senior Member

    37,402
    14,758
    113
    Nov 27, 2007
    new york ciity
    The two quarterbacks were developed very differently. Luck was told to just sling it and tannehill was basically handcuffed to a more disciplined and avoid mistakes approach. What that has resulted in is luck is an interception machine right now while tannehill who developed slower is a more rounded and smart QB. If you went into week one I would say tannehill is better but knowing how good luck is and knowing he has had a full offseason knowing he has to cut down on the turnovers I would bet that luck once again jumps ahead of tannehill. Luck has been groomed to be a QB from day one. It would surprise me if he can't fix his int problem
     
    cbrad and resnor like this.
  6. roy_miami

    roy_miami Well-Known Member

    1,385
    560
    113
    Oct 11, 2013
    For my money the 4 best coaches besides Belichick of late are Bruce Ariens, Jim Harbaugh, Marvin Lewis and John Harbaugh.

    Jim Harbaugh's record in 0-7 point games: 21-9 (10-4 with Smith)
    Alex Smith in his two previous seasons prior to Harbaugh: 1-9
    Alex Smith in his last two seasons with Reid: 7-7

    Bruce Ariens record in 0-7 point games: 13-4 in Arizona and 5-1 as interim with Luck in 2013

    Marvin Lewis with Palmer: 20-26
    Marvin Lewis with Dalton: 22-13

    John Harbaugh with Flacco: 31-33

    Palmer w/ Ariens: 11-4
    Palmer w/ Hue Jackson: 4-1
    Palmer w/ Dennis Allen: 2-2
    Palmer w/ Marvin Lewis: 20-26

    I don't know what conclusions can be drawn based on this. Am I wrong about John Harbaugh being one of the best coaches and maybe Flacco doesn't get the credit he deserves? How do you explain Marvin Lewis' record with the two different QBs? Luck is 5-1 with Ariens as coach but is 15-6 with Pagano which is still pretty good relative to the league.

    Caldwell was 12-4 in 0-7 point games with Manning and 10-8 wihout him, but he's 10-5 with Stafford who is 23-25 overall.

    So you may be able to look at this stat and say it shows Ariens is an elite coach, but elite coaches are so rare and obvious that I don't know how useful it is. And on top of that the stat doesn't give the results you would expect for coaches such as John Harbaugh and Marvin Lewis. Marvin Lewis is the most interesting study, he is the second most tenured coach in the league and as the stats show has looked bad to elite for periods over that time. Is it vaiance? Is it simply having the right fit between QB and coach? Just plain ole' experiance?
     
    Pauly, cbrad and Finster like this.
  7. Thunderkyss

    Thunderkyss Member

    53
    31
    18
    Sep 10, 2012
    Completely agree... well, almost.

    We're at that point. He's got four seasons behind him. He's on his second contract. I know he's had a slew of different coordinators & what not, but it's time for him to reach that level between player & management that QBs need to take, as leaders of the team.

    His OC needs to see him as an equal, a partner & that has more to do with how Tannehill carries himself than anything else. I'm not saying he should be telling the OC how things are going to be, but he & the OC need to be together on how they plan their offense.
     
    dolphin25 likes this.
  8. dolphin25

    dolphin25 Well-Known Member

    4,037
    1,101
    113
    Nov 22, 2014
    Marvin Lewis was hated in Cincinnati, he had some not so good teams for a long time, only the past few year (Dalton) have they started be really strong.
     
  9. cbrad

    cbrad . Club Member

    7,521
    8,912
    113
    Dec 21, 2014
    Good data!

    However.. there's one outlier in that group: Roethlisberger. Take away that outlier and the correlation between win % and <4 minute trailing rating is 0.43, which is really good in football. I mean that's as good as YPA, which is known to be a good predictor: http://www.footballperspective.com/correlating-passing-stats-with-wins/

    And Roethlisberger is really an outlier given his career rating of 76.1 in <4 minute trailing stats. In fact, I hope your data is accurate because that seems really odd for him to jump up that high. Can you double check Roethlisberger again?
     
    Pauly likes this.
  10. cbrad

    cbrad . Club Member

    7,521
    8,912
    113
    Dec 21, 2014
    Your argument holds only if the final win/loss records are equal and if passer rating captures all aspects of QB performance. Since Luck has a far better W/L record in close games, we know there is something in the W/L record the passer rating isn't capturing, meaning you can't really say Tannehill is playing better than Luck for the first 56 minutes. That is, they could be playing similarly, and Luck's <4 minute trailing performance might be the difference maker in the W/L record.
     
    Thunderkyss likes this.
  11. cbrad

    cbrad . Club Member

    7,521
    8,912
    113
    Dec 21, 2014
    What's surprised me is that Luck's decision-making hasn't improved more over the last 4 years. Move this discussion away from Tannehill for a moment and you can identify a bunch of decision-making issues with Luck that don't seem to be getting fixed. So I'm not so sure he's going to fix the INT issue. That's also a prove-it-to-me first thing.
     
  12. Finster

    Finster Finsterious Finologist

    3,087
    2,038
    113
    Jul 27, 2013
    If we throw out last year, due to the injury, his INTs went way down from his rookie year, in his 2nd and 3rd years combined he threw 2 1/2 times more TDs(63) than INTs(25), which is upper tier in the NFL, and even with his rookie season and injury season, he's still nearly at 2 to 1.
     
    dolphin25 likes this.
  13. cbrad

    cbrad . Club Member

    7,521
    8,912
    113
    Dec 21, 2014
    Yeah, his TD/INT ratio is good. But Luck gambles a lot (fits his name), often in situations where it's ill-advised. His INT totals are higher than they would otherwise be because of that. Luck (like Tannehill) has decision-making issues he needs to work on, and at least on this one issue I haven't seen much improvement from him over the years (taking into account I don't watch Indy nearly as much, but when I do see him play I see those ill-advised throws once again).
     
  14. Unlucky 13

    Unlucky 13 Team Rosen Staff Member Club Member

    26,406
    27,557
    113
    Apr 24, 2012
    Troy, Virginia
    I agree with you. Luck makes bad choices almost on the level of Fiedler or Henne, he just has the supreme talent to bounce back from them a lot more often, and to make it work more often when he's too gutsy. Almost like Bret Favre in that way, just with a better head on his shoulders off the field.
     
  15. Thunderkyss

    Thunderkyss Member

    53
    31
    18
    Sep 10, 2012

    Is that true?

    I believe one of the reasons Luck has such high numbers is that he's allowed to keep throwing. If Fiedler or Henne did some of the things Luck has, they'd be handing the ball off the rest of the night & most likely replaced for the next game.


    I mean Luck came into the league throwing the ball 50 times per game.
     
  16. Pauly

    Pauly Season Ticket Holder

    3,173
    3,111
    113
    Nov 29, 2007
    Aaron Rodgers numbers seem screwy. I will go back and double check, but yesterday my wife was suggesting that I spend less time on the computer and more time helping her with the household chores.
     
  17. Pauly

    Pauly Season Ticket Holder

    3,173
    3,111
    113
    Nov 29, 2007
    I've been trying to keep coaching/team mates out of my analysis. Mainly because I don't have any meaningful opinion on just what level of support Luck has gotten from his coaching and team mates.

    I want to do a 'no excuses' analysis. No offsets for injury, coaching, OL play whatever. Just what the QB has actually produced on the field.
     
  18. Pauly

    Pauly Season Ticket Holder

    3,173
    3,111
    113
    Nov 29, 2007
    I did a double take on Big Ben. Basically he's been a totally different QB in the last 4 or 5 years compared to early in his career.

    Also the data is incomplete, I was sampling some other QBs who are accused of being clutch, not analysing the league.

    One thing I did note with the season by season data though is when you had a QB with a really high average passer rating (110ish) in 0-7 games their win % was really high. For example Aaron Rodgers went 5-0 in 2014 with a 110 rating. Peyton Manning went 4-1 with a 108 rating. They were the only seasons I noticed with a really high passer rating.
     
  19. Pauly

    Pauly Season Ticket Holder

    3,173
    3,111
    113
    Nov 29, 2007
    Double checked Rothlesbirger's and Rodger's numbers.

    Minor change to Rodgers' TD and Int totals. No change to Rothlisbirger.
     
  20. Pauly

    Pauly Season Ticket Holder

    3,173
    3,111
    113
    Nov 29, 2007
    One thing I should say. I totally agree with Cbrad that if it is Luck's performance that causing Indy to have such a great W-L record in divisional games and 0-7 point games the Luck is doing something exceptional.

    Our area of divergence is purely in whether not not it's Andrew Luck that's causing Indy's record.
     
  21. cbrad

    cbrad . Club Member

    7,521
    8,912
    113
    Dec 21, 2014
    OK, I checked it myself. You got the TD's wrong for Roethlisberger. It's not 10 TD and 5 INT, but instead 6 TD and 5 INT:
    http://www.pro-football-reference.com/players/R/RoetBe00/splits/2012/
    http://www.pro-football-reference.com/players/R/RoetBe00/splits/2013/
    http://www.pro-football-reference.com/players/R/RoetBe00/splits/2014/
    http://www.pro-football-reference.com/players/R/RoetBe00/splits/2015/

    The other numbers were correct. So that gives us 87.16 rating for Roethlisberger, which is still an outlier, but not like 99.28 :wink2:. The reason I became suspicious is that, given the other data, 99.28 is like 4 sigma away from expected, meaning you'd expect that once in 15,000+ sets of 4 years!!! 87.16 is closer to 1 in 100, which is within reason.
     
    Pauly likes this.
  22. Pauly

    Pauly Season Ticket Holder

    3,173
    3,111
    113
    Nov 29, 2007
    Sorry for putting up a barrage of posts but I prefer to try to keep it one topic per post, rather than a huge mess of different ideas in one post.

    What things can move the needle as far getting a better (or worse) than .500 result in 0-7 point games?

    1) Coaching.
    Bill Belichek is the poster boy for a coach making a difference. By making better down and distance decisions and more efficient time management a coach can shift the win%

    2) Opponent.
    As discussed about GWDs. An opponent that has a poor pass defense and a poor passing game is more likely to allow a GWD, because each of your drives is more likely to score and each of their drives is less likely to score. So if you face a disproportionate number of 'bottom feeders' you should have a better than average record (PS random distribution is not even distribution so what matters most is the total number of bottom feeders faced, not which ones in particular).

    3) QB play.
    You'd have to do a league wide analysis over 4 or 5 years to get strong results. I'm not sure what is more important - QB play in the last 4 minutes or QB play over the game. Cases can be made for both situations.

    4) Fortune
    Random distribution is not even distribution. I basically view football as a series of connected random events, you've got 22 players, 7 referees, dozens of coaches, field condition etc. all of which affect the outcome of every single given play. Over a long run random events should even out, but not always.

    5) Exceptional D
    Having a great D (1 or 2 in the league kind of thing) should be giving you the ball more often and in better field position than average, which should help you in your GWD attempts. Also the opposing QB is going to be less likely to steal a GWD against your D
     
  23. Pauly

    Pauly Season Ticket Holder

    3,173
    3,111
    113
    Nov 29, 2007
    Thanks for that. I see what I did wrong, I mixed up 1ds and Tds in one year.

    It's tough to double check your own work because you see what you think should be there, not what actually is there.
     
    cbrad likes this.
  24. cbrad

    cbrad . Club Member

    7,521
    8,912
    113
    Dec 21, 2014
    Yeah, my rebuttal earlier was in the context of Luck vs. Tannehill. If ANY portion (probably a decent amount) of the W/L record in close games was due to Luck, that right there is the key difference between Tannehill and Luck, at least based on the stats we've looked at so far.

    I think the story is similar to how many would analyze the two QB's without stats: Luck has underperformed as a passer, but can be counted on to help win games that many other QB's (including Tannehill) cannot be counted on to win.

    As far as how much of the W-L in close games is due to Luck, hard to say but the default should be based on "variance explained", which is just the square of any correlation between your favorite stat and winning. Passer rating has a correlation of 0.51 with wins, so 25% of the W-L should be your default assumption.

    Of course, that's 25% of the entire team effort. If you can show there was a GWD, then that's arguably too low because the offense was more likely instrumental when it counted (even without adjusting for that suggestion by Fin D). Variance explained for the offense will of course be around double that. So I'm estimating you're looking at around 35% of the W-L due to Luck, but no way to directly measure this of course (btw.. this is where GWD and W-L record differ in terms of how they matter for analysis).
     
    Pauly likes this.
  25. Pauly

    Pauly Season Ticket Holder

    3,173
    3,111
    113
    Nov 29, 2007
    I think my main objection is that given Tannehill's numbers Miami's W-L totals lie about where they should, but Luck's numbers and Indy's W-L totals seem to me to be out of line.
    I expect Luck's and Indy's W-L numbers to be better because of divisional differences, but Indy's W-L record is better than Luck's raw numbers and strength of division explain.
     
  26. cbrad

    cbrad . Club Member

    7,521
    8,912
    113
    Dec 21, 2014
    Yeah I know, but remember you're basing all that off of using passer rating to adjust for different conditions. It's not accounting for the effect of all the other stats.

    I mean while we're at it, probably the best way to adjust for opponent is NOT to use passer rating, but to use W-L directly. If one had historical data on the probability one team with a given W-L record beats another team with a different W-L record, that's your adjustment right there. That type of adjustment will more naturally take into account what's being missed with passer rating. Still, good you did the work you did!
     
  27. Pauly

    Pauly Season Ticket Holder

    3,173
    3,111
    113
    Nov 29, 2007
    Well I started this rabbithole because I wanted to compare QB performance (relatively easy), not team W-L records (much more difficult).
     
    cbrad likes this.
  28. jdang307

    jdang307 Season Ticket Holder Club Member

    39,161
    21,798
    113
    Nov 29, 2007
    San Diego
    100% how I feel
     
  29. Pauly

    Pauly Season Ticket Holder

    3,173
    3,111
    113
    Nov 29, 2007
    This is something that I really want to dig deeper into.

    1) If a QB performs better in the first 56 minutes that should skew where his 0-7 point games fall. I know many other variables come into play but if your team's QB performs better for the first 56 minutes generally your games that are 0-7 in the last 4 will be against tougher opponents than if your QB performs at a lower level.

    2) It shifts the size of the lead and whether or not you have the lead. It shifts between having to drive for a TD and having to drive for a FG. It may also be the difference, for example between having a 1 point lead or a 2 point deficit.

    3) Strategy for playing with a lead is different to the strategy in playing trying to catch up. When you have a lead you want first downs to burn up time, when you have a loss yardage gained and not running off time are more important. Which means you can't treat the stats for playing with a lead and playing from behind equally.

    I'm not going to do that deep dive into the stats, because I have a job and a wife.
     
  30. cbrad

    cbrad . Club Member

    7,521
    8,912
    113
    Dec 21, 2014
    This one isn't difficult to do. Just plot passer rating vs. point differential. Here are the graphs for Tannehill (blue) and Luck (red):
    http://postimg.org/image/6qzaizms1/
    http://postimg.org/image/e6gn56xw1/

    Anyway, the best-fitting lines have slopes 0.93 for Tannehill and 0.82 for Luck, meaning that's how much increase in passer rating you need for each extra point on the point differential axis. Correlations are about 0.6 for Tannehill and 0.57 for Luck, showing a good relationship between passer rating and point differential.

    So I think that quantifies how much extra passer rating points (for the whole game, but that's dominated by the first 56 min anyway) affect point differential.
     
    Pauly likes this.
  31. Pauly

    Pauly Season Ticket Holder

    3,173
    3,111
    113
    Nov 29, 2007
    Just to put in perspective I've looked at the career TD leaders and checked their win% in 0-7 point games.

    We also have to remeber there is a fair degree of overlap with HoF caliber coaching as well as QB performance.

    Joe Montana. 43-34 (55.8%)
    Brett Favre 68-50 (57.6%)
    Dan Marino 65-45 (59.1%)
    Peyton Manning 77-40 (65.8%)
    Tom Brady 63-28 (69.2%)
    John Elway 66-51 (56.4%)
    Drew Brees 52-46 (53.1%)
    Fran Tarkenton 49-44 (52.6)
    Warren Moon 43-47 (47.7%)
    Johnny Unitas 37-21 (61.6%)

    Total: 606-406 (59.8%)
    my comment. HoF caliber QB play does give you an advantage in the W-L column. Probably in the region on 55% once you discount for HoF coaching. Then you have two outliers, Peyton Manning and then further out Tom Brady, both GOAT contenders, and Brady doing it with proabably the best coaching help of all time.

    Andrew Luck. 20-7 (74.1%)
     
    cbrad likes this.
  32. dolphin25

    dolphin25 Well-Known Member

    4,037
    1,101
    113
    Nov 22, 2014
    You much not remember fiedler much...... I got to where I would cringe every time he went back to pass expecting a pick 6
     
    Unlucky 13 likes this.
  33. Pauly

    Pauly Season Ticket Holder

    3,173
    3,111
    113
    Nov 29, 2007
  34. Pauly

    Pauly Season Ticket Holder

    3,173
    3,111
    113
    Nov 29, 2007
    Sorry for the thread necromancy but I thought it appropriate.

    At the end of 2019 career stats.
    Tannehill: 3197 attempts, 2030 completions, 23,176 yards, 145 TDs, 81 Ints. Passer rating 89.8
    Best season: 2019 - led league in passer rating, yards/attempt

    Luck: 3290 attempts, 2000 completions, 23,671 yards, 171 TDs, 83 Ints, passer rating 89.5
    Best season: 2014 led league in TDs.

    So their net career efficiency after a similar number of attempts is eerily similar.
     
    Unlucky 13, KeyFin, xphinfanx and 2 others like this.
  35. Puka-head

    Puka-head My2nd Fav team:___vs Jets Club Member

    6,560
    2,843
    113
    Nov 25, 2007
    Slightly left of center
    I'll say it again, been sayin it since ever since. Out of all the 1st round QBs drafted that year Tannehill will have the best career. Check that box. And other than Wilson he's had the best career of ANY QB drafted that year. And still ballin'.
     
    The_Dark_Knight and resnor like this.
  36. Pauly

    Pauly Season Ticket Holder

    3,173
    3,111
    113
    Nov 29, 2007
    Kirk Cousins is still ahead of Tanny, but not by too much. Cousins stats come out as slightly better than Tanny in every metric that goes into the passer rating formula.
    3145 attempts, 2104 completions, 24,107 yards, 155 TDs and 71 Ints. Passer rating 96.8

    I find it really interesting that all 3 have such similar number of attempts despite the various ways their careers have gone.
     
    resnor likes this.
  37. Unlucky 13

    Unlucky 13 Team Rosen Staff Member Club Member

    26,406
    27,557
    113
    Apr 24, 2012
    Troy, Virginia
    First off, you had me at the term "Thread Necromancy"

    Secondly, I'm always in favor of these things. As long as its not to pick a fight, going back to old discussions is good, because you can see what everyone said back then, rather than just starting a new one.

    And Third, I've been making a similar argument for a long time. Tannehill had the deck stacked against him his entire time in Miami. Now that he has a team that's willing to put him in the right offense and use him properly, look what happened. Its so vindicating that after I made this argument for years and was shot down relentlessly by people who don't like RT, to see it work when it finally happens. Good for RT and I hope he's the Titans QB for a long time.
     
    The_Dark_Knight, Pauly and resnor like this.
  38. Bumrush

    Bumrush Season Ticket Holder Club Member

    22,781
    21,696
    113
    Nov 25, 2007
    It's a shame my favorite memory of him in Miami was when he just let it rip and was unleashed his rookie year. That AZ game was a freaking marvel. I remember thinking Miami got lucky that they didn't suck for Luck. For a variety of reasons (I'm also looking at you Mike Wallace and Philbo the Clown) it just never came together in Miami. I'm not sure if he'll sustain his success in Ten but I'm happy for 17 and that Miami will likely draft his replacement this draft. A win win for all parties involved.
     
    resnor and Puka-head like this.
  39. The_Dark_Knight

    The_Dark_Knight Defender of the Truth

    8,656
    7,034
    113
    Nov 24, 2007
    Melbourne, FL
    Its a win win if Miami does draft a quarterback and IF said quarterback is indeed what they feel he is.

    Trubisky was supposed to be better than Mahomes and Watson. Oops
    Goff had a flash of brilliance and Wentz? Jesus, he gets hurt tripping over a blade of grass
    Jameis Winston? Mariota?

    If we do draft a QB, I just pray we don't draft a QB for the sake of drafting one. With the exception of Tannehill, we've been let down at the QB position so many times in recent memory.
     
  40. The Guy

    The Guy Well-Known Member

    1,731
    861
    113
    Oct 1, 2018
    Any comparison between Tannehill and Luck (and between Tannehill and any QB) needs to consider the difference between them in terms of their ability to sense and evade pressure and avoid sacks.

    We know for example that Tannehill has had a comparatively high sack rate throughout his career, despite numerous configurations of offensive linemen, coaches, receivers, running backs, and two different teams. We also know the Titans in 2019 had the fourth-best pass blocking win rate in the league, and that the team scored a high number of points per game, and so Tannehill's continued high sack rate with them was very likely not a function of 1) a poor pass blocking offensive line, or 2) being behind on the scoreboard and susceptible to defensive pressure. We also know that sack rate is one of the more stable statistics associated with quarterbacks when they change teams.

    So, Tannehill has had a problem with sensing and evading pressure and avoiding sacks throughout his career, and that appears to be attributable to him and not to his surroundings.

    We also know that 84% of offensive drives are stalled when they involve a sack, and that there is a -0.60 correlation between points scored and sacks in the league between 2012 and 2018, the seasons in question for both Tannehill and Luck.

    So, we need to take a look at what both quarterbacks did in terms of sacks during those seasons, because avoiding sacks is an important part of playing the quarterback position, and sacks have quite the deleterious effect on NFL offensive performance.
     

Share This Page