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

Ryan Tannehill best games compilation

Discussion in 'Miami Dolphins Forum' started by Brasfin, May 15, 2016.

  1. cbrad

    cbrad .

    10,659
    12,657
    113
    Dec 21, 2014
    2nd link first paragraph in post #269. Not all events have to have equal probability for something to be random.
     
  2. resnor

    resnor Derp Sherpa

    16,329
    9,874
    113
    Nov 25, 2007
    New Hampshire
    Yet, a prediction is black or white. For instance, I say, "I predict that tomorrow the sun will rise at 5:45am" Then, tomorrow the sun rises at 5:43am. Was my prediction right? Was it partially right? No, it was wrong, or it was right.

    The reason you can predict with varying degrees of success, is because, in football for instance, some events have higher occurrences than others, giving you more data to identify patterns, etc, allowing you to more precisely predict. But the more data, the less random. Truly random events cannot be predicted because they are, well, random.
     
  3. Fin D

    Fin D Sigh

    72,252
    43,684
    113
    Nov 27, 2007
    In statistics, (not in layman's terms) "random" and "random variable" are different though right? I mean a given event's result could be random because of random variables...no?
     
  4. cbrad

    cbrad .

    10,659
    12,657
    113
    Dec 21, 2014
    In math, a random variable is basically something you can use in a mathematical argument (or equation etc..) that outputs something "random". So the only difference is that "random" is a concept while "random variable" actually outputs a randomly chosen event (number) from some pre-specified set of numbers. Do that repeatedly and you get what is called a "random process" or a "stochastic process":
    https://en.wikipedia.org/wiki/Stochastic_process
     
  5. Finster

    Finster Finsterious Finologist

    3,087
    2,038
    113
    Jul 27, 2013
    Now, if you do that 100 times and are right 70 times, it could then be said that you have a 70% probability of correctly predicting the time of sunrise.
     
  6. Pauly

    Pauly Season Ticket Holder

    3,696
    3,743
    113
    Nov 29, 2007
    OK To start with I will put some links to where I originally posted the data.

    1) http://www.thephins.com/forums/show...ehill-should-be-a-better-QB-now-Lazor-is-gone
    2) http://www.thephins.com/forums/showthread.php?88889-RT17-and-the-run-game
    3) post 211 in this thread http://www.thephins.com/forums/show...-best-games-compilation&p=2801074#post2801074

    We started talkinh about probability and predictions etc. when we were discussing whether Lazor's offense was predictable and whether defenses were able to respond to it (post 211)
    Short summary:
    In games where Miami rushes more than average (36%+of the time) Tannehill's ypa is 7.9. Interceptions 1.9%
    Rushing ypa is 4.67

    In games where Miami rushes less than average (36% of the time) Tannehill's ypa is 6.0. Interceptions 2.1%
    Rushing ypa is 4.32



    Another control to check if it is Lazor's playcalling is to look at NFL averages for behind -v- tied/ahead.

    NFL Starting QBs (all)
    Tied/Ahead
    4649 completions of 7269 attempts for 54,310 yards. 374 TDs 151 Ints.
    63.9%; 7.5 ypa; 5.1 TD%; 2.1 Int%. 93.0 rating
    Behind
    4663 completions of 7161 attempts for 54,556 yards. 341 TDs 188 Ints.
    65.1%; 7.6 ypa; 4.7 TD%; 2.6 Ints. 93.0 rating

    Taking the NFL as a whole you can say that OCs do a good job of being unpredictable according to game situation.

    Looking at the top tier QBs.
    Tied/ahead
    1808 completions of 2743 attempts for 21,955 yards. 168 Tds, 50 Ints.
    65.9%; 8.0 ypa; 6.1% TDs; 1.8% Ints. 103.2 rating
    behind
    1006 completions of 1539 attempts for 12,192 yards. 80 TDs, 33 Ints.
    66.4%; 7.9 ypa; 5.1% TDs; 2.1% Ints; 98.0 rating
    The top Qbs had very similar performance according to game situation.

    Middle tier QBs
    tied/ahead
    2030 completions of 3204 attempts for 23,422 yards. 159 TDs and 65 Ints
    63.4%; 7.3 ypa; 5.0 TD%; 2.0% Ints; 93.4 rating
    Behind
    2197 completions of 3506 attempts for 25,178 yards. 156 TDs and 78 Ints.
    62.7%; 7.2 ypa; 4.5 TD%; 2.2% Ints; 89.8 rating
    Again with the mid tier QBs the playcalling seems stable according to game situation.

    Bottom Tier QBs
    Tied/ahead
    810 completions of 1322 attempts for 8,933 yards. 47TDs and 36 Ints.
    62.2%; 6.7 ypa; 3.5 TD%; 2.& Int %; 82.6 rating
    Behind
    1468 completions of 2418 attempts for 17,186 yards. 105 TDs and 77 Ints
    60.4%; 7.1 ypa; 4.3 TD%; 3.2 Int%; 83.2 rating
    As a group the bottom tier QBs (Tannehill's tier for 2015) were throwing lower % deep balls when behind for more TDs and more Ints. I would say this is most probably a function of bottom tier QBs playing from further behind than the other QB tiers, and their OCs calling deeper passes.


    My conclusion is that Lazor's playcalling became predictable to the point where it was an exploitable weakness.
    If Tannehill was giving defenders clues to the passes I would expect his interception% to increase dramatically.

    The data, in my view, supports the conclusion that
    (1) The NFL average OC's predictability does not alter according to behind/not behind game situation.
    (2) Tannehill's drop in ypa is consistent with defenses taking away expected route combinations forcing him to check down

    Just for fun I looked at Jay Cutler under Gase in 2015.
    Tied/ahead
    62 of 149 for 949 yards. 3 TDs and 2 Ints
    41.6%, 6.4 ypa. 2.0 TD%, 1.3 Int%. 63.0 rating
    Behind
    219 of 337 attempts for 2710 yards. 18 TDs and 10 Ints
    64.8%, 8.0 ypa; 5.3 TD%; 3.0 Int% 95.0 rating

    Cutler is well known as an idiosyncratic QB, and I'd want a lot more data to draw conclusions about Jay Cutler or Gase's offense. However, predictability of play calling when behind wasn't one of Gase's problems last year.
     
    resnor, Pandarilla and Fin D like this.
  7. Pauly

    Pauly Season Ticket Holder

    3,696
    3,743
    113
    Nov 29, 2007
    If we take the numbers from Cbrad's post 214, where he calculated the hypothetical effects of Miami's "unpredictable offence" (36%+ rushing attempts) and "predictable offence (less than 36% rushing)

    Predictable offense; 51.2% chance of getting another first down
    Unpredictable offense: 59.1% chance of getting another first down.
     
  8. Fin-O

    Fin-O Initiated Club Member

    11,375
    11,392
    113
    Sep 28, 2015

    ----- wife is a sure thing, yet i would never touch it.
     
  9. roy_miami

    roy_miami Well-Known Member

    1,385
    560
    113
    Oct 11, 2013
    I don't know about all that but my prediction is with Tannehill at the helm our offense will never be as good as it was in 2014 under Lazor when it ranked 8th overall and was 11th in passing and 2nd in rushing, according to Football Outsiders. In fact, with Lazor over his year and a half our offense was as good as its been since the Marino days.

    Lazor's offense was ranked 8th in 2014 and 16th until he was fired in 2015. We were also 8th and 16th in 2008/2009 and the next best was 11th and 17th in 2002/2003. See, people need to have a little perspective when declaring Lazor was one of the worst ever. The fact is he's the best Dolphin fans have seen in a long time, but because Tannehill doesn't look like Russell Wilson then Lazor must be horrendous.

    When I read the tea leaves what I see is Lazor had a good football mind, probably well above average, but was a prick and was not well liked, so overall was an average coach.
     
    dolphin25 and Pauly like this.
  10. Stringer Bell

    Stringer Bell Post Hard, Post Often Club Member

    44,356
    22,480
    113
    Mar 22, 2008
    By this definition, everything is a gamble, aside from death.

    Sent from my SAMSUNG-SM-G920A using Tapatalk
     
  11. Stringer Bell

    Stringer Bell Post Hard, Post Often Club Member

    44,356
    22,480
    113
    Mar 22, 2008
    Most would differentiate risk assessment from gambling. There is a risk that one's employer goes bankrupt and can't pay its employees, but most would not characterize going to work as gambling.

    Sent from my SAMSUNG-SM-G920A using Tapatalk
     
  12. Stringer Bell

    Stringer Bell Post Hard, Post Often Club Member

    44,356
    22,480
    113
    Mar 22, 2008
    What are probabilities used for?

    Sent from my SAMSUNG-SM-G920A using Tapatalk
     
  13. cbrad

    cbrad .

    10,659
    12,657
    113
    Dec 21, 2014
    Are you sure you're calculating those numbers correctly? Reason I ask is because 51.2% is lower than ALL values you're taking a weighted average of, which is mathematically impossible.

    Just so it's clear, what you need to do is to take each of those numbers, multiply them by some weight, then sum them, then divide that sum by the sum of the weights. The weight for say run-pass-run should be run%*pass%*run% (except it's not in percentages). So if run % is 36% (let's say), the weight would be 0.36*(1-0.36)*0.36.
     
  14. cbrad

    cbrad .

    10,659
    12,657
    113
    Dec 21, 2014
    So to summarize the key points of your argument, you're comparing 2014-2015 Tannehill (under Lazor) vs. 2012-2013 Tannehill and showing evidence of higher risk plays when trailing with Sherman and lower risk plays (i.e more predictable) under Lazor. You've showed multiple times that the run/pass ratio changed in favor of heavy pass under those conditions. So your conclusion is that Lazor's pass heavy offense was more predictable.

    I'm not sure I buy that argument for two reasons even though it's certainly a possibility.

    First reason is that the logic you used to argue Lazor was more predictable than Sherman when trailing can be used to argue Lazor was less predictable than Sherman when tied or ahead. I find that highly implausible as it's usually the case people are relatively stable in their predictability (which you yourself suggested with the overall averages).

    Second reason is that even if Lazor was more predictable, I'm not sure how the run/pass ratio matters. That is, if you run 2/3 of the time, you're as predictable in terms of "run" vs. "pass" as if you pass 2/3 of the time. Of course, the calculus changes if there are multiple run or pass plays, and it might be worth taking into account the number of possible run or pass plays Sherman and Lazor called in those situations (can anyone provide that info for us?). Point is, the ratio itself tells you nothing if the ratios are the same, which they are according to your stats: 2/3 run vs. pass when tied or ahead in the "top 8" vs. 2/3 pass vs. run in the "bottom 8".


    I'd venture to suggest a different interpretation: Lazor thought what Tannehill would be good at is making lots of shorter, high-percentage throws. He devised an offense around that philosophy that made Tannehill more efficient when ahead, but just happened to hurt him when behind because you need to be willing to take a lot more risks when behind. This fits with the idea that predictability is relatively stable for a person and also fits your stats. The run/pass ratio I'd say is simply an artifact of game situation. That is, heavy run is a consequence of being ahead, not a cause, and heavy pass is a consequence of trailing. So while Lazor might prefer pass in general, the ratio itself as a function of game situation isn't the cause of Tannehill's efficiencies.
     
    Pauly likes this.
  15. resnor

    resnor Derp Sherpa

    16,329
    9,874
    113
    Nov 25, 2007
    New Hampshire
    Um, yeah. For instance, every time you drive on the road, you are gambling that probability will be in your favor, and you won't be in an accident. Every been in a car accident? You lost your gamble that day. Every time you sit on a chair, it's a gamble that the chair will hold you. Ever sat on a chair that broke, or tipped over? You lost that gamble when that happened.
     
  16. resnor

    resnor Derp Sherpa

    16,329
    9,874
    113
    Nov 25, 2007
    New Hampshire
    Good question. Do you believe that stating that someone has a 1:1,000,000 chance at winning the Powerball is the same as predicting when someone will win it? Powerball goes months without a winner, even though there is always a statistical probability that someone will win every week.
     
  17. Stringer Bell

    Stringer Bell Post Hard, Post Often Club Member

    44,356
    22,480
    113
    Mar 22, 2008
    If your argument is that literally everything is a gamble, then what purpose does its distinction make?

    Every common use of the term relies upon the participant having no impact on the outcome, and skill is not a factor. When you are choosing a side with a statistical advantage, you are becoming a participant in your outcome, rather than a bystander. This is an important distinction, because it is precisely what the NFL argue as to why betting on NFL games is gambling - those betting will never achieve a statistical advantage.

    To the car anology - there is a high level of skill involved. Not only in assessing the risk (don't drive into a hurricane), but also in how one controls a car and exerts judgement.

    Sent from my SAMSUNG-SM-G920A using Tapatalk
     
  18. roy_miami

    roy_miami Well-Known Member

    1,385
    560
    113
    Oct 11, 2013
    I predict that you will not win the powerball. I would even be willing to "gamble" and put $100 on that prediction.
     
    dolphin25 likes this.
  19. roy_miami

    roy_miami Well-Known Member

    1,385
    560
    113
    Oct 11, 2013
    Piddly little like being able to change from run to pass or pass to run on most plays? Yet the QB is not culpable at all in the imbalance of the pass/run ratio? The way I see it the QB pretty much decided the what run/pass ratio would look like.

    Its entirely possible, probable even, that if you want to "fix" the run/pass ratio you need to strip Tannehill of even more aubible power, not add to it.
     
    Finster likes this.
  20. resnor

    resnor Derp Sherpa

    16,329
    9,874
    113
    Nov 25, 2007
    New Hampshire
    But I might win. And you didn't answer my question. Really, you aren't predicting anything. You're making a gamble, based on probability, that I won't win. It's a smart gamble, but it's still a gamble.
     
  21. resnor

    resnor Derp Sherpa

    16,329
    9,874
    113
    Nov 25, 2007
    New Hampshire
    Can you factor the odds, the probability, of being involved in a car accident? Yes you can. Therefore, every time you drive, you are making a decision, based on experience, and other data, that you will be safe in your car. Yet, every day, people are in accidents. It's a gamble. There's absolutely no guarantee that you won't be in an accident. Therefore, driving in your car means you are taking risk. It's a gamble.
     
  22. Stringer Bell

    Stringer Bell Post Hard, Post Often Club Member

    44,356
    22,480
    113
    Mar 22, 2008
    Taking risk != gambling. A game of chance cannot involve any skill. If you can impact the odds, you are exerting skill.


    Sent from my SAMSUNG-SM-G920A using Tapatalk
     
    Finster likes this.
  23. resnor

    resnor Derp Sherpa

    16,329
    9,874
    113
    Nov 25, 2007
    New Hampshire
    Whatever you say.
     
  24. roy_miami

    roy_miami Well-Known Member

    1,385
    560
    113
    Oct 11, 2013
    Yes you might. But I predict you won't. Notice I didn't say I'm certain you won't. There is a reason predict and certain are different words with different meanings.
     
    Finster likes this.
  25. Finster

    Finster Finsterious Finologist

    3,087
    2,038
    113
    Jul 27, 2013
    I was just about to point that out, that prediction is not an absolute, because for some reason he has got it in his head that to predict, is to be 100% correct, but prediction is not tied to 100% positive outcome.
     
    dolphin25 likes this.
  26. Finster

    Finster Finsterious Finologist

    3,087
    2,038
    113
    Jul 27, 2013
    Taking risk = gambling, is 100% correct, and it's taking a relatively high risk that makes it gambling, or simply placing wagers on anything is also gambling.

    Dice, poker, pool, etc, are all well known gambling pass times, however they are games that can have skill, but are still gambling based on wagering, however if you play any of these games without putting anything of value on the line, then they are not gambling.

    Driving to work while observing the traffic laws, and watching the road is not a gamble, although anything can happen, but the margin is so low that it is not considered a gamble, but if you speed, tailgate, text, etc, you are now entering into the gambling area, gambling on getting a ticket, or worse.
     
  27. cbrad

    cbrad .

    10,659
    12,657
    113
    Dec 21, 2014
    Just want to make sure you understand that when Stringer Bell wrote "Taking risk != gambling" that the "!=" was not a typo. That means "does not equal" in some programming languages like python. So I hope you understand your position is vastly different than his :wink2:
     
  28. resnor

    resnor Derp Sherpa

    16,329
    9,874
    113
    Nov 25, 2007
    New Hampshire
    Gambling doesn't have to have high risk.

    But, I agree, Finster, taking risk is gambling.

    Roy, you're not predicting. You're restating probability. Predicting would be you telling me that based on my buying tickets history, I will win on X ticket, given an assumption that I could live that long.
     
  29. resnor

    resnor Derp Sherpa

    16,329
    9,874
    113
    Nov 25, 2007
    New Hampshire
    Definition of gamble:

    Driving a car fits the definition. So does playing Powerball. So does any number of things. The difference is, perceived risk of many events is so low, and they're so commonplace to us, that we don't recognize it as taking a chance.
     
  30. Finster

    Finster Finsterious Finologist

    3,087
    2,038
    113
    Jul 27, 2013
    I was not aware of that !=, however, I'll bet our stances aren't vastly different, oh wait, did I just gamble?
    :shifty:
     
  31. Finster

    Finster Finsterious Finologist

    3,087
    2,038
    113
    Jul 27, 2013
    Gambling by definition, if we are throwing out placing wagers, is taking a high risk, something with minimal risk isn't gambling.
     
  32. resnor

    resnor Derp Sherpa

    16,329
    9,874
    113
    Nov 25, 2007
    New Hampshire
    I don't find any definition that relegates gambling to only taking high risk. If someone only has $10 to their name, and they bet $5 on a roll of dice, that is a large risk to them, but it wouldn't even cause Bill Gates to blink.

    Regardless, if you're playing poker, using pennies for wagers, it's still gambling, even if the risk is miniscule, i.e., you're only losing a few cents here and there. Playing nickel slots is still gambling. Going biggie jumping is gambling, because there is a perceived risk of bodily harm (for most people), with a gain of adrenaline boat and excitement.

    Risk is all relative to the person taking the risk.
     
  33. CaribPhin

    CaribPhin Guest

    We've really trotted out some JAGs for this team man:

    Jeron Mastrud
    Marlon Moore
    Marcus Thigpen
    Brian Hartline
    Brandon Gibson
    85% of the O-line since 2012
     
    Hiruma78 likes this.
  34. Fin D

    Fin D Sigh

    72,252
    43,684
    113
    Nov 27, 2007
    Wait, things are getting weird.

    The original comment about gambling was:
    How did that change to this:
     
    resnor likes this.
  35. Stringer Bell

    Stringer Bell Post Hard, Post Often Club Member

    44,356
    22,480
    113
    Mar 22, 2008
    Assessing risk is a skill.
     
  36. resnor

    resnor Derp Sherpa

    16,329
    9,874
    113
    Nov 25, 2007
    New Hampshire
    The problem is, people are ignoring the actual definitions of words, and attributing other characteristics to the words.
     
  37. resnor

    resnor Derp Sherpa

    16,329
    9,874
    113
    Nov 25, 2007
    New Hampshire
    Assessing risk doesn't make something not risky, or make it not gambling.

    For instance, when roy bet $100 that I wouldn't win the Powerball, he was gambling, based on his assessment that the risk was vastly in his favor. But he was still gambling.
     
  38. roy_miami

    roy_miami Well-Known Member

    1,385
    560
    113
    Oct 11, 2013
    So I'm not allowed to predict that you won't win the lottery? Where do you draw the line, am I allowed to predict the Patriots won't win the Super Bowl? The AFC? The AFC East?
     
  39. Fin D

    Fin D Sigh

    72,252
    43,684
    113
    Nov 27, 2007
    Knowing the odds and impacting the odds are two completely different things.
     
  40. Dummuck

    Dummuck Member

    55
    25
    8
    Oct 9, 2015
    Forest City N.C.
    I agree.

    It is the coaching giving him just too small of a window or bad calling completely for his style of play..I loved Coach Campbell but he did not change the way that was set up from the other coaches.
     

Share This Page