Right, the more relevant parameters (that actually exist) that require more data.....makes the model less predictive. Again, I ask you, if making the model more accurately reflect the relevant parameters in the subject matter and thereby increasing the data needed makes the model less predictive, then at what point do you realize the model is bad and you're just trying like hell to subjectively prove an already held belief through pseudo-objectivity?
By testing its accuracy. Choose the model given the data available that best predicts what you want to predict. In football, it's wins.
I would say that whatever Gase says has to be taken with a grain of salt, dissemination is part of the job. The only thing that matters, is what he "thinks" of RT now, and none of us know for sure what he's thinking, he may be disappointed, because he has had to reign in his O because RT can't really handle it? That's possible, as they have had to reign some things in, and this is a recurring RT story line, so there is precedence. Maybe he's just getting his back because "no one ever had his back before", which was one of the supposed "bugaboos" of RT's career, and something that Gase really hammered home as much as he could to the media, so is Gase just keeping his word by getting his back? Is it Gase just trying to get the most out of RT that he can? I would find it hard to believe that he doesn't have doubts about RT, perhaps he does believe he can fix RT, but there is no way I believe that he is satisfied with the QB play, because it has been a very vanilla, short based pass attack from a guy who is supposedly an offensive genius that is predicated on QB play.
Yeah, the model can't change each time you add new data. So you choose the model that is most robust at predicting the data you want over time. In practice that usually means simplicity works well. I mean.. if you have as many parameters as data points you can perfectly predict any data! And there are a bunch of guys in science that make careers off of what is essentially nothing more than curve fitting (adding more parameters as necessary). It's not good.
We're going round and round because you won't (or can't) admit you're wrong. You used part of a quote from Gase that had nothing to do with WHY he took the job, to try to prove that Gase had doubts about Tannehill when he took the job.
So now all he needs is the unicorn line. Doesn't need good coaching, run game commitment or audible powers. Then what the hell have you been blathering on about the past few years? So we don't need Gase, oline or run game. Lazor, Philbin, Tanny and the unicorn line is all we need.
LOL.. what a joke! I just gotta respond to this. You're arguing now that we can't use information a person gives about himself unless he was answering the specific question at hand? Tell that to prosecutors! So let's say a person accused of a crime says he was never at location X when asked directly. But we find tape of that person where things he says implies he was at location X in a completely unrelated conversation where a different question was asked. You'd say we can't use that info because it wasn't a response to the precise question asked. LOL No I'm not wrong dude.
He was asked if he thinks he can win a Super Bowl with Tanny. He just answered a question by saying he's gonna have Tanny's back this year, then is asked that and he says, oh I haven't seen him much. If you don't think you can win a SB with a guy you lost before you even got started. Eli has two rings. Flacco has one. Rex Grossman made it. Matt Hasselbeck made it and some argue they were robbed. "I got Tanny's back this year" Think you can win it allwith him? "Oh well I haven't seen him all that much so I can't answer that." Now, "I'm pretty sure he's the reason I came here."
No he's not. He said yesterday, the reason he took the job was because of Tanny. When asked asked before the season what he thought of Tanny, he could have said, of course we can win with him it's why I took the job. But he didn't. He said, oh I barely seent him. So we don't have any opinions yet. You have to strain the rules of logic to argue these two statements don't contradict each other.
Wait....wtf? Are you know pretending like we didn't point to that stretch of games has proof Thill is above average and you guys argued against it? Are you also going to pretend I haven't said 1000's of times on this site that the oline has been Thill's biggest roadblock and that when you have an oline that bad then you at least need to commit to the run and let the QB audible? Is that really what the **** you're doing? You are literally going to sink that ****ing low, that its not enough to contradict yourself time and time again, but now you're trying to spread that I am doing it to?
Its real simple. A good model for you guys has a 75%-ish success rate at predicting results. That still means 25% you're wrong. All that we've done is explain why the model fails 25% of the time and why its likely Thill resides in that 25%. You and your ilk, instead of acknowledging that, do everything in power to argue against that and pass it off as objective observation.
2 things: 1. 30% luck is not accurate at all. 2. If the success rate is at 70% or less, then my point is stronger.
Hilarious. Pulling a quote from an answer to a question about a Super Bowl, to reinforce the idea that Gase isn't sold on Tannehill isn't being disingenuous? If someone asks me my favorite color, I'd say blue. If they asked me if I love the color of my dog, I'd say yes, and she's gray and white. Would that mean I was lying about my favorite color?
I haven't. I also haven't seen any proof it is. I do know there's a metric **** ton of data your stats don't or can't account for that has **** all to do with stats.
Don't you guys ever get sick and tired of arguing every little word of each others posts? My gosh.. I mean, you're all arguing now over interpreting something Adam Gase said, now we're comparing it to colors and dogs, and colors of dogs... just so you can say each other is wrong apparently. Here's this... you will never know the right answer. None of you will ever be proven right. None wrong. At least not until Gase clarifies it for you, and while he's employed here, that wont happen. So, move on. All that's being accomplished here is you guys finding a new way to go after one another about a new post every other page. I hope Ryan Tannehill retires tomorrow so I don't have to see this stuff anymore and we don't lose more members over bans, or them making a choice to leave from frustration with this stuff. Don't get me wrong, I'm not saying discussing Tannehill, or things are off-limits, but this is all you guys do. However, I'm sure someone will tell me I'm wrong, or I really should have said this, or that, or that I really meant to say something else, or that there's a 72.34 % chance that I'm an idiot, or how dare I tell you what to think or post. The constant nitpicking, and nagging over little details in a post though is pretty mind-numbing. I applaud you guys for being committed though. lol Just my two cents... which equals about 1.4 cents really with the currency conversion probably.
So he's said literal words defending him, and saying he took this job knowing it was a good place because of RT.....but you're gunna go with hypotheticals of what he may be "thinking" instead. Oh....ok...
Yeah don't confuse correlation, which is post-hoc analysis, with predictive power. No one is good at predicting. Correlations can be very high though. However, it's the square of the correlation that matters, so it really goes down to something like at best 50% of the variance explained (from say a 0.7 correlation). Anyway, I've always acknowledged limitations of stats. Ton of posts in a few threads here doing that. What you haven't acknowledged is that all you guys offer is a possible explanation for what happened. You're not providing any hard evidence the portion of variance unexplained was primarily due to what you suggest.
FWIW, the best analysis I've seen re: luck vs. skill was from Brian Burke: http://archive.advancedfootballanalytics.com/2007/08/luck-and-nfl-outcomes.html http://archive.advancedfootballanalytics.com/2007/08/luck-and-nfl-outcomes-2.html http://archive.advancedfootballanalytics.com/2007/08/luck-and-nfl-outcomes-3.html Burke's analysis finds that luck is less than 30% - it is 26%.
On the what percent is luck issue.. no one knows, but it is interesting to see coach win% level off at around 70%: https://i.imgsafe.org/4a62d6fe89.png What that suggests is that the ability to effect a particular outcome is limited at around that level for the best teams (coaches tend to have longer tenure the more they win). Now, that data is consistent with a higher percentage for luck than 30% but it's a bit hard to argue based on this data alone that luck is less than 30% when no one really demonstrates it. One thing to point out: this is different from predictive power. It's theoretically possible to predict better than how well the best coaches/teams can affect the outcome of a game if you know how to use the data better than they do.
I'm sorry dude, but did you even read what I wrote? All I did was put out hypotheticals, and in conclusion said that I'm sure he wants more from the QB. Do you have proof that he means what he said, or is he just towing the company line? One thing is for certain, neither you or I know for sure what he's thinking, which was the entire point of the comment.
It can get tiring and sure it would be better if things were different. But one thing won't change: we're all on the same side come gameday. Good enough for me!
I get it. The difference is there's evidence of one and not the other. You're right, its hypotheticals. Adam Gase has been nothing but upfront and honest, not sure why he'd bull**** about Tannehill, he's not wrong....longest winning streak since 2008, and hes been clean as can be throughout the streak with a line and running game. So he's correct that he's not the ONLY issue over the years, and we can win with him....because we ARE.
Gase had a chance to put his full support behind Tanny at his press conference, he did not give him a vote of confidence. He gives him a vote of confidence after a 5 game win streak, and says that confidence was there all along. Do what you want with that, but that's what happened.
There is a 95% chance that you leave the Horseshoe on Saturday upset. Harbaugh has poisoned Mich fans souls by feeding them hope, watching them clamor in the sun...you can watch Ohio State destroy an entire fanbase. Then, when it is finished...and your playoff hopes are.....ashes....then you have Urban's permission to be eliminated. Had to. Go Bucks
You sir, need to quit drinking this early on a Tuesday. However, that being said... I don't doubt Michigan drops the game. I didn't like their QB to begin with, and I sure don't like their backup either.
So...first time head coach comes in, takes over struggling franchise, doesn't guarantee Super Bowl win, and that's portrayed as not confident in the quarterback? Perhaps he wasn't confident because he wasn't sure we had a good enough team? Perhaps he wasn't sure because we didn't have a good enough oline? Nope, it's gotta be Tannehill.
We have provided hard evidence. You guys ignore it or discount it based on....what appears to be at best 50%-70% likelihood of X. The current streak we're on is proof enough actually.
Again, all that is happening there, is that a bunch of numbers either fall through the cracks (ability of a given receiver on a type of route, etc.) or can't be gotten (play calling, etc.) so what's left is called "luck".
Saying you can win with Tanny is now guaranteeing a Super Bowl? Keep stretching it buddy. The rest of your post is irrelevant. The question was simple and clear. Does he think he can with with Tanny. His answer? He didn't answer. Said he has barely seen Tanny play. Look, I love Gase. I think Tanny has shown some real stuff the past couple weeks. No need to make things up. Gase's words are crystal clear and they are in 100% contradiction. Where was this proclamation during the losing streak when people asked if he was going to be benched?
What are you going on about? THIS is the quote we're discussing: So, Gase was asked specifically about playoffs and Super Bowl. He talked about Tannehill...and the rest of the team. Yet somehow, you guys are spinning this into about Tannehill. He wasn't asked about generically winning games with Tannehill. He was asked a specific question, and answered it. Not surprisingly, you guys are using his answer to answer a different question.
Also, I'm pretty sure that Gase was out in the media saying Tannehill was not at all at fault during the losing streak. In fact, you may remember my "apology" thread...right around that time, Gase came out swinging at critics of Tannehill, and I said, If Gase still is backing Tannehill, then I'm still all in on him. Gase never, ever intimated that Tannehill was in danger of being benched. He's backed him this entire season.
"I can win the Super Bowl With Tanny." "We are going to win the Super Bowl, with Tanny." Are not the same thing. He was asked the first (and he declined to answer). You defend him by saying he wouldn't say the second (which he wasn't asked). I don't really think this is a big deal. It's just funny. Gase is not a deity here. No need to form a cult of personality and take everything he says as the final word. He wasn't as confident in Tanny before taking the job as he put forth the other day. He's lying. We have proof. His own words. Simple as that. Who cares. Only thing that matters is what is put on the field on game day.