Suggs acuses Goodell of Superbowl Blackout.

[PatriotSeven]

Yeah, it's statements like that that once again prove you have no clue what you are talking about. Sample size ALWAYS matters when you are trying to establish what you are saying.

You are in line with Terrell Suggs, and you believe the Patriots SB36 victory was rigged. Keep it up bro and you'll be the laughingstock of the forum.

No it doesn't. Statements like yours is what tells me what I'm discussing is too far over your head.

And the only thing you have to come back with is that I'm a laughing stock. From where I stand, it doesn't look like that.

It's as stupid as saying you need a large sampling size to figure out if you hit the Powerball your chances of hitting it were 1 in 175,223,510.

You can look at a sampling size of 10000 years, and the chances of that happening won't change that it's a 0.000x% occurrence. I don't need to actually wait for that to happen to figure it out.

The fact that it's such a small time period, less than 30 years, and it popped up is actually what makes it even more unlikely.

You need a large data set if you're searching for the Higgs Boson and need a high confidence level to get there. But if the freaking thing came out and said "hi I'm a Higgs Boson" in the first 10 minutes you started your test, you may want to ask some questions or consider yourself incredibly lucky. Or maybe, someone's pulling a prank on you.

 

[SB39]

No it doesn't. Statements like yours is what tells me what I'm discussing is too far over your head.

I have a Master's Degree in statistics. I analyze numbers for a living. I have almost about every SAS certification they make (which was pretty convenient considering I used to work there). I'm not even going to start talking to you about concepts I know you have no understanding of whatsoever like proper hypothesis testing, confidence intervals and p-values.

I wrote up a quick SAS program and I ran the numbers. I even used the numbers you provided, whereby the higher seed has a 60% chance of winning any game (and I included the condition that the top 2 seeds will always "win" round 1). The p-value for the 2005-2012 results comes out to 0.08. I know you have no idea whatsoever what that means so let me put it in words you'll understand:

If we analyze the past 7 years which you are obsessing over, there is a 8% chance we get the results we've seen or results which are more extreme.

Is 8% unusual? Sure. 3 #5 and #6 seeds winning 7 Super Bowls is an unlikely result. However, there is not a single hypothesis test in the world that would accept a p-value of 0.08 as statistical proof that something is afoot other than what would normally be expected. There is absolutely nothing statistically noteworthy or unusual about the actual results relative to what we would expect to see. Just try getting a drug approved by the FDA with a clinical trial resulting in a p-value of 0.08. you'd get laughed out of town.

EDIT: I had to update my numbers after realizing I failed to account for 1 of the seasons.

 

[DisgruntledTunaFan]

Baltimore lost a key OL at the end of the year, and put Bryant McKinney at LT. They were among the league leaders in penalties, sacks allowed, and among the bottom in sacks made, and only won 10 regular season games.

But somehow they won the Super Bowl? I'm not saying it's rigged, but I've seen stranger things happen.

 

[PatriotSeven]

I have a Master's Degree in statistics. I analyze numbers for a living. I have almost about every SAS certification they make (which was pretty convenient considering I used to work there). I'm not even going to start talking to you about concepts I know you have no understanding of whatsoever like proper hypothesis testing, confidence intervals and p-values.

I wrote up a quick SAS program and I ran the numbers. I even used the numbers you provided, whereby the higher seed has a 60% chance of winning any game (and I included the condition that the top 2 seeds will always "win" round 1). The p-value for the 2005-2012 results comes out to 0.08. I know you have no idea whatsoever what that means so let me put it in words you'll understand:

If we analyze the past 7 years which you are obsessing over, there is a 8% chance we get the results we've seen or results which are more extreme.

Is 8% unusual? Sure. 3 #5 and #6 seeds winning 7 Super Bowls is an unlikely result. However, there is not a single hypothesis test in the world that would accept a p-value of 0.08 as statistical proof that something is afoot other than what would normally be expected. There is absolutely nothing statistically noteworthy or unusual about the actual results relative to what we would expect to see. Just try getting a drug approved by the FDA with a clinical trial resulting in a p-value of 0.08. you'd get laughed out of town.

EDIT: I had to update my numbers after realizing I failed to account for 1 of the seasons.

Burn the freaking thing. And your program.

 

[SB39]

Burn the freaking thing. And your program.

Sure thing. But that ain't going to change the fact that you have no clue what you are talking about and I am 100% correct. The ONLY thing unusual in our 7 year sample is that a #6 seed won twice, but even that is not so ridiculously unusual as to be statistical evidence of anything whatsoever.

Sorry, brah, the numbers don't lie. You don't even know what a p-value is or what hypothesis testing is but you think you can argue with someone who does? Nice try, brah.

 

[PatriotSeven]

Sure thing. But that ain't going to change the fact that you have no clue what you are talking about and I am 100% correct. The ONLY thing unusual in our 7 year sample is that a #6 seed won twice, but even that is not so ridiculously unusual as to be statistical evidence of anything whatsoever.

Sorry, brah, the numbers don't lie. You don't even know what a p-value is or what hypothesis testing is but you think you can argue with someone who does? Nice try, brah.

Dude, your numbers were more right the first time you said it, BEFORE you figured out variation based on probability. You were actually correct 7 pages ago when you said 1.5%.

If you realize that's without a bye week, equal probability, no homefield advantage if you got any result over that number you need to go back and re-do it. In your attempt to "prove me wrong" with a program you lost your common freaking sense!

Unless your number was actually .08% not .08 and you messed up. In that case, I'll buy it, but that's not exactly what I would define as just a "small chance" anymore is it?

 

[SB39]

Unless your number was actually .08% not .08 and you messed up. In that case, I'll buy it, but that's not exactly what I would define as just a "small chance" anymore is it?

You don't even have the first clue what a p-value is, but my numbers and calculations are 100% accurate. I used the assumption you provided us with, that being a 60% chance for the higher seed to win (though I think that number is way too high - I even used it in the Super Bowl where there is no HFA). I also included the first round byes.

The p-value was 0.08. There isn't a statistician on the planet that would accept a p-value that high as statistical evidence proving something. Like I said above, try running a clinical trial and going to the FDA with a p-value of 0.08. They'll laugh you out of the building. Kinda like how the whole forum is laughing at you right now.

 

[PatriotSeven]

You don't even have the first clue what a p-value is, but my numbers and calculations are 100% accurate. I used the assumption you provided us with, that being a 60% chance for the higher seed to win (though I think that number is way too high - I even used it in the Super Bowl where there is no HFA).

The p-value was 0.08. There isn't a statistician on the planet that would accept a p-value that high as statistical evidence proving something. Like I said above, try running a clinical trial and going to the FDA with a p-value of 0.08. They'll laugh you out of the building. Kinda like how the whole forum is laughing at you right now.

Dude I don't really care about a p-value because there is nothing to prove. It happened. It's about figuring out the probability of a perfect mix happening, and if you're telling me you came up with 8% I guarantee you your 8% is 100% WRONG.

There are over 46,656 freaking possibilities, out of which only 720 are acceptable, under the condition of equal probability of advancement, and you're actually claiming you expect to see a perfect mix of winning seeds, in a time span of 6 years, 8 times in 100 years! Stop and think how ridiculous that sounds.

The chances of just a 6 seed alone even winning 6 times in 100 years is 6%! And not only do you expect it 8 times, but you expect a perfect mix of a 6 string, non-repeatable number(perfect parity), showing up 8 times. LOL. Ridiculous.

 

[SB39]

Dude I don't really care about a p-value because there is nothing to prove.

If you're trying to prove something statistically, then you darn well better care what a p-value is. These are the methods that grown-ups use to determine if there is a cause-and-effect present, or if something is just a cool coincidence.

It's about figuring out the probability of a perfect mix happening, and if you're telling me you came up with 8% I guarantee you your 8% is 100% WRONG.

Again, you're demonstrating that you just don't have the first clue about statistics.

I took the 7 year period you keep harping on, 2005-2012. (BTW, that's actually an 8-year period, but I am focusing on the 7 years you keep harping on). Using the assumption that the higher seed wins 60% of the time (50% when same seed meet in the Super Bowl) and factoring in first round byes, there is an 8% chance we get the results we've observed (or results which are more extreme).

And 8% ain't statistical proof of anything.

 

[PatriotSeven]

If you're trying to prove something statistically, then you darn well better care what a p-value is. These are the methods that grown-ups use to determine if there is a cause-and-effect present, or if something is just a cool coincidence.
Again, you're demonstrating that you just don't have the first clue about statistics.

I took the 7 year period you keep harping on, 2005-2012. (BTW, that's actually an 8-year period, but I am focusing on the 7 years you keep harping on). Using the assumption that the higher seed wins 60% of the time (50% when same seed meet in the Super Bowl) and factoring in first round byes, there is an 8% chance we get the results we've observed (or results which are more extreme).

And 8% ain't statistical proof of anything.

Because it's not freaking 8%. That's your problem in the first place. And the point at hand is that of a perfect mix of winning seeds, so I don't know what the hell you are calculating.

Calculate the odds for a perfect mix of 6 winning strings over a 6 time period. It's not 8%! Not even under equal conditions of success.

 

[SB39]

There are over 46,656 freaking possibilities, out of which only 720 are acceptable, under the condition of equal probability of advancement, and you're actually claiming you expect to see a perfect mix of winning seeds, in a time span of 6 years, 8 times in 100 years! Stop and think how ridiculous that sounds.

Problem is that's not what I am saying. I'm using terms and vernacular which you clearly don't understand. I am far and away more educated than you on the concepts of hypothesis testing, how to prove things statistically, etc.

I used to do hypothesis testing for a living (but now work in predictive modeling). I have used the assumptions you provided. I have performed a statistical analysis - a real life, grown up statistical analysis, not just saying "wow that's crazy!" The results are indisputable to any educated individual who knows what they're talking about. Your evidence taken from that 7 year span does not rise to the level of statistical proof, period, end of story.

 

[PatriotSeven]

Problem is that's not what I am saying. I'm using terms and vernacular which you clearly don't understand. I am far and away more educated than you on the concepts of hypothesis testing, how to prove things statistically, etc.

I used to do hypothesis testing for a living (but now work in predictive modeling). I have used the assumptions you provided. I have performed a statistical analysis - a real life, grown up statistical analysis, not just saying "wow that's crazy!" The results are indisputable to any educated individual who knows what they're talking about. Your evidence taken from that 7 year span does not rise to the level of statistical proof, period, end of story.

The problem is you brag about being educated, having a diploma, a program and you are not even capable of answering a straight question.

The chances of a 6 string number, 0-6, non-repeatable, happening in a consecutive 6 year time span, is not 8%. It's 1.5%. In addition if the probability is skewed for certain digits to appear more often, which in our case it is, then your numbers should be far lower than that.

I have enough knowledge to know you are wrong. You claim your program spits out some number despite the fact you have shown no proof to back it up.

In addition a simple self-check should tell you, that if a #6 seed only has a 6% chance of showing up, then a 6 digit string, of unique numbers is far, far, far, lower than 8%.

 

[SB39]

The problem is you brag about being educated, having a diploma, a program and you are not even capable of answering a straight question.

I don't see a question. I see a bunch of ignorant statements made by someone who doesn't even have an high school student's understanding of statistics, but no question. If you have a question that is not based on a faulty assumption then I will be happy to answer it.

 

[PatriotSeven]

I don't see a question. I see a bunch of ignorant statements made by someone who doesn't even have an high school student's understanding of statistics, but no question. If you have a question that is not based on a faulty assumption then I will be happy to answer it.

I've already addressed the point multiple times, explained it logically so that even someone with a highschool understanding of statistics can figure out your numbers are incorrect, and considering your ignorance to that fact is as egregious as your results and your insults, there's really not much of a point to this conversation.

You clearly don't want to admit you were wrong.

 

[SB39]

I've already addressed the point multiple times, explained it logically so that even someone with a highschool understanding of statistics can figure out your numbers are incorrect, and considering your ignorance to that fact is as egregious as your results and your insults, there's really not much of a point to this conversation.

You've demonstrated repeatedly that you have no clue what you are talking about. You harp about your 1.5% number as if that demonstrates anything whatsoever (newsflash: it doesn't). You don't understand the first thing about statistics, let alone advanced concepts like hypothesis testing, statistical proof, the importance of sample size, p-values and confidence intervals.

(NOTE: I don't consider hypothesis testing, p-values and confidence intervals to be "advanced concepts" but they are certainly light years ahead of anything you understand).

 

[JoeSixPat]

Suggs also thinks Goodell ordered the hits on Hoyer and Manuel tonight.

 

[PatriotSeven]

You've demonstrated repeatedly that you have no clue what you are talking about. You harp about your 1.5% number as if that demonstrates anything whatsoever (newsflash: it doesn't). You don't understand the first thing about statistics, let alone advanced concepts like hypothesis testing, statistical proof, the importance of sample size, p-values and confidence intervals.

(NOTE: I don't consider hypothesis testing, p-values and confidence intervals to be "advanced concepts" but they are certainly light years ahead of anything you understand).

1.5% represent the chance of finding a unique 6 digit combination, with non-repeatable digits(1,2,3,4,5,6) in any order(6,2,3,4,5,1), possible from 6 seeds with an equal chance of advancement. Which means each number has an equal chance of being randomly generated. Which means it also allows for repeatable digits(1111111).

n!/n^n = 6!/6^6 = 1.5%

Ok so if you don't know how to calculate it when the seeds have a variable chance of appearance(seed probability) say so.

But "your hypothesis test" for whatever formula you used for your 8%, can easily be disproven by the simple fact you're not going to even see a 6 seed appear more than 6%, let alone a unique string of 6 digits.

 

[Joker]

I accuse YOU, Joe SixPat!

 

[Joker]

"Suggs also thinks...."


uh...no, no he doesn't...that would require brain cells

 

[SB39]

1.5% represent the chance of finding a unique 6 digit combination, with non-repeatable digits(1,2,3,4,5,6) in any order(6,2,3,4,5,), possible from 6 seeds with an equal chance of advancement. Which means each number has an equal chance of being randomly generated. Which means it also allows for repeatable digits(1111111).

n!/n^n = 6!/6^6 = 1.5%

You are saying all 6 seeds have an equal chance so that is the assumption I am proceeding with in this post. So now here's the thing that is going to blow your mind:

You're right that there is a 1.5% chance that all 6 seeds win it once in a 6-year span. But guess what??? Even though there is only a 1.5% chance that all 6 seeds win it exactly once, that is still the most likely outcome...!!! That is still, statistically speaking, the most expected outcome. And not a single intelligent human being would ever claim that observing the most expected outcome is evidence of bias.

Let me ask you a very simple question (which I know you won't have the guts to answer): If you rolled a 6-sided die 6 times and got one 1, one 2, one 3, one 4, one 5 and one 6 (in any order), would you say "holy crap! The probability of that happening is only 1.5% so the die must be rigged!!!!!"?