I Have The Power! (To Decide The Madden Cover)

[fieryprophet]

I don't disagree with that within the context of your system, but the whole reason athletes are considered elite, and thereby susceptible to a "curse" in the first place, is because their performances have shown to not be entirely random but predictable and consistently measuring as superior to other subjects. For this reason, one can also look at the drops in relative production as well as the also chances of injury to cull significance as well. Part of the issue is that most Madden athletes are selected after an elite season, making a similar achievement difficult. However, one would need to really figure out if the degree of statistical drops in key categories (TDs, yards, etc.) are significant as compared to a standard year-over-year variability determined by many players to really get a feel for significance.

To compare the performance of an athlete, who has already shown themselves to be able to continually "buck the odds" so to speak and perform at a high level throughout multiple samplings, to an infinite number of possibilites is logical but not entirely appropriate in this context, as history has shown that past performance can indeed be indicative of future expectations. If you're going to weight probabilities, you have to weight the ones that indicate the ability to excel as well.

Again, however, the sampling itself is statistically biased, because the cover athletes chosen are almost always coming off a career year that sheer regression to the mean will look like "failure" compared against it. The only way the curse could possibly be correlated with failure (and remember, correlation does NOT equal causation, which is a separate topic entirely) is if the cover athlete was randomly chosen at a random point in the prime of their career (when they should be most likely to be performing at a high level.)

Since they are not, and since the criteria for failure is so multi-faceted (with injuries, statistical regression, personal failings and dumb luck all playing major roles in the definition) and success is poorly described (again, how does Drew Brees PRO BOWL season NOT qualify as a success?) the entire discussion is extremely skewed towards "curse" and far from reality.

[KSpan]

But you have to consider the reason athletes are on the cover of Madden in the first place, and that is largely because they are coming off of a great year that, whether they land the cover or not, will be hard to duplicate.

There's a reason that guys like Tom Brady and Peyton Manning are rarely on the cover, and that's because they have long and consistently successful careers, rather than coming out of nowhere and grabbing the kind of attention that lands you the Madden cover.

Compare it to Tiger Beat Magazine. How many flash in the pan, 1 hit wonders have they had on their cover over the years? Hanson, Backstreet Boys, etc. Did their careers fizzle because of some "Tiger Beat Curse," or were they on the cover of Tiger Beat precisely because they were the latest sensation likely destined to fizzle out?

Covered that in this part of my post:

"Part of the issue is that most Madden athletes are selected after an elite season, making a similar achievement difficult. However, one would need to really figure out if the degree of statistical drops in key categories (TDs, yards, etc.) are significant as compared to a standard year-over-year variability determined by many players to really get a feel for significance."

My angle is that if sports were a truly random system, there would be no real difference between the Peyton Mannings and Curtis Painters of the world, and that's simply not the case. However, Fiery is right in that there are a TON of variables that go into it and it would be nigh impossible to get a 100% accurate formula in place. It's kind of like pharmacokinetics, or how your body handles drugs - there is no way to possibly account for every single factor and individual, so sometimes the most economical and effective approach is to just treat it as a binary unit instead of a million moving parts.

And I still vote for Cam.

[KSpan]

Again, however, the sampling itself is statistically biased, because the cover athletes chosen are almost always coming off a career year that sheer regression to the mean will look like "failure" compared against it. The only way the curse could possibly be correlated with failure (and remember, correlation does NOT equal causation, which is a separate topic entirely) is if the cover athlete was randomly chosen at a random point in the prime of their career (when they should be most likely to be performing at a high level.)

Since they are not, and since the criteria for failure is so multi-faceted (with injuries, statistical regression, personal failings and dumb luck all playing major roles in the definition) and success is poorly described (again, how does Drew Brees PRO BOWL season NOT qualify as a success?) the entire discussion is extremely skewed towards "curse" and far from reality.

The joy of statistics - the answer always depends on the question :) It's nice to geek out over this stuff every now and again.

[fieryprophet]

The joy of statistics - the answer always depends on the question :) It's nice to geek out over this stuff every now and again.

It's fun to think about, until you realize that the fundamental inability to understand basic statistical correlation and the nature of randomness accounts for the multitudes of lotteries, casinos, economic bubbles and crashes, and total political stupidity around our world, then it's not so much fun anymore. Seriously, human beings could not have been cursed with a more debilitating cognitive advantage; our inclination to see "patterns" helps us navigate a dangerous world, but leaves us looking and believing in patterns that have no basis whatsoever in reality, with dire consequences.

[TruCatzFan]

Fieryprophet is my ALL TIME FAVORITE HUDDLER!

[SgtJoo]

It's fun to think about, until you realize that the fundamental inability to understand basic statistical correlation and the nature of randomness accounts for the multitudes of lotteries, casinos, economic bubbles and crashes, and total political stupidity around our world, then it's not so much fun anymore. Seriously, human beings could not have been cursed with a more debilitating cognitive advantage; our inclination to see "patterns" helps us navigate a dangerous world, but leaves us looking and believing in patterns that have no basis whatsoever in reality, with dire consequences.

Totally this.

Posted this link before, but people seriously need to read it.

http://en.wikipedia.org/wiki/Texas_sharpshooter_fallacy

[jtnc]

If Cam Newton ends up on the Madden cover, he will be on my 2012 fantasy bust list. High expectations + curse = trouble.

https://twitter.com/#!/Michael_Fabiano/status/190864787033423872

[KendrickPanther]

Ok I read the comments on ESPN for like 20 seconds. Curse be damned I want those whiny San Francisco fans to lose. I forgot how much I hated San Fran with realignment.

[Tarheels23]

Every 1 Matters!

[KSpan]

Again, however, the sampling itself is statistically biased, because the cover athletes chosen are almost always coming off a career year that sheer regression to the mean will look like "failure" compared against it. The only way the curse could possibly be correlated with failure (and remember, correlation does NOT equal causation, which is a separate topic entirely) is if the cover athlete was randomly chosen at a random point in the prime of their career (when they should be most likely to be performing at a high level.)

Since they are not, and since the criteria for failure is so multi-faceted (with injuries, statistical regression, personal failings and dumb luck all playing major roles in the definition) and success is poorly described (again, how does Drew Brees PRO BOWL season NOT qualify as a success?) the entire discussion is extremely skewed towards "curse" and far from reality.

From a strictly statistical standpoint, year in and year out player statistics and performances can be charted into a bell curve, which tells you exactly what it will take in any given year to have a statistically-significant performance as compared to a closed and complete sampling of every other player that meets a given criteria of entry (say, 25 carries or 25 pass attempts) - in this sense, sports statistics are a bit unique because one can always sample 100% of the subject population. A null hypothesis of "Athletes appearing on the cover of Madden suffer a statistically-sgnificant drop in their performance as compared to prior years" or "Athletes appearing on the cover suffer... compared to the rest of NFL players in their category" does not imply causation, but is a perfectly relevant statement that the data can accurately reject or fail to reject.

By sampling every player in the performance category, all possible random outcomes that actually occurred for a given year are accounted for because, unlike other systems, the sample size is fixed. Plotting the data mentioned above and testing changes for statistical significance compared to the average change observed in the entire system is relevant because all players have the same exposure to randomness - in this sense, randomness is a common variable and can be an inferred factor, as testing against the entire system accounts for trands observed across the entire league (i.e. receiver numbers jumping due to a rules change).

In this type of test system statistical significance fails to reject the null, and someone choosing to make a faulty and unproven attribution does not impact the relevance of the analysis as it relates to the null.

[KSpan]

From a strictly statistical standpoint, year in and year out player statistics and performances can be charted into a bell curve, which tells you exactly what it will take in any given year to have a statistically-significant performance as compared to a closed and complete sampling of every other player that meets a given criteria of entry (say, 25 carries or 25 pass attempts) - in this sense, sports statistics are a bit unique because one can always sample 100% of the subject population. A null hypothesis of "Athletes appearing on the cover of Madden suffer a statistically-sgnificant drop in their performance as compared to prior years" or "Athletes appearing on the cover suffer... compared to the rest of NFL players in their category" does not imply causation, but is a perfectly relevant statement that the data can accurately reject or fail to reject.

By sampling every player in the performance category, all possible random outcomes that actually occurred for a given year are accounted for because, unlike other systems, the sample size is fixed. Plotting the data mentioned above and testing changes for statistical significance compared to the average change observed in the entire system is relevant because all players have the same exposure to randomness - in this sense, randomness is a common variable and can be an inferred factor, as testing against the entire system accounts for trands observed across the entire league (i.e. receiver numbers jumping due to a rules change).

In this type of test system statistical significance fails to reject the null, and someone choosing to make a faulty and unproven attribution does not impact the relevance of the analysis as it relates to the null.

Accidentally hit quote instead of ediit... apparently, posts can not be deleted. Learn something new every day.

[Zod]

All that server power and you are using it to vote on video games.

I'd be a porn producer millionaire .

[Cary Kollins]

1. This thread proves not every huddler is an inbred dumbass.

2. I hope some snitch doesn't run and alert the authorities.

[Doc Holiday]

3 things in the world that you don't fug with, Chuck Norris, Sasquatch and the Madden curse.

You guys have no idea the powers that you are dealing with...................