102 replies to this topic

[BBQ&Beer]

Science FTW!

You get pie just for Billy Dee.

[fieryprophet]

Science FTW!

Randomness is one my pet peeves because many people find it very difficult to understand something so fundamentally simple.

My favorite misconception is the gambler's fallacy: the belief that the outcome of a series of unrelated events influences the outcome of a future one. To jump back on the example of a coin, if you flipped a coin twenty times and it landed on heads every single time, then asked the average person what the likelihood of it landing on heads on the next flip was, they would either answer:

1) More likely to land on heads, because it had landed on heads the previous 20 flips, or

2) More likely to land on tails, because it was "due" to land tails because of all the landings on heads before.

The correct answer is (for an unbiased coin, obviously) that every flip has a 50/50 chance no matter what the previous 20 results were, because each flip is an unrelated event on which prior outcomes have no bearing.

But try telling that to a gambler sitting at a slot machine they've sunk a hundred bucks into without winning and thinking they're "due" for a jackpot any spin now.

[Miguel O'Hara 2.0]

Randomness is one my pet peeves because many people find it very difficult to understand something so fundamentally simple.

My favorite misconception is the gambler's fallacy: the belief that the outcome of a series of unrelated events influences the outcome of a future one. To jump back on the example of a coin, if you flipped a coin twenty times and it landed on heads every single time, then asked the average person what the likelihood of it landing on heads on the next flip was, they would either answer:

1) More likely to land on heads, because it had landed on heads the previous 20 flips, or

2) More likely to land on tails, because it was "due" to land tails because of all the landings on heads before.

The correct answer is (for an unbiased coin, obviously) that every flip has a 50/50 chance no matter what the previous 20 results were, because each flip is an unrelated event on which prior outcomes have no bearing.

But try telling that to a gambler sitting at a slot machine they've sunk a hundred bucks into without winning and thinking they're "due" for a jackpot any spin now.

But, but, its a CURSE!!! They defy all logic! Black magic!!!

Im with ya dude. Maybe if every Madden cover athlete got hurt in the exact same way, maybe then i might give it a second look, maybe. Possibly. The fact that each case can be argued logically, imo, disproves any notion of a curse.

[Pantherman]

Cam is Superman. He ain't scared of no cover curse. Cam will send the Madden Cover Curse packing back to John Madden himself and say shove it up your arse.

[rodeo]

I am very aware of that. I run a software company. I have an entire server farm at my disposal.

Then carry on, sir. Just try to keep it to a realistic margin.

[fieryprophet]

The margin might actually be an issue, lol. Anyone have any good numbers on how many votes this vote has been averaging? I'd like to keep it realistic if possible. . .

[KSpan]

I have written a pseudo-random number generator for a custom cryptographic suite. One of the stringent demands on such a system is the ability to create numbers that are as absolutely random as possible. What people don't understand about randomness is that if you flip a coin 20 times and it comes up heads 17 times, that is a perfectly logical random result. The 50/50 distribution only has to hold true when dealing with a massive number of permutations (on the order of millions) but a random number generator that produced a steady series of heads/tails combinations in order would be viewed as functionally broken.

In the same regards the Madden curse, which has multiple "bad" definitions but very few good ones (so the probability of a bad result is closer to 90% than 50%) having a run of 9-12 "bad" years out of 12 is not only statistically insignificant, it's entirely flawed by the faulty weighting of the probabilities involved (I mean, who considers Drew Brees Pro Bowl season to have been cursed? Idiots, that's who.)

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.

[haccess]

The margin might actually be an issue, lol. Anyone have any good numbers on how many votes this vote has been averaging? I'd like to keep it realistic if possible. . .

Clicking on previous rounds shows their numbers, usually about 450k total votes in each matchup.

[Ken]

The margin might actually be an issue, lol. Anyone have any good numbers on how many votes this vote has been averaging? I'd like to keep it realistic if possible....

In the 1st round Cam beat Blount with 87% there was 575,908 total votes
In the second he beat Gates with 83 % of the 454, 770 votes
Last round he beat Fitz with 70% of the 447,395 votes.

As of this morning he had leading with 52%.

[rodeo]

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.

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?

[SOJA]

being on the cover of Madden literally has nothing to do with performance, its called regression to the mean

[fieryprophet]

In the 1st round Cam beat Blount with 87% there was 575,908 total votes
In the second he beat Gates with 83 % of the 454, 770 votes
Last round he beat Fitz with 70% of the 447,395 votes.

As of this morning he had leading with 52%.

That few?!? Holy crap, I better shut it down before I lap the field. . .

[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.