Statistical anomoly?

[qjakal]

Someone sent this to me today...looking for feedback and comments from the numerical side of the TMF on your feelings as to its validity...

"If you toss a penny 10000 times, it will not be heads 5000 times, but more like 4950. The heads picture weighs more, so it ends up on the bottom more."


I soooooo hope this is false..... Q

 

[guitman69]

When you flip a coin, the odds are that it will turn up heads is 1:2 or 50%. This does not mean that after flipping 10000 times it will come up heads exactly 5000 times. Giving the odds that something will happen does not mean it WILL happen with that regularity. You could end up with it only coming up heads 1/3 of the time. Remember, this is a random event. There is NO way to predict a random event. You can only give a stastical probability of that event.

If you could predict a random event through math, then the weather man would be right all the time.

 

[qjakal]

Yup...but nope...lol

No argument from me <G>, but this gent is claiming that there is a variance in the proportions of weighting, which will affect the "odds" due to the change in the actual physics of the act of flipping the object....hence, the statistical imbalance I'm concerned about. Perhaps i should be looking for a coin collector, or someone with the proper equipment to measure something this small accurately.... Q

 

[guitman69]

Well, if you want to get down to the milligrams, you could also say that a pair of dice is more likely to land 6 up because the 1 on the opposite side has less material removed to make up the numbers and is therefore heavier. If something this small changed the probability, it would surprise me if it was more than a thousanth of a persent or less.

 

[Frink]

The assumptions for these probability problems always begin with, "Assuming a fair die..." (for dice calculations) or "assuming a fair coin..." (for coin-tosses) or "assuming a fair, randomly-shuffled deck..." (for playing cards calcs), etc. None or these "fair" assumptions actually hold 100%, but they tend to be close enough for all practical purposes.

These types of probablilities are also to be interpreted as being taken "in the limit to infinity." In other words, your "tosser" might toss the coin 5000 times and get 4950 heads, but he could toss it 100,000 times and get 49,992 heads....and after 1 billion times get 0.5 billion minus 2 heads....and so on. So for the probability to equal 0.5 (exactly) he'd have to make a "fair toss" of the "fair coin" an infinite number of times.

Basically, given a *real* coin is not actually 100% fair (it cannot be, but it's generally VERY close), and a "toss" is not 100% fair (or repeatable), and you're only (realistically) going to toss it a relative few times, without knowing the outcome beforehand, I'd say just call it a completely random toss (P(Heads)=0.5). lol

As an aside.....If you're worried about being cheated, I'd worry MUCH more about how the tossing is being done than anything. If a human's doing it, it's way too easy to skew results for repeated attempts--especially if it's intentional.

😎

 

[Madman]

I once flipped a penny -- it bounced, rolled, and stopped. On its edge! I guess that buggers the whole thing up now, doesn't it? 🙂

Madman

 

[JoBelle]

Reminds me of the opening scene to ROSENCRANTZ AND GUILDENSTERN ARE DEAD! A wicked cool movie!!!

Anyone know it?
Joby

 

[guitman69]

I once flipped a penny -- it bounced, rolled, and stopped. On its edge! I guess that buggers the whole thing up now, doesn't it? 🙂

Madman

It's times like that I'll bet you wish you had a video camera rolling!

 

[venray]

Just follow the link.........

http://www.sciencentre.qld.gov.au/sciencentre/exhibits/permanent/math/toss.htm


...and click the button 10,000 times and the question will be answered.......

 

[Dave2112]

There was an old question in one of my math classes about probability. "If you flip a coin, and it comes up heads nine times, what is the probability that the tenth flip will be heads?"

The answer, of course, is 50-50. You are reffering to the final flip, not the nine preceeding it. Any individual flip has a 50-50 chance either way, but the odds differ for more flips. The chances of one flip out of 1,000 are different than the 1,000 flips as a whole.

😎

 

[Cyrano]

Ah, don't worry about these sorts of things, I always bet on tails and I never bet the roller is going to crap out. And so all I can do is pray luck be a lady tonight.

 

[MadKalnod]

Well, if you want to get down to the milligrams, you could also say that a pair of dice is more likely to land 6 up because the 1 on the opposite side has less material removed to make up the numbers and is therefore heavier.

Tell that to the contents of my dicebag, and make them explain all those critical fumbles in Dungeons & Dragons and Warhammer 40,000!

 

[Alex Warfield]

When you flip a coin, the odds are that it will turn up heads is 1:2 or 50%. This does not mean that after flipping 10000 times it will come up heads exactly 5000 times. Giving the odds that something will happen does not mean it WILL happen with that regularity. You could end up with it only coming up heads 1/3 of the time. Remember, this is a random event. There is NO way to predict a random event. You can only give a stastical probability of that event.

If you could predict a random event through math, then the weather man would be right all the time.

Actually, I can predict a random event....Of course that doesn't mean I'll be right!