Silicon Investor (SI) -- The First Internet Community

STOCKTALK

We've detected that you're using an ad content blocking browser plug-in or feature. Ads provide a critical source of revenue to the continued operation of Silicon Investor. We ask that you disable ad blocking while on Silicon Investor in the best interests of our community. If you are not using an ad blocker but are still receiving this message, make sure your browser's tracking protection is set to the 'standard' level.

Politics : Formerly About Advanced Micro Devices -- Ignore unavailable to you. Want to Upgrade?

To: jcholewa who wrote (128011)	11/10/2000 1:11:03 PM
From: Scumbria	Read Replies (2) \| Respond to of 1570913

JC, Thanks for the math lesson. It makes sense to me. Scumbria

To: jcholewa who wrote (128011)	11/10/2000 2:30:00 PM
From: andreas_wonisch	Read Replies (2) \| Respond to of 1570913

JC, Re: I believe this is pretty easy. The chances of flipping 1500 heads where all other flips can be any combination of heads and tails should be something like: 0.5 ^ 1500 x 1.0 ^ 500 = 0.5 ^ 1500 = 2.85 x 10^(-452) No, that's not right. With your solution the chances of getting more than one head would be 0.5, more than two 0.25, more than three 0.125. The problem with your solution is that you compute the probability for getting 1,500 head first and then any combination of heads and tails. You don't consider that heads can be arranged in any order. The correct approach to this problem is with the help of the binomial distribution. The probability of getting exactly k heads is: p(X=k) = (n over k) * p^k * q^(n-k) In this case the probability is p = q = 0.5 since head and tails are equally probable. Further we have n = 2000. But we want to get the probability of getting more than k heads. So we need to sum up: p(X>=k) = sum_{k=0}^{n} p(k) = (n over k) * p^k * q^(n-k) It's very difficult (from a computation standpoint) to sum up all these small numbers. But since 2000 is a fairly big n we can approximate the binomial distribution with the gaussian distribution. With the normalized gaussian distribution we have: P(X>=k) = 1 - Phi[(k - np)/(sqrt(npq))] where Phi(z) is int_-z^z 1/sqrt(2pi) * e^(-x^2/2) dx. With k = 1500 this probability is so extremely small that it comes very close to your number. With k = 1000 it's -- like expected -- 0.5. The probablity for k = 1100 would already be below 0.01. Andreas

To: jcholewa who wrote (128011)	11/10/2000 2:35:35 PM
From: Ali Chen	Read Replies (1) \| Respond to of 1570913

JC, <The chances of flipping 1500 heads where all other flips can be... Apologies if I misdid the math. I think I have it more or less correct. :)> Apology accepted. :) :) If you slightly change the initial condition to 1000 heads (which apparently does not change the reasoning behind your formula), you will still get the miserable probability of (1/2)^1000=10^(-300), while in reality the result is 1. Regards, - Ali