Take the 2-minute tour ×
Mathematics Stack Exchange is a question and answer site for people studying math at any level and professionals in related fields. It's 100% free, no registration required.

Let's say I have a probability distribution $F(x)$, which is a cdf. Then, I have a value $y$ at some place on the support of $F$. Then, there will be some number $n$ random independent draws from $F$, and I want to know the probability that at least one of them is $> y$.
I would think that to find this I would do $n(1-F(y))$, since the pr. of any one draw being $> y$ is $1-F(y)$. But say $F(y)$ is $.5$. Then, is $n=3$, then I would get $1.5$, which of course doesn't make any sense.

What am I doing wrong?

share|improve this question
3  
Try to think about the complementary event: what is the probability that all of the values drawn are smaller (or equal) than $y$? –  Fabian Jan 9 '12 at 19:58
1  
If all the values are smaller, then the pr is $F(y)^n$... so I guess the answer is $1-F(y)^n$ ... thanks, @Fabian! (hopefully I'm not still wrong...) –  Ac3 Jan 9 '12 at 20:01
    
you are welcome... –  Fabian Jan 9 '12 at 20:02
    
Ac3: You could write this as an answer and possibly, a little later on, accept it. –  Did Jan 9 '12 at 20:07
    
I think the credit should go to Fabian... but if s/he doesn't add an answer after a while, I will write one so I can close this. –  Ac3 Jan 9 '12 at 20:11

1 Answer 1

up vote 1 down vote accepted

(concluding the comments to the original question)

The problem is rather easy if one thinks about the complementary event: what is the probability that all of the values drawn are smaller (or equal) than $y$? The probability that a single draw is smaller than $y$ is given by $F(y)$. As the $n$ draws are independent the total probability for the complementary event is $F(y)^n$.

Concluding, the probability that at least one of the draws is larger than $y$ is given by $$1- F(y)^n$$ (which of course is always between 0 and 1).

share|improve this answer

Your Answer

 
discard

By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.