Find the probability that a random variable is "good" and find the expected value of the number of "good" random variables.

65 Views Asked by At

Suppose that $X_1,\cdots,X_n$ are i.i.d. continuous random variables. $X_k$ is called good if we have $X_i < X_k$ for all $i<k$.

1) Find the probability that $X_k$ is good.

2) Find the expected value of the number of good random variables. Show that this value approaches infinity when $n \to \infty$.

Please do not give me a full solution first. Just hints. My major issue with this problem is that I do not understand what the question is asking mathematically. So, a random variable is just a function from some probability space to real numbers. What kind of probability density are we assuming on the space of random variables? What is our probability space here?

1

There are 1 best solutions below

2
On BEST ANSWER

If we sort the variables $X_1\dots X_k$ in increasing order, there are $k!$ possible orders. Only the ones where $X_k$ comes last make $X_k$ good. There are $(k-1)!$ such arrangements. Since, by symmetry, all orders are equally likely, the probability that $X_k$ is good is $${(k-1)!\over k!}={1\over k}$$

We ignore the possibility that two of the variables have the same value. Since the X_i are continuous r.v.'s, the probability that two of them are equal is $0.$

The expected number of good $X_k$ is just $$\sum_{k=1}^\infty P(X_k\text{ is good})=\sum_{k=1}^\infty{1\over k} = \infty$$