10
$\begingroup$

Let $X$ and $Y$ be two independent random variables having the same uniform distribution $U(0,1)$ with density

$f(x)=1$ if $0≤x≤1$ (and $0$ elsewhere).

Let $Z$ be a real random variable defined by:

$Z=X-Y$ if $X>Y$ (and $0$ elsewhere).

  1. Derive the distribution of $Z$.

  2. Compute the expectation $E(Z)$ and variance $V(Z)$.

Improve this question
$\endgroup$
7
  • 3
    $\begingroup$ Homework? What have you tried and where are you stuck? Do you know how to find the distribution of a sum of independent random variables. If you do then, hint: $X - Y = X + (-Y)$. That said, your question doesn't seem to be asking about the distribution of a (pure) subtraction. So, providing some details on your thought process will help users here guide you in the right direction. $\endgroup$ – cardinal Nov 9 '11 at 17:44
  • $\begingroup$ i am preparing for an exam after leaving university for 5 years and working in a totally different field that has nothing to do even with numbers. $\endgroup$ – Majed Hijazi Nov 9 '11 at 18:15
  • 2
    $\begingroup$ The distribution is more than just the mean and variance, so you should review the distinction among the three. Then consider relying on first principles. E.g., drawing a picture of the joint distribution of $(X,Y)$ in the $x,y$-plane along with level curves of $Z=X-Y$ will provide an immediate (and easy) geometric derivation of the distribution of $Z$. $\endgroup$ – whuber♦ Nov 9 '11 at 18:29
  • 1
    $\begingroup$ Is this exactly how the problem is originally stated? Or does the problem ask for the conditional distribution of $Z$ given that $X\gt Y$? $\endgroup$ – whuber♦ Nov 9 '11 at 18:34
  • 3
    $\begingroup$ Hint: Since $P\{X < Y\} = \frac{1}{2}$ (think about why this must be so), $Z$ has value $0$ with probability $\frac{1}{2}$. Thus, $Z$ is what is sometimes called a mixed random variable which takes on some values with nonzero probability and behaves as a continuous random variable for some values. As @whuber does, I too ask whether you have mis-stated the problem. It leads to more complications than one would expect from a typical end-of-chapter problem at the apparent level of the book you are using. $\endgroup$ – Dilip Sarwate Nov 10 '11 at 12:50
5
$\begingroup$

Check p. 18 of Probability Distributions as Program Variables, by Dimitrios Milios.

He has discussed the problem in an fairly detailed manner.

Improve this answer
$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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