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.

I have three independent uniforms variables, $U, V, W$ that take their values in $[0,1]$. I also have $X = \max(U, W)$ and $Y = \max(V,W)$.

I need to compute the conditional expectation of $W$ knowing (the $\sigma$-algebra generated by) $X$ and $Y$. I have proven that it can be written as $f(X,Y)$ with $f$ measurable.

Is there any way to compute this without too much calculations? I have already computed $E(W \mid X)$ but I can't find any simple way to use this. Do I have to make the full calculation?

share|improve this question
add comment

1 Answer

up vote 1 down vote accepted

Here is a hands-on approach. One looks for the mean value of $W$ conditionally on $(X,Y)=(x,y)$, with $x$ and $y$ in $[0,1]$.

  • If $x=y$, then almost surely $U\lt W$ and $V\lt W$, in which case $W=x$.

  • If $x\lt y$, either $U\lt W\lt V$ (case 1), or $W\lt U\lt V$ (case 2).

    • In case 1, $(X,Y)=(W,V)$ hence $W=x$.
    • In case 2, $(X,Y)=(U,V)$ and $W$ is conditioned to be in the interval $[0,U]$ hence the conditional mean of $W$ is $\frac12U=\frac12x$.
    • Cases 1 and 2 are equiprobable by exchangeability of $(U,V,W)$ hence when $x\lt y$, the conditional mean of $W$ is $\frac12(x+\frac12x)=\frac34x$.
  • If $y\lt x$, by symmetry the conditional mean of $W$ is $\frac34y$.

  • Hence, when $x\ne y$, the conditional mean of $W$ is $\frac34\min\{x,y\}$.

Finally, $$ \mathbb E(W\mid X,Y)=\tfrac14\min\{X,Y\}\cdot(3+\mathbf 1_{X=Y}). $$

share|improve this answer
add comment

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.