Trouble with this probability identity: $\operatorname{E} \bigl[ (Z - E[Z]) \bigr]^2 = \operatorname{var}(X) + \operatorname{var}(Y)$

49 Views Asked by At

My professor recently surprised us with a pop quiz on probability, and I'm struggling to prove a particular identity involving random variables. Suppose we have two independent random variables, $X$ and $Y$, and define another random variable $Z$ as $Z = X - Y$. I am trying to prove the following expression: $$ \operatorname{E} \bigl[ (Z - E[Z]) \bigr]^2 = \operatorname{var}(X) + \operatorname{var}(Y) $$

The presence of the square outside the expected value is confusing me, even though I confirmed with my professor that it should be there. Can someone guide me through this proof? Any help would be greatly appreciated. Thank you!

1

There are 1 best solutions below

1
On

The square outside is definitely wrong, because $E(Z-E(Z))^2=(E(Z)-E(E(Z))^2=(E(Z)-E(Z))^2=0$ which can only be equal to $Var(X)+Var(Y)$ if both $X$ and $Y$ are almost surely constant. If it is inside, we have \begin{align*}E((Z-E(Z))^2)&=E((X-Y)^2-(E(X-Y))^2)\\ &=E((X-Y)^2)-(E(X-Y))^2\\ &=E(X^2-2XY+Y^2)-(E(X)-E(Y))^2\\ &=E(X^2)-2E(XY)+E(Y^2)-(E(X)^2-2E(X)E(Y)+E(Y)^2)\\ &=Var(X)+Var(Y)-2(E(XY)-E(X)E(Y))\\ &=Var(X)+Var(Y), \end{align*} where the last equality holds because a family of independent random variables is uncorrelated.