Define $Z = \min(X, Y)$ and the joint pdf of $X$ and $Y$ as $f_{XY}(x,y)$.

I saw an approach that said

$$ E[Z] = \int \int \min(x,y) f_{XY}(x,y) \,dy\,dx $$

Is this readily obvious, or do you need to convert the following: $$ E[Z] = \int \min(x,y)f_Z(z) \,dz $$

to the above?

  • $\begingroup$ Obvious. For any function $g(x,y)$, the expected value of $g$ is $\int \int g f_{XY}dxdy$$. $\endgroup$
    – YJT
    Apr 19, 2021 at 13:32
  • $\begingroup$ @YJT Is this the law of unconscious statistician? If so, I've only seen that for the one variable case, but it seems it applies to any number of variables? $\endgroup$
    – David
    Apr 19, 2021 at 13:34
  • $\begingroup$ @David Yes. Consider the (generically 2-to-1) change of variables $(Z,W)=(\min(X,Y),\max(X,Y))$. $\endgroup$ Apr 19, 2021 at 13:42
  • $\begingroup$ Yes.it applies to more than one variable. $\endgroup$
    – YJT
    Apr 19, 2021 at 13:42

2 Answers 2


You do not need to do that. In general you have $$ \operatorname E(g(W)) = \int_{\mathcal W} u\cdot f_{g(W)}(u)\, du = \int_{\mathcal X} g(x)f_W(x)\, dx. $$ Using the second integral is often simpler because you are given $f_W$ and there is no need to find $f_{g(W)}.$

Some people call this the "law of the unconscious statistician," and if you google that term you will find it.


The two definitions are the same, but I think that in a real problem you have to use the second one....

For example, let $X,Y$ be i.i.d. rv's following a $Exp(1)$ distribution.

to find

$$E(Z)=\int_0^{\infty}2t e^{-2t}dt=\frac{1}{2}\int_0^{\infty}w e^{-w}dw=\frac{1}{2}$$

  • $\begingroup$ This is quite incorrect. There is not in fact a need to find the density function of $Z. \qquad$ $\endgroup$ Apr 19, 2021 at 15:39

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.