I have two independent random variables X and Y with given pmf. I want to calculate joint probability of W=X+Y and Z=XY, is the below formula correct for calculating it?

$P(W=w,Z=z) = \sum_{k}P(X=k)P(Y=w-k \ \land Y = \frac{z}{k})$

and by this, I want to show whether these two RV are dependent or independent.

I have to mention that I already calculate pmf for both W and Z.


The formula is "almost" correct. It is not unthinkable that $P(X=0)>0$ and in that case $k$ ranges over a set that contains $0$. That gives troubles because $\frac{z}{k}$ is part of the expression.

With more caution it can be deduced that:

$$P\left(W=w,Z=z\right)=P\left(X+Y=w,XY=z\right)=\sum_{k}P\left(X=k\right)P\left(X+Y=w,XY=z\mid X=k\right)=$$$$\sum_{k}P\left(X=k\right)P\left(k+Y=w,kY=z\mid X=k\right)=\sum_{k}P\left(X=k\right)P\left(Y=w-k,kY=z\right)$$where the last equality is based on independence. Observe that $Y=\frac{z}{k}$ is replaced by $kY=z$.

Further note that for almost every $k$ the event $\{Y=w-k,kY=z\}$ is empty, so that you can hardly speak of "summation" because almost all terms are $0$.

A route that avoids the summation is:$$P\left(W=w,Z=z\right)=P\left(X+Y=w,XY=z\right)=P\left(X\left(w-X\right)=z\right)$$ You know the PMF of $X$ and based on that knowledge you can find the PMF of $X(w-X)$ for every $w$.

  • $\begingroup$ Thanks for your explanations. $\endgroup$ – Peyman Tahghighi Feb 22 at 9:09
  • $\begingroup$ You are welcome. $\endgroup$ – drhab Feb 22 at 9:10

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.