I have a joint distribution of 2 random variables $X$ and $Y$: $$ f(x, y) = \frac{3}{2}x + \frac{y}{2} $$ if $x,y \in [0;1]$ and $0$ otherwise. What's the following probability:$$ P\left(XY < \dfrac{1}{2}\right) $$ I thought about sum of probabilities: $P(X < \frac{1}{2Y}) + P(Y < \frac{1}{2X})$, but appropriate integrals don't converge: $$ \int^{1}_{0}\int^{\frac{1}{2y}}_{0}\left(\frac{3}{2}x + \frac{y}{2}\right) dxdy $$ $$ \int^{1}_{0}\int^{\frac{1}{2x}}_{0}\left(\frac{3}{2}x + \frac{y}{2}\right) dydx $$

Then I imagined space of $X, Y$ as a square on a plane with center in $(0.5, 0.5)$. Then the area when $X > 0.5$ and $Y > 0.5$ violate $ XY < \frac{1}{2}$, therefore I calculated $$ 1 - P(0.5 < x < 1, 0.5 < y < 1)$$ as $$ 1 - \int_{\frac{1}{2}}^{1}\int_{\frac{1}{2}}^{1}\left(\frac{3}{2}x + \frac{y}{2}\right) dxdy = \frac{5}{8}$$

But I'm not sure that it's right and therefore I'm here. I will be grateful for any help you can provide.


Actually, if $X>\frac{1}{2}$ and $Y>\frac{1}{2}$, you may still have $XY<\frac{1}{2}$. For instance, if $X=Y=\frac{5}{8}$, $XY=\frac{25}{64}<\frac{1}{2}$.

What you really need is $Y<\frac{1}{2X}$. So you can integrate the joint density of $X$ and $Y$ over the region $R$ where $0 \leq x \leq 1$, $0 \leq y \leq 1$, and $y < \frac{1}{2x}$. The reason your integrals did not converge is because you did not account for the fact that $X$ and $Y$ are always $ \leq 1$. You will have to break $R$ up into two regions to write down the integral, for instance $$P\left(Y< \frac{1}{2X}\right)=\int_R f(x,y) \ dy \ dx = \int_0^\frac{1}{2}\int_0^1 f(x,y) \ dy \ dx + \int_\frac{1}{2}^1 \int_0^\frac{1}{2x} f(x,y) \ dy \ dx.$$

Finally, you do NOT want $P\left(Y<\frac{1}{2X}\right)+P\left(X< \frac{1}{2Y}\right)$ because these are the same event; if $Y<\frac{1}{2X}$ then $X<\frac{1}{2Y}$ and vice versa. So just calculate one of them.


To be found is:$$\Pr\left(XY<\frac12\right)=\int^1_0\int^1_0f(x,y)g(x,y)dxdy$$ where $g$ denotes the indicator function of set $\{\langle x,y\rangle\in[0,1]^2\mid xy<\frac12\}$.

Observe that:


Try it out.


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.