A point $(x, y)$ inside of a unit circle is picked uniformly at random. Then a radius $r$ is picked at random such that a circle $(x, y, r)$ is inside the unit circle. What is the probability that unit circle's origin is inside of circle $(x, y, r)$?

I'll assume that the unit circle is centered at the point $(0, 0)$.

I tried to fix the $y$ coordinate of the random point to $0$ and then try to think about all possible $x, r$ values. This has lead me to the following inequalities: $$0 \leq x \leq 1$$ $$0 \leq r \leq 1 - x$$

This corresponds to all radii such that a new circle is inside of unit one. Now, for a new circle to contain $(0, 0)$ $$x \leq r$$

must hold.

This has lead me to calculating the probability of $\frac{1}{2}$ but it seems to be incorrect. Could you help me please?

  • $\begingroup$ So the success probability for a fixed $x$ is $(1-2x)/(1-x)$. Now integrate over $x$. $\endgroup$ – Marcus Ritt May 16 '19 at 11:43
  • $\begingroup$ I think you must be thinking $x\leq r\leq 1-x$ implies $x\leq\frac12.$ True enough, but this is a necessary condition, not a sufficient one. If $x=.4$ say, $r$ has to be between $.4$ and $.6$. $\endgroup$ – saulspatz May 16 '19 at 11:51
  • $\begingroup$ Don't forget to work out the distribution of $x$ within the interval $[0,1].$ It isn't uniform. $\endgroup$ – David K May 16 '19 at 12:00

First we need to find the probability of $x,y$ to be on distance $R$ from the center: CDF is $P(d<R) = R^2$, so PDF is $\omega(R) = 2R$.

Assuming that “radius picked at random” means, uniform distribution from 0 to 1-R, then the probability of $r-R>0$: $$ \int_0^{1/2} \int_R^{1-R}\omega(R)\omega(r)drdR=\int_0^{1/2} \int_R^{1-R}\frac{2R}{1-R}drdR=\int_0^{1/2}\frac{2R(1-2R)}{1-R}dR\\ = \frac32 - \log 4 $$


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.