2
$\begingroup$

$U$ stands for the number of trials to get the first head, $V$ stands for the number of trials to get two heads.

I used hand-waving proof, saying that you could not have the two heads trials without having the first head occur, but I don't know how to mathematically prove that the two random variables must be dependent.

$\endgroup$
  • $\begingroup$ Find functions $f$ and $g$ so that $Ef(U)G(V)\ne Ef(U) Eg(V)$. $\endgroup$ – kimchi lover Oct 15 '18 at 12:08
  • 2
    $\begingroup$ What's the probability that $U=3$ and $V=2$? What's the probability of those events separately? $\endgroup$ – lulu Oct 15 '18 at 12:10
3
$\begingroup$

The variables are independent if $P(U=u, V=v)=P(U=u)\cdot P(V=v)$ for all values $u,v$.

Therefore, to prove that the variables are not independent, all you need to do is find one pair $u,v$ such that

$P(U=u, V=v)\neq P(U=u)\cdot P(V=v)$

To do that, you can go one of two ways:

Option $1$: Try a little brute force search. Plug in a couple values $u,v$ and see if the equation holds.

Option $2$: Since the right side of the equation will always be nonzero, think about how you could make the left side zero. Can you think of some pair of numbers $u,v$ such that it's impossible to get two heads in $v$ trials, but only one head in $u$ trials?

$\endgroup$
  • $\begingroup$ I attempted to do option 1 but don't know how to derive the joint distribution of u,v...which is why I posted this question to see if someone could give me some suggestion on the most generic way of solving this question... $\endgroup$ – Chloe Zhou Oct 15 '18 at 14:16
  • $\begingroup$ @ChloeZhou Maybe try option 2? $\endgroup$ – 5xum Oct 16 '18 at 7:40

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.