5
$\begingroup$

I am struggling with how to properly approach this problem. The wording is as follows:

Suppose that $X$ has a binomial$(n,p)$ distribution and let $Y$ have a negative binomial$(r, p)$ distribution (below) . Show that $F_X (r − 1) = 1 − F_Y (n − r)$.

I have been attempting a solution for several hours, but all of my work seems to basically go in circles without coming up with an actual proof.

My understanding is that $F_X (r - 1)$ is the CDF that demonstrates probabilities that $r-1$ successes have occurred before some amount of trials that have been performed, and $F_Y (n-r)$ is the probability that $n-r$ failures have occurred before a given amount of successes. Together, they add up to the sample space of possibilities, i.e. within $r-1+n-r$ trials (i.e., $n-1$ trials), a given number of successes and failures have occurred, with the $r^\text{th}$ success coming on the $n^\text{th}$ trial. I cannot see how to mathematically demonstrate these two distributions as complements, however.

$\endgroup$
3
$\begingroup$

Consider $n$ trials on which $r-1$ or fewer successes have occurred. This event can be described as $(X\leq r-1)$ and its probability is $P(X\leq r-1) = F_X(r-1)$. But this event can also be described as the event that at least $n-r+1$ failures have occurred before the $r$-th success has been observed, and so its probability can also be expressed as $$P(Y \geq n-r+1) = P(Y>n-r) = 1-F_Y(n-r)$$. Hence we have that $F_X(r-1) = 1-F_Y(n-r)$.

| cite | improve this answer | |
$\endgroup$

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.