I have an urn with $N$ red balls and $m$ blue balls. I sample with replacement from this urn until I encounter a blue ball, which I then discard. What is my expectation and variance for the number of total sampling events (with and without replacement) before all $m$ blue balls are discarded?


This answers half of the question.

Let $X_0$ be $1$ more than the number of red balls drawn before the first blue ball is drawn. For $k=1,\dots,m-1$ let $X_k$ be $1$ more than the number of red balls drawn after the $k$-th blue ball but before the $(k+1)$-st blue ball. Then $X=\sum_{k=0}^{m-1}X_k$ is the total number of draws needed to get rid of the blue balls.

When there are $k$ blue balls, the probability of drawing a blue ball is $\frac{k}{N+k}$, so the expected number of draws up to and including the first draw of a blue ball is $\frac{N+k}k=\frac{N}k+1$. Thus,

$$\Bbb E(X_k)=\frac{N}{m-k}+1$$

for $k=0,\dots,m-1$, and by linearity of expectation

$$\Bbb E(X)=\sum_{k=0}^{m-1}\Bbb E(X_k)=\sum_{k=0}^{m-1}\left(\frac{N}{m-k}+1\right)=N\sum_{k=1}^m\frac1k+m=NH_m+m\;,$$

where $H_m$ is the $m$-th harmonic number. Thus,

$$\Bbb E(X)\approx N\left(\ln m+\gamma+\frac1{2m}-\frac1{12m^2}\right)+m\;,$$

where $\gamma\approx0.5772156649$ is the Euler-Mascheroni constant.

| cite | improve this answer | |

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.