91Ó°ÊÓ

The joint probability mass function is a function that fully describes a discrete random vector's distribution.

The number of balls that must be observed in order to achieve a special ball is $Y_{k+1}$.

$Y_{k+1}=n+1-\underset{i=1}{\overset{k}{∑}}{Y}_i$

$Y_{k+1}$stands for the amount of balls that must be observed in order to achieve a special ball.

Let $Y_2$be the number of further balls that are removed until the second special ball arrives, and let $Y_1$signify the selection number of the first special ball that was withdrawn.

After that, the joint mass function,

$$\begin{gathered} P(Y_j=i_{j,}j=1,2,....,k+1) \\ =P(Y_j=i_{j,}j=1,2,....,k)P(Y_{k+1}=i_{k+1}:Y_j=i_j,j=1,2,....k) \\ =\frac{k!\left(n-k\right)!}{n!}P\left\{n+1-\underset{i=1}{\overset{k}{∑}}{Y}_i=i_{k+1}:Y_j=i,j=1,2,....k\right\} \\ =\frac{k!\left(n-k\right)!}{n!},\left\{\underset{j=1}{\overset{k+1}{∑}}{i}_j=n+1\right\} \\ =P(y_1,y_2,....y_n)y_i=0,1,\underset{i=1}{\overset{n}{∑}}{y}_i=k \end{gathered}$$

As a result, the joint mass function is symmetric, proving the conclusion. That is to say, the random variables are exchangeable.