In my textbook, a clear proof that the Geometric Distribution is a distribution function is given, namely
$$\sum_{n=1}^{\infty} \Pr(X=n)=p\sum_{n=1}^{\infty} (1-p)^{n-1} = \frac{p}{1-(1-p))}=1.$$
Then the textbook introduces the Negative Binomial Distribution; it gives a fairly clear explanation for why the PMF of a Negative Binomial random variable $N$ with parameter $r$ is
$$p\binom{n-1}{r-1}p^{r-1}(1-p)^{n-r} = \binom{n-1}{r-1}p^{r}(1-p)^{n-r} $$
But to show
$$\sum_{n=r}^{\infty} \Pr(N=n)=\sum_{n=r}^{\infty}\binom{n-1}{r-1}p^{r}(1-p)^{n-r}=1$$
the textbook gives (in my opinion) a wordy and informal argument that is nowhere near as clear.
What is a straightforward algebraic way to prove the above statement; that the Negative Binomial is a distribution function? I also looked at a different probability textbook, plus wolfram.com's definition before asking.
Best Answer
It's evident that $\Bbb{P}(N=n)\ge 0$ for $n\ge r$. So you have to prove that $\sum_{n\ge r}\Bbb{P}(N=n)=1$: $$ \begin{align} \sum_{n\ge r}\Bbb{P}(N=n)&=\sum_{n\ge r} \binom {n-1} {r-1} p^r \left({1-p}\right)^{n-r}\\ &=\sum_{n\ge r} \binom {n-1} {n-r} p^r \left({1-p}\right)^{n-r}\;\;\quad\quad\text{(symmetry})\\ &=p^r\sum_{j\ge 0} \binom {r+j-1} {j} \left({1-p}\right)^{j}\qquad\text{(substituting }j=n-r)\\ &=p^r\sum_{j\ge 0} (-1)^j \binom{-r}{j}\left({1-p}\right)^{j}\qquad\text{(identity}\tbinom{j+r-1}{j}=(-1)^j \tbinom{-r}{j})\\ &=p^r\sum_{j\ge 0} \binom{-r}{j}\left({p-1}\right)^{j}\\ &=p^r\left(1+(p-1)\right)^{-r} \qquad\qquad\qquad\text{(binomial theorem) }\\ &=1 \end{align} $$ using the identity $$ \begin{align} \binom{j+r-1}{j}&=\frac{(j+r-1)(j+r-2) \cdots r}{j!}\\ &=(-1)^j \frac{(-r-(j-1))(-r-(j-2)) \cdots (-r)}{j!} \\&=(-1)^j \frac{(-r)(-r-1) \cdots (-r-(j-1))}{j!} \\&=(-1)^j \binom{-r}{j} \end{align} $$