1

I have to show that $$\displaystyle\sum\limits_{k=0}^n (-1)^k\binom{m+1}{k}\binom{m+n-k}{n-k} = \begin{cases} 1\ \text{if}\ n=0 \\ 0\ \text{if}\ n>0 \end{cases}$$

My try: I have tried to use snake oil method $$\sum_{k=0}^n \binom{m+1}{k}\binom{m+n-k}{n-k}$$ $$=\sum_m{\sum_{k=0}^n \binom{m+1}{k}\binom{m+n-k}{n-k}}x^m$$ $$={\sum_{k=0}^n(-1)^k \binom{m+1}{k}\sum_m\binom{m+n-k}{n-k}}x^m$$ We know that the second sum $\frac{1}{(1-x)^{1+n-k}}=\sum_m\binom{m+n-k}{n-k}x^m$ $$=\sum_{k=0}^n(-1)^k\binom{m+1}{k}\frac{1}{(1-x)^{1+n-k}}$$$$=\frac{1}{(1-x)^{1+n}}\sum_{k=0}^n(-1)^k\binom{m+1}{k}(1+x)^k$$ How should I proceed after that?

Edit: My method might be absolutely wrong so if anyone might show me some other method I would really appreciate it

Chill2Macht
  • 20,920
  • What is $\sum_m$? 2) If $m<n$ I think the last sum doesn't exist in closed form, 3) How did you get $x$ in the sum over $m$?
  • – Alex May 21 '16 at 18:06
  • @Alex By $\sum_{m}$ I mean sum over all $m$.And I have also edited the question – Sayan Chattopadhyay May 21 '16 at 18:11
  • 1
    I'm curious about a combinatorial proof of this statement, and so I've started a question to that end here: http://math.stackexchange.com/questions/1794539/combinatorial-proof-of-a-certain-alternating-sum-of-binomial-coefficients?lq=1 – Semiclassical May 21 '16 at 19:15
  • @Albas: Hint: It is not valid to factor out $\binom{m+1}{k}$, since this binomial coefficient depends on the inner sum with index $m$. – Markus Scheuer May 21 '16 at 21:05
  • @MarkusScheuer Does that mean it will be better to take $n$ as the free variable here? – Sayan Chattopadhyay May 22 '16 at 04:46
  • @Albas: When checking this problem again, I saw that the snake oil method does also work. Regards, – Markus Scheuer May 23 '16 at 16:23