0

I tried to prove $$\sum\limits_{r=0}^n\binom{n}{r}=\sum\limits_{r=0}^n \frac{n!}{r!(n-r)!} =2^n.$$

I used induction method:

Assuming $$\sum\limits_{r=0}^n \frac{n!}{r!(n-r)!} =2^{n}.$$

Proving $$\sum\limits_{r=0}^{n+1} \frac{{(n+1)}!}{r!(n+1-r)!} =2^{n+1}.$$

I got $$\sum\limits_{r=0}^{n+1} \frac{{(n+1)}!}{r!(n+1-r)!} =\sum\limits_{r=0}^{n+1} \Bigl(\frac{n!}{r!(n-r)!}+\frac{n!}{(r-1)!(n-r+1)!}\Bigr),$$ and an invalid term appeared: $(-1)!$.

How does one prove this identity, and what's wrong with my method?

John He
  • 559

1 Answers1

0

The identity you are using in the last equality has some hypothesis: it holds when $r > 0$.

There are many approaches to prove this fact. I particulary like a combinatorial proof via the following counting argument: if $A_k$ is the set of subsets of $\{1, \dots, n\}$ of size $k$, then by (one possible) definition, $|A_k| = {n \choose k}$. Moreover, we have that the subsets of $\{1, \dots, n\}$, that is the elements of $\mathcal{P}(\{1,\dots,n\})$, are the disjoint union of $A_0, \dots, A_n$.

Thus, $\mathcal{P}(\{1,\dots,n\}) = \sqcup_{0 \leq i \leq n}A_i$ and

$$ 2^n = |\mathcal{P}(\{1,\dots,n\})| = |\sqcup_{0 \leq i \leq n}A_i| = \sum_{i=0}^n|A_i| = \sum_{i = 0}^n{n \choose i}. $$

qualcuno
  • 17,557