0

I know that this problem might qualify as a duplicate, but I didn't find any satisfactory answers among the similar posts.

We want to show by induction that \begin{equation} \sum_{i=0}^{n} \binom{i}{m} = \binom{n+1}{m+1} \end{equation} for all $n \geq 0$ and $m \geq 0$. In the basis step, we put $n=1$ and get \begin{align*} &\sum_{i=0}^1 \binom{i}{m} = \binom{0}{m} + \binom{1}{m} \end{align*} We see that $\binom{0}{m} = 0$ if $m = 0$ and $\binom{0}{m} = 1$ if $m > 0$. Lets first look at the case $m=0$: \begin{align*} &\sum_{i=0}^{1} \binom{i}{0} = 1 + \frac{1!}{0!1!} = 2 = \binom{2}{1} \end{align*} Thus, the equation holds for $n=1$ when $m = 0$. When $m>0$ we get \begin{align*} &\sum_{i=0}^{1} \binom{i}{m} = \binom{1}{m} = \frac{1!}{m!(1-m)!} \end{align*}

I don't know where to go from here. I want to end up with $\frac{2!}{(1+m)!(1 - m)!}$, so I tried to multiply with $\frac{(m+1)!}{(m+1)!}$, but that just got me here: \begin{align*} \sum_{i=0}^{1} \binom{i}{m}&= \frac{1}{m!(1-m)!} \cdot \frac{(m+1)!}{(m+1)!} \\ &= \frac{(m+1)}{(1-m)!(m+1)!} \end{align*}

I didn't get anywhere using Pascals rule, either. And this is just the basis step of the induction. The induction step, with $n = k+1$, seems just as difficult to me.

SBS
  • 459
  • $\displaystyle \binom00 = \frac{0!}{0!0!} = 1$, not $0$. Not sure if that will help resolve the overall issue, though. –  Sep 14 '16 at 13:53
  • 2
    Have a look at: http://math.stackexchange.com/questions/1490794/proof-of-the-hockey-stick-identity-sum-t-0n-binom-tk-binomn1k1 – StubbornAtom Sep 14 '16 at 13:55
  • Notice that $\binom{n}{k}=0$ if $n<k$ for $n,k\in\Bbb N_0$. So $\binom{0}{m}=0$ if $m>0$. – Masacroso Sep 14 '16 at 14:04
  • The exact same question was asked before but was closed (improperly, in my opinion) as a duplicate: http://math.stackexchange.com/questions/1682754/proof-by-induction-that-sum-k-nm-binomkn-binomm-1n-1 – David K Sep 14 '16 at 14:06
  • 1
    Another issue with your attempt at the proof is that you are trying to prove this for all $n\geq 0$ so... your basis should start with $n=0$ not $n=1$ (unless the argument is different for $n=0$ than for $n=1$ somehow and the $n=1$ case doesn't follow easily) – JMoravitz Sep 14 '16 at 14:08
  • http://math.stackexchange.com/questions/1490794/proof-of-the-hockey-stick-identity-sum-t-0n-binom-tk-binomn1k1 has one inductive answer, but generally doesn't ask for inductive proof. – David K Sep 14 '16 at 14:08
  • I realize now that there's a mistake in the text. We have $\binom{0}{m} = 1$ when $m=0$ and $\binom{0}{m} = 0$ when $m>0$, not the other way around. – SBS Sep 14 '16 at 17:55
  • @SBS Are you sure about $\sum \limits_{i= \color {red}0}^{n}$? because that would force $m=0$ – Airdish Sep 20 '16 at 16:37
  • @SBS Inductive proof below. – ClownInTheMoon Sep 20 '16 at 17:08

3 Answers3

2

An inductive proof rests entirely on Pascal's formula. Note the above formula can as well be written: $$\sum_{i=m}^n \binom{i}{m} = \binom{n+1}{m+1}\quad\text{or}\quad \binom{n+1}{m+1}=\sum_{k=0}^{n-m} \binom{n-k}{m}$$ We'll prove by induction on $l$ that, for all $l\;$ ($0\le l<n-m$): $$ \binom{n+1}{m+1}=\sum_{k=0}^{l}\binom{n-k}{m}+\binom{n-l}{m+1} \tag{1}$$

Indeed, if $l=0$, this is simply Pascal's formula.

For the inductive step, suppose formula $(1)$ is true for some $l$. We have \begin{align} \binom{n+1}{m+1}&=\binom{n}{m}+\binom{n-1}{m}+\dots+\binom{n-l}{m}+\binom{n-l}{m+1}\\ &=\binom{n}{m}+\binom{n-1}{m}+\dots+\binom{n-l}{m}+\binom{n-l-1}{m}+\binom{n-l-1}{m+1}\\ &=\sum_{k=0}^{l+1}\binom{n-k}{m}+\binom{n-l-1}{m+1}. \end{align}

Setting $l=n-m-1$, we obtain \begin{align} \binom{n+1}{m+1}&=\binom{n}{m}+\binom{n-1}{m}+\dots+\binom{m+1}{m}+\binom{m+1}{m+1}\\ &=\binom{n}{m}+\binom{n-1}{m}+\dots+\binom{m+1}{m}+\binom{m}{m}. \end{align}

Bernard
  • 175,478
0

For the inductive step, assume:

$$\sum_{i=0}^n \dbinom{i}{m} = \dbinom{n+1}{m+1}$$

Add $\dbinom{n+1}{m}$ to both sides to give:

$$\sum_{i=0}^{n+1} \dbinom{i}{m} = \dbinom{n+1}{m}+\dbinom{n+1}{m+1}=\dbinom{n+2}{m+1}$$

JMP
  • 21,771
  • I see that you have suggested a synonym between (binomial-theorem) and (binomial-coefficients) tag. I wanted to let you know that there is a post on meta about this synonym. In general, I think that it is better to discuss synonyms first on meta. (As I tried to tell you before, but it is possible that my previous messages did not reach you.) – Martin Sleziak Oct 22 '16 at 15:03
0

Here is how you prove this via Induction. Let me know if you have questions on a particular step. Under the inductive step it's basically:

1) Perform induction on $n$. You could substitute $k$ for $n$ if you'd like but I did not.

2) Apply Pascal's Identity to split the sum into two sums.

3) Factor out $n+1$ from your sums since $i=n+1$ yields ${(n+1)-1 \choose m}$ on the first sum and ${(n+1)-1 \choose m-1}$ on the second sum, we can write the factors we pull out as ${n \choose m}$ and ${n \choose m-1}$.

4) Regroup

5) Apply Pascal's Identity to your two factored out terms and your two sums.

6) Apply your inductive hypothesis to $\sum_{i=0}^{n}{i\choose m}$.

7) Finally, apply Pascal's Identity one last time.

Here is the work:

Proposition. $\sum_{i=0}^n{i\choose m}={n+1\choose m+1}$ where $0\le m\le n$.

$Proof.$ We use mathematical induction.

Base Case. Let $n=0$. Then $\sum_{i=0}^{n=0}{i\choose m}= {0\choose 0}=1$ and ${n+1\choose m+1}={0+1\choose 0+1}={1 \choose1} = 1$. Thus for $n=0$ our proposition holds.

Inductive Step. Let $n\ge0$. Assume $\sum_{i=0}^n{i\choose m}={n+1\choose m+1}$. Then

$$\begin{align*}\sum_{i=0}^{n+1}{i\choose m}&=\sum_{i=0}^{n+1}{i-1\choose m}+\sum_{i=0}^{n+1}{i-1\choose m-1}\\ &={n\choose m}+\sum_{i=0}^{n}{i-1\choose m}+{n\choose m-1}+\sum_{i=0}^{n}{i-1\choose m-1}\\ &={n\choose m}+{n\choose m-1}+\sum_{i=0}^{n}{i-1\choose m}+\sum_{i=0}^{n}{i-1\choose m-1}\\ &={n+1\choose m}+\sum_{i=0}^{n}{i\choose m}\\ &={n+1\choose m}+{n+1\choose m+1}\\ &={n+2\choose m+1}\\ &={(n+1)+1\choose m+1} \end{align*} $$

It follows by mathematical induction that $\sum_{i=0}^n{i\choose m}={n+1\choose m+1}$ where $0\le m \le n$. $\Box$