# Truncated alternating binomial sum

It is easily checked that $\displaystyle\sum_{i\ =\ 0}^{n}\left(\, -1\,\right)^{i} \binom{n}{i} = 0$, for example by appealing to the binomial theorem.

I'm trying to figure out what happens with the truncated sum $\displaystyle\sum_{i\ =\ 0}^{D}\left(\, -1\,\right)^{i}\binom{n}{i}$.
How far away from $0$ can this get, as a function of $D$ ?.

I'm mostly interested in the case of when $D \ll n$, such as $D \sim \,\sqrt{\,n\,}\,$.

Thanks !

-
I'm still working on the proof myself, but look at $D=n-2$ and then use induction from there, the result is surprisingly simple. – Silynn Aug 5 '14 at 10:13

Let $n\ge 2\in\mathbb N$ (since the case $n=1$ is trivial).

For $0\le D\lt n$, we can prove the following by induction: $$\sum_{i=0}^{D}(-1)^i\binom{n}{i}=(-1)^D\binom{n-1}{D}.$$

For $D=0$, it holds trivially.

For a $D$ such that $0\le D\le n-2$, suppose that it holds. Then, \begin{align}\sum_{i=0}^{D+1}(-1)^i\binom{n}{i}&=(-1)^{D+1}\binom{n}{D+1}+\sum_{i=0}^{D}(-1)^i\binom{n}{i}\\&=(-1)^{D+1}\binom{n}{D+1}+(-1)^D\binom{n-1}{D}\\&=(-1)^{D+1}\left\{\binom{n}{D+1}-\binom{n-1}{D}\right\}\\&=(-1)^{D+1}\binom{n-1}{D+1}\end{align} Hence, it holds when $D+1$.

Therefore, it holds for any $0\le D\lt n$. Q.E.D.

From this, you'll also see how far away from $0$ it can get.

-

First, let us write $$\sum_{i=0}^{D} (-1)^i \binom{n}{i}=\sum_{i=0}^D \binom{i-n-1}{i}$$ This step can be proven by using the definition of binomial coefficient and pulling a $-1$ out of each term.

Next, $$\sum_{i=0}^D \binom{i-n-1}{i}=\binom{D-n}{D}$$ can be proven inductively.

And finally, this can be simplified using the same result in the first step. $$\binom{D-n}{D}=(-1)^D\binom{n-1}{D}$$

-

Use the following:

$$(1-x)^{n-1} = (1-x)^n \times \frac{1}{1-x} = (1-x)^n (1 + x + x^2 + \dots) =$$ $$\left(1 + n(-x) + \binom{n}{2}(-x)^2 + \dots + (-x)^n\right)(1+x+x^2 + \dots)$$

Now, mutiplying any polynomial (or power series) by $1 + x + x^2 + \dots$ has the effect of giving you the truncated sums of the coefficients of the polynomial as the coefficients of the powers of $x$ in the resulting power series.

In your case, the resulting series is itself a polynomial, $(1-x)^{n-1}$, giving you a neat closed form answer.

-

$\newcommand{\angles}[1]{\left\langle\, #1 \,\right\rangle} \newcommand{\braces}[1]{\left\lbrace\, #1 \,\right\rbrace} \newcommand{\bracks}[1]{\left\lbrack\, #1 \,\right\rbrack} \newcommand{\ceil}[1]{\,\left\lceil\, #1 \,\right\rceil\,} \newcommand{\dd}{{\rm d}} \newcommand{\ds}[1]{\displaystyle{#1}} \newcommand{\expo}[1]{\,{\rm e}^{#1}\,} \newcommand{\fermi}{\,{\rm f}} \newcommand{\floor}[1]{\,\left\lfloor #1 \right\rfloor\,} \newcommand{\half}{{1 \over 2}} \newcommand{\ic}{{\rm i}} \newcommand{\iff}{\Longleftrightarrow} \newcommand{\imp}{\Longrightarrow} \newcommand{\pars}[1]{\left(\, #1 \,\right)} \newcommand{\partiald}[3][]{\frac{\partial^{#1} #2}{\partial #3^{#1}}} \newcommand{\pp}{{\cal P}} \newcommand{\root}[2][]{\,\sqrt[#1]{\vphantom{\large A}\,#2\,}\,} \newcommand{\sech}{\,{\rm sech}} \newcommand{\sgn}{\,{\rm sgn}} \newcommand{\totald}[3][]{\frac{{\rm d}^{#1} #2}{{\rm d} #3^{#1}}} \newcommand{\verts}[1]{\left\vert\, #1 \,\right\vert}$\begin{align} &\color{#66f}{\large\sum_{k = 0}^{D}\pars{-1}^{k}{n \choose k}} =\sum_{k = 0}^{D}\pars{-1}^{k}\ \overbrace{\oint_{0\ <\ \verts{z}\ =\ a\ <\ 1} {\pars{1 + z}^{n} \over z^{k + 1}}\,{\dd z \over 2\pi\ic}} ^{\ds{=\ {n \choose k}}} \\[3mm]&=\oint_{0\ <\ \verts{z}\ =\ a\ <\ 1}{\pars{1 + z}^{n} \over z} \sum_{k = 0}^{D}\pars{-\,{1 \over z}}^{k}\,{\dd z \over 2\pi\ic} =\oint_{0\ <\ \verts{z}\ =\ a\ <\ 1}{\pars{1 + z}^{n} \over z} {\pars{-1/z}^{D} + z \over 1 + z}\,{\dd z \over 2\pi\ic} \\[3mm]&=\pars{-1}^{D}\ \underbrace{\oint_{\verts{z}\ =\ a\ <\ 1} {\pars{1 + z}^{n - 1} \over z^{D + 1}}\,{\dd z \over 2\pi\ic}} _{\ds{=\ {n - 1 \choose D}}}\ +\ \underbrace{\oint_{0\ <\ \verts{z}\ =\ a\ <\ 1}\pars{1 + z}^{n - 1} \,{\dd z \over 2\pi\ic}}_{\ds{=\ 0}} \\[3mm]&=\color{#66f}{\large\pars{-1}^{D}{n - 1 \choose D}} \end{align}

-
I'll admit that this would not have been the way I would've tried to prove this :-) – Henry Yuen Aug 6 '14 at 11:53
@HenryYuen However, we'll go straightforward to the end. It's quite useful in more complicated cases. You can check my answers and there you'll find a lot of applications of this method. Thanks. – Felix Marin Jan 7 at 12:00