While deriving a elementary argument for the inverse of the Hilbert Matrix, I encountered the following quantity

If $n \in \mathbb Z_+$ and $i \in \{0,\dots,n-1\}$, then $$ \frac{2n+1}{n-i} \binom{n+i}{n-1-i} = \frac{2n+1}{2i+1} \binom{n+i}{n-i}$$ is an integer.

Numerical evidence seems to suggest that it is true (verified for $n$ up to $1024$). But there are several subtleties:

  1. If the $2n+1$ coefficient is removed, this statement is no longer true.
  2. It might be tempting to separate the binomial coefficient into two using the Pascal Triangle: $$ \binom{n+i}{n-1-i} = \binom{n-1+i}{n-2-i} + \binom{n-1+i}{n-1-i}$$ but each component after separating is not an integer.
  3. By the substitution $n + i = a,n - i = b$, $$ \frac{2n+1}{2i+1} \binom{n+i}{n-i} = \frac{a+b+1}{a-b-1} \binom{a}{b} $$
  • 3
    $\begingroup$ Your integer is $\dbinom{n+i}{2i} + 2 \dbinom{n+i}{2i+1}$. $\endgroup$ – darij grinberg Jan 27 '18 at 0:50

Recalling the binomial identity $\binom{n}{k}=\frac{n}{k}\binom{n-1}{k-1}$ we obtain \begin{align*} \color{blue}{\frac{2n+1}{n-i}\binom{n+i}{n-i-1}}&=\frac{(n-i)+(n+i+1)}{n-i}\binom{n+i}{n-i-1}\\ &=\left(1+\frac{n+i+1}{n-i}\right)\binom{n+i}{n-i-1}\\ &\color{blue}{=\binom{n+i}{n-i-1}+\binom{n+i+1}{n-i}\in\mathbb{Z}} \end{align*}


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.