How do i show this identity: $ \sum\limits_{n=0}^{a_1} \dfrac{a_2 +n \choose a_2}{a_1+a_2 \choose a_2} = \sum\limits_{n=0}^{a_1} \dfrac{a_1 \choose n}{a_1+a_2 \choose n} $

I don't think the partial sums match $ \dfrac{a_2 +n \choose a_2}{a_1+a_2 \choose a_2} ≠ \dfrac{a_1 \choose n}{a_1+a_2 \choose n}$

I was trying to understand why $\sum ^{a_{1}+a_{2}}_{n=1}\dfrac {\begin{pmatrix} a_{1} \\ n \end{pmatrix}}{\begin{pmatrix} a_{1}+a_{2} \\ n \end{pmatrix}}=\dfrac {a_{1}}{1+a_{2}}$,

Somebody kindly outlined the proof, using $\sum\limits_{m=0}^b {c+m \choose c} ={b+c+1 \choose c+1} = \frac{b+c+1}{c+1} {b+c \choose c}$ as $\sum\limits_{n=1}^{a_1+a_2} \dfrac{a_1 \choose n}{a_1+a_2 \choose n} = \sum\limits_{n=0}^{a_1} \dfrac{a_1 \choose n}{a_1+a_2 \choose n} - 1$ = $= \sum\limits_{n=0}^{a_1} \dfrac{a_2 +n \choose a_2}{a_1+a_2 \choose a_2} - 1$ = $= \dfrac{a_2+a_1+1}{a_2+1}-1$

I understand apart from that identity.


We start with the right-hand side and obtain \begin{align*} \color{blue}{\sum_{n=0}^{a_1}}&\color{blue}{\binom{a_1}{n}\binom{a_1+a_2}{n}^{-1}}\\ &=\sum_{n=0}^{a_1}\frac{a_1!}{n!(a_1-n)!}\cdot\frac{n!(a_1+a_2-n)!}{(a_1+a_2)!}\\ &=\frac{a_1!}{(a_1+a_2)!}\sum_{n=0}^{a_1}\frac{(a_1+a_2-n)!}{(a_1-n)!}\\ &=\frac{a_1!a_2!}{(a_1+a_2)!}\sum_{n=0}^{a_1}\frac{(a_2+n)!}{n!a_2!}\tag{1}\\ &\,\,\color{blue}{=\binom{a_1+a_2}{a_2}^{-1}\sum_{n=0}^{a_1}\binom{a_2+n}{a_2}} \end{align*} and the claim follows.


  • In (1) we change the order of summation $n\to a_1-n$ and expand with $a_2!$.

The other identity
\begin{align*} \sum_{n=1}^{a_1}\binom{a_1}{n}\binom{a_1+a_2}{n}^{-1}=\frac{a_1}{1+a_2} \end{align*} can be derived for instance using the Beta function with the identity \begin{align*} \binom{p}{q}^{-1}=(p+1)\int_0^1x^q(1-x)^{p-q}\,dx \end{align*}

We obtain \begin{align*} \color{blue}{\sum_{n=1}^{a_1}}&\color{blue}{\binom{a_1}{n}\binom{a_1+a_2}{n}^{-1}}\\ &=(a_1+a_2+1)\int_0^1\sum_{n=1}^{a_1}\binom{a_1}{n}x^n(1-x)^{a_1+a_2-n}\,dx\\ &=(a_1+a_2+1)\int_0^1(1-x)^{a_1+a_2}\sum_{n=1}^{a_1}\binom{a_1}{n}\left(\frac{x}{1-x}\right)^n\,dx\\ &=(a_1+a_2+1)\int_0^1(1-x)^{a_1+a_2}\left[\left(1+\frac{x}{1-x}\right)^{a_1}-1\right]\,dx\\ &=(a_1+a_2+1)\int_0^1\left[(1-x)^{a_2}-(1-x)^{a_1+a_2}\right]\,dx\\ &=(a_1+a_2+1)\left[\frac{1}{a_2+1}-\frac{1}{a_1+a_2+1}\right]\\ &\,\,\color{blue}{=\frac{a_1}{a_2+1}} \end{align*} and the claim follows.

  • 1
    $\begingroup$ @ Yes, step (1) was what I could not get, changing the order of summation back to front. Thank you very much, it seems a very long winded way to prove $\sum ^{a_{1}+a_{2}}_{n=1}\dfrac {\begin{pmatrix} a_{1} \\ n \end{pmatrix}}{\begin{pmatrix} a_{1}+a_{2} \\ n \end{pmatrix}}=\dfrac {a_{1}}{1+a_{2}}$ $\endgroup$ – Tinatim Jun 27 '18 at 13:50
  • $\begingroup$ Of course, my apologies. Ok, what's the straight-forward approach then? $\endgroup$ – Tinatim Jun 27 '18 at 14:28
  • $\begingroup$ @Tinatim: Everything's fine. Proof of the other identity added. Regards, $\endgroup$ – Markus Scheuer Jun 27 '18 at 14:51
  • $\begingroup$ Nice, I assume this should read $\begin{align*} \binom{p}{q}^{-1}=(p+1)\int_0^1x^q(1-x)^{p-q}\,dx \end{align*}$ I will look at the beta function. $\endgroup$ – Tinatim Jun 27 '18 at 16:56
  • 1
    $\begingroup$ @Tinatim: This approach is often useful when working with reciprocals of binomial coefficients. Thanks for pointing at the mistake. Corrected. $\endgroup$ – Markus Scheuer Jun 27 '18 at 17:03

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.