MathOverflow is a question and answer site for professional mathematicians. Join them; it only takes a minute:

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

While looking for a closed form of a expression I worked myself to a formula that resembles the Vandermonde convolution, but is summed over even binomial coefficients only.


I'm at a loss as to what to do with it. I can re-write it in several ways, but the principal problem remains. Is there a known technique to attack such sums? Thanks.

share|cite|improve this question
up vote 15 down vote accepted

We have $$\sum_l\binom{a+l}lx^l=\frac1{(1-x)^{a+1}},$$ hence the generating function for the even terms of the sequence is $$\sum_l\binom{a+2l}{2l}x^{2l}=\frac12\left(\frac1{(1-x)^{a+1}}+\frac1{(1+x)^{a+1}}\right).$$ Consequently, \begin{multline*}\sum_l\binom{a+2l}{2l}\binom{b+2(n-l)}{2(n-l)}=\\\\ [x^{2n}]\frac14\left(\frac1{(1-x)^{a+1}}+\frac1{(1+x)^{a+1}}\right)\left(\frac1{(1-x)^{b+1}}+\frac1{(1+x)^{b+1}}\right),\end{multline*} where $[x^n]f$ denotes the $n$th coefficient of the power series for $f$. Plugging in the actual values for $a$ and $b$ and summing over $k$ gives \begin{align*} \sum_k\sum_l&\binom{2(k+l)}{2l}\binom{2(n-k)+2(n-l)}{2(n-l)}\\\\ &=[x^{2n}]\sum_{k=0}^n\frac14\left(\frac1{(1-x)^{2k+1}}+\frac1{(1+x)^{2k+1}}\right)\left(\frac1{(1-x)^{2(n-k)+1}}+\frac1{(1+x)^{2(n-k)+1}}\right)\\\\ &=[x^{2n}]\left[\frac{n+1}4\left(\frac1{(1-x)^{2n+2}}+\frac1{(1+x)^{2n+2}}\right)+\frac{1-x^2}{8x}\left(\frac1{(1-x)^{2n+2}}-\frac1{(1+x)^{2n+2}}\right)\right]\\\\ &=\frac{2n^2+4n+1}{2n+1}\binom{4n}{2n}=(2n^2+4n+1)C_{2n}. \end{align*}

share|cite|improve this answer
For $n=1$, we have $\sum_k \sum_l$ ${2k+2l}\choose{2l}$ ${4n-2k-2l}\choose{2n-2l}$ $= \sum_l$ ${2l}\choose{2l}$ ${4-2l}\choose{2-2l}$ + ${2+2l}\choose{2l}$ ${2-2l}\choose{2-2l}$ = ${4}\choose{2}$ + ${2}\choose{2}$ + ${2}\choose{0}$ + ${4}\choose{2}$ = $14$. – Zack Wolske Nov 5 '11 at 8:19
Thanks! In retrospective I should have figured that out myself. After fixing numerical mistakes after the second-to-last $=$ it evaluates to $\frac{n+1}{2}{4n+1 \choose 2n+1}-\frac{1}{4}{4n \choose 2n+1}+\frac{1}{4}{4n+2 \choose 2n+1}$ – Rasto S. Nov 7 '11 at 15:41
I am sorry for the errors, I hope it is correct now. – Emil Jeřábek Nov 8 '11 at 14:49
Yep, it's fixed. – Zack Wolske Nov 8 '11 at 16:13

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.

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