If we want a central binomial coefficient, for n greater than zero we have a ‘nice’ expression $${2n \choose n} =\prod_{k=1}^{n}{(4-\frac{2}{k})}$$ which is term-wise rational and produces each binomial coefficient matching ${2n \choose n}$ along the way.

But supposed we desired a similar expression for some other ‘linear pattern’ in Pascal’s triangle, like ${3n \choose n}$, or ${5n \choose 2n}$. Perhaps it is term-wise algebraic rather than rational, sans any ‘bespoke encoding of a sequence into a magic constant’ horsery. Can our desire be satisfied under any conditions? If so, how, and if not, why?


Here is one other such expression I have been able to derive, in case anyone comes across this and is interested:

$${4n \choose 2n} = \prod_{k=1}^{n}{\frac{2(4k-1)(4k-3)}{k(2k-1)}}$$

And generally:

$${an \choose n} = \prod_{k=1}^{n}{\frac{a(k)_{a}}{(a-1)(k)_{(a-1)}}} $$

  • 1
    $\begingroup$ Something like ${3n\choose n}=\prod_{k=1}^n\frac{3n+1-k}{k} =\prod_{k=1}^n(\left(\frac{3n+1}{k}-1\right)$? $\endgroup$ – Hagen von Eitzen Apr 16 at 3:15
  • $\begingroup$ That is a nice one, thank you for sharing! I am interested to know how that can be derived, not only for ${3n \choose n}$ but for any such ${an \choose bn}$ with $b < a$, especially if it is possible without the term depending on $n$. $\endgroup$ – tuespetre Apr 16 at 4:11
  • $\begingroup$ I now see the general ‘multiplicative formula’ using a falling factorial, and how I could plug in the $a$ and $b$. I would still like to see if there is a way to avoid having the terms depend upon the upper bound. $\endgroup$ – tuespetre Apr 16 at 5:25

From $${2n \choose n} =\prod_{k=1}^{n}{(4-\frac{2}{k})}$$ one can obviously get $${4n \choose 2n} =\prod_{k=1}^{2n}{(4-\frac{2}{k})}$$ and

$${6n \choose 3n} =\prod_{k=1}^{3n}{(4-\frac{2}{k})}$$ Which are not new formulas.

In general we have $$ {m \choose n} =\prod_{k=0}^{n-1}{(1+\frac{d}{n-k})} $$ where $d=m-n.$


It appears that for any $${an \choose bn}$$ with $0 \le b \le a$ we have the degree $(a-1)$ rational curve $$R_{a,b}(x) = \frac{(ax)^{\underline{a}}}{(bx)^{\underline{b}}((a-b)x)^{\underline{a-b}}}$$ such that $${a(n-1) \choose b(n-1)}R_{a,b}(n) = {an \choose bn} = \prod_{k=1}^{n} R_{a,b}(k).$$ I do not have a rigorous proof for this; I just discovered the pattern through computation.


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.