Are these well known properties of binomial coefficients?

257 Views Asked by At

I apologize for the number of definitions. I did not know how to state these ideas any simpler. If anyone can help me simplify the definitions, I will be glad to shorten the details.

Let:

  • $x,n$ be integers with $n > 1$ and $x \ge 2n$

  • $f(x,n) = \dfrac{(x+n)!}{x!}$

  • lcm$(a,b)$ be the least common multiple of $a$ and $b$.

  • $v_p(x) = $ the maximum power of $p$ that divides $x$ so that $p^{v_p(x)} | x$ but $p^{v_p(x)+1} \nmid x$

  • $g(x,n) = \prod\limits_{p\text{ prime & }p \le n} \dfrac{p^{v_p(\text{lcm}(x+1, \dots, x+n))}}{p^{v_p(\text{lcm}(2,3,\dots,n))}}$

  • $h(x,n) = \prod\limits_{p\text{ prime &}p > n\text{ & }p^t | f(x,n)}p$

  • $t = \dfrac{n!}{\text{lcm}(2,3,\dots,n)}$

  • $u = \dfrac{f(x,n)}{\text{lcm}(x+1, x+2, \dots, x+n)}$

  • $w = \dfrac{u}{t}$

Does it now follow that the following relations hold for binomial coefficients:

  • lcm$(x+1, x+2, \dots, x+n) = g(x,n)h(x,n)\text{lcm}(2,3,\dots,n)$
  • ${{x+n} \choose n} = g(x,n)h(x,n)w$
  • From the Sylvester-Schur theorem, $h(x,n) > n$
  • $w \le \dfrac{(n-1)!}{t}$ [I am still working on the argument.]
  • lcm$(x+1, x+2, \dots, x+n) \le {{x+n}\choose{n}}$lcm$(2,3,\dots,n)$

For me, the most surprising result is that $g(x,n)h(x,n)$ always seem to divide ${{x+n}\choose{n}}$.

Are these well known properties of the binomial coefficient? Am I wrong in my analysis? Can anyone find a counter example that shows my observations are wrong?


Edit:

Fixed a mistake so that ${x \choose n}$ is now ${{x+n}\choose n}$. Thanks to Collag3n for catching that.

Added 1 more property:

$$\text{lcm}(x+1, x+2, \dots, x+n) \le {{x+n}\choose{n}}\text{lcm}(2,3,\dots,n)$$

Edit 2:

It looks like I had not properly defined $g(x,n)$.

I had written $g(x,n) = \prod\limits_{p\text{ prime & }p \le n \text{ & }p^{v_p(n!)} | f(x,n)} \dfrac{p^{v_p(f(x,n))}}{p^{v_p(n!)}}$ but it should be this:

$$g(x,n) = \prod\limits_{p\text{ prime & }p \le n}\dfrac{p^{v_p(\text{lcm}(x+1, \dots, x+n))}}{p^{v_p(\text{lcm}(2,3,\dots,n))}}$$

I also updated my definition of $v_p(x)$ to accommodate the modified definition of $g(x,n)$.

A big thanks to Collag3n for noticing that the previous definition did not work.

Edit 3:

My definition for $h(x,n)$ was also not properly defined.

I had written $h(x,n) = \prod\limits_{p\text{ prime &}p > n\text{ & }p | f(x,n)}p$ but it should be:

$$h(x,n) = \prod\limits_{p\text{ prime &}p > n\text{ & }p^t | f(x,n)}p$$

For example, $h(47,6) = 49*17*13*53 = 573,937$.