9
$\begingroup$

Suppose $\binom{n}{k}=\binom{n'}{k'}$ with $k \geq 2$, $k' \geq 2$, $n \geq 2k$ and $n' \geq 2k'$. Does it follow that $n=n'$ and $k=k'$?

EDIT: Yup, $\binom{16}{2}=\binom{10}{3}=120$.

Now I want to ask if there are infinitely many such pairs, but I should probably ask that in a separate question. Thanks!

EDIT 2: for future reference, yes there are infinitely many such coincidences. See Singmaster's Conjecture.

$\endgroup$
2
  • $\begingroup$ Rephrased: "Are there numbers in Pascal's triangle that are equal, apart from the obvious right-left symmetry?" I haven't checked, but I am convinced that there are quite a few. $\endgroup$
    – Arthur
    Nov 25, 2011 at 7:01
  • 3
    $\begingroup$ A related question on MO: mathoverflow.net/questions/17058/factorials-in-pascals-triangle $\endgroup$
    – KCd
    Nov 25, 2011 at 7:29

5 Answers 5

7
$\begingroup$

$$ \binom{3003}{1} = \binom{78}{2} = \binom{15}{5} = \binom{14}{6}. $$

Nobody knows whether any number besides 3003 appears at least eight times in Pascal's triangle.

$\endgroup$
6
$\begingroup$

Let $f(t)$ be the number of ways in which $t$ can be written as a binomial coefficient. Then the question can be restated equivalently as asking whether $f(t) \leqslant 2$ for all $t$. As the other answers point out, there are numerous counterexamples to this strong claim.

It is however likely that a weaker form of this statement is true. Singmaster's conjecture states that there exists some $M \lt \infty$ such that $f(t) \leqslant M$ for all $t$. This is still open. The best known upper bound, due to D. Kane, is $$f(t) = O \left(\frac{(\log t) \cdot (\log \log \log t) }{(\log \log t)^3} \right) .$$ On the other hand, it is consistent with our current knowledge that $f(t)$ never exceeds $8$.

Reference: This MO post by Michael Hardy asks for recent progress on this problem.

$\endgroup$
3
$\begingroup$

There are one or two (nontrivial) choices of $(m,n)$ where $\binom{n}{2} = \binom{m}{3}$. Comparing the sequences on OEIS might be faster than coming up with the correct words to find it in a search engine.

Edit: 120 is in both sequences, with $m=10$ and $n=16$. In OEIS the list of triangular numbers ( http://oeis.org/A000217 ) and tetrahedral numbers ( http://oeis.org/A000292 ) shows that there are no other small examples and I am pretty sure that this pair is known to be the largest integer solution of $3y(y-1)=x(x-1)(x-2)$, and that this was first proved in an elementary way that does not use the theory of elliptic curves.

$\endgroup$
1
  • $\begingroup$ No, there are two more examples of ${n\choose2}={m\choose3}$. See oeis.org/A027568 $\endgroup$ May 2, 2021 at 6:59
3
$\begingroup$

10-choose-4 = 21-choose-2.${}{}{}{}$

$\endgroup$
5
  • $\begingroup$ Did you mean to write $\binom{10}{4} = \binom{21}{2}$ ? $\endgroup$
    – Sasha
    Nov 25, 2011 at 6:59
  • 1
    $\begingroup$ Trying to guess what you meant, since it obviously isn't $\binom{10}{2}=\binom{21}{4}$... $\endgroup$
    – Alon Amit
    Nov 25, 2011 at 7:01
  • $\begingroup$ ${10\choose 4} = {21 \choose 2}$ $\endgroup$ Nov 25, 2011 at 7:02
  • 1
    $\begingroup$ Oops. Miscopied from my source. Thanks for spotting the error and figuring out what I meant. $\endgroup$ Nov 25, 2011 at 10:13
  • $\begingroup$ @Gerry I took the liberty to correct the typo. Hope it's ok. Regards, $\endgroup$
    – Srivatsan
    Nov 25, 2011 at 13:42
2
$\begingroup$

$$\binom{10}{3}=\binom{16}{2}.$$

$\endgroup$
0

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.