← All problems

Erdős Problem #849 — Prescribed multiplicities of binomial coefficients

29 years old · documented by

For an integer aa, let M(a)M(a) be the number of integer pairs (n,k)(n,k) satisfying n≥2n\ge 2, 1≤k≤n/21\le k\le n/2, and (nk)=a\binom{n}{k}=a; equivalently,

M(a)=#{(n,k)∈Z2:n≥2, 1≤k≤⌊n2⌋, (nk)=a}.M(a)=\#\left\{(n,k)\in\mathbb Z^2:n\ge2,\ 1\le k\le\left\lfloor\frac n2\right\rfloor,\ \binom nk=a\right\}.

Is it true that for every integer t≥1t\ge1 there is an integer aa such that M(a)=tM(a)=t? Since all the binomial coefficients in this range are positive, any such aa must be positive. Erdős Problems #849

Background

This is a multiplicity question for entries of Pascal’s triangle, with the symmetry (nk)=(nn−k)\binom nk=\binom n{n-k} removed by restricting to k≤n/2k\le n/2. The problem record attributes the question to Erdős and Gordon, who had considered it “many years” earlier, and identifies it with Singmaster’s conjecture. The examples a=120a=120 and a=3003a=3003 give multiplicities three and four, respectively, in this half-triangle convention. Erdős Problems #849 Erdős, Some problems I presented or planned to present in my shorttalk (1996)

The motivating expectation behind Singmaster’s conjecture is that multiplicities should be bounded by an absolute constant, which would rule out the proposed examples for every tt. The question asks for the stronger opposite phenomenon: whether every positive multiplicity occurs. In the cited account, no example with multiplicity at least five is recorded. Erdős Problems #849

References

Progress summary

Never refreshed

Nothing recorded yet. Refresh searches the literature and the public web for attempts on this problem, and writes the first summary here.

Solutions 0

No solutions have been posted yet.