← All problems

Common-root conjecture for truncated binomials

22 years old · traced to

For integers dd and 0<k<m<d0<k<m<d, let Pk;d(x)P_{k;d}(x) and Pm;d(x)P_{m;d}(x) denote the truncated binomial polynomials used in the paper. Common-root conjecture. The polynomials Pk;d(x)P_{k;d}(x) and Pm;d(x)P_{m;d}(x) do not have common roots. This is the additional condition needed for non-degeneracy in the case N=2N=2; the source does not establish it for all 0<k<m<d0<k<m<d.

References
  1. I. Scherbak, “Intersections of Schubert varieties and highest weight vectors in tensor products of sl_N+1-representations”, arXiv:math/0409329 (2005).

Progress summary

Refreshed
Open · solution reported

A proof has been reported, but no retrieved source independently checks it, so the conjecture is not verified.

The conjecture says that the truncated binomials Pk;dP_{k;d} and Pm;dP_{m;d} have no common root whenever 0<k<m<d0<k<m<d. A 2004 paper reports that Inna Scherbak later resolved the related question using representation theory, but does not give the proof.

Known results

  • The 2004 paper proves that, for prime dd, each Pd,kP_{d,k} is irreducible for 1≤k≤d−11\le k\le d-1.
  • It also proves that the number of distinct roots is asymptotic to n2/2n^2/2 as n→∞n\to\infty.
  • Scherbak's reported representation-theoretic resolution is not supplied or independently assessed in the retrieved record.

March 2026 claimed solution

A March 5, 2026 report claims that Carl Pomerance gave a proof, with ChatGPT producing an argument; this remains unverified. A submission dated October 24, 2026 cites Peikai Qi's paper and claims a proof covering the interior and boundary cases, but that paper was not independently retrieved.

Community submission (unverified)

A submitted argument claims that Qi's Theorem 3.5 gives coprimality after the stated change of variables, with Remark 3.6 handling ℓ=d−1\ell=d-1; its correctness is unverified.

Current status (as of October 2026): The conjecture has reported proofs, including a recent submitted claim, but no retrieved source verifies them, so the general statement remains open.

Sources

Solutions 1

Scherbak's common-root conjecture is proved in Peikai Qi, Partial Results on Common Roots of Positive Derivatives of Geometric Sums (2026). Theorem 3.5 establishes coprimality for a more general family of polynomials. The conjecture follows by a change of variables, with the remaining boundary case handled in Remark 3.6.See full solutionHide full solution

The conjecture is resolved in the following paper:

Peikai Qi, Partial Results on Common Roots of Positive Derivatives of Geometric Sums (2026).

Full paper (PDF)

Define

Hd,r(z)=∑j=0d(r+jj)zj.H_{d,r}(z)=\sum_{j=0}^{d}\binom{r+j}{j}z^j.

Theorem 3.5 establishes that

gcd⁡(Hd,r,He,s)=1(d<e, r≥s≥1).\gcd(H_{d,r},H_{e,s})=1 \qquad(d<e,\ r\ge s\ge1).

Using the identity

PN,k(w)=(1+w)kHk,N−k−1(w1+w),P_{N,k}(w)=(1+w)^kH_{k,N-k-1}\left(\frac{w}{1+w}\right),

the result implies that PN,kP_{N,k} and PN,ℓP_{N,\ell} have no common roots whenever 0<k<ℓ≤N−20<k<\ell\le N-2.

The remaining boundary case ℓ=N−1\ell=N-1 is established in Remark 3.6.

Consequently, Scherbak's common-root conjecture holds for all 0<k<ℓ<N0<k<\ell<N.