← All problems

Liu–Xu conjecture on denominators of Hodge integrals

For each integer g≥0g\ge 0, let DgD_g be the least common denominator of all genus-gg Hodge integrals ∫M‾g,nψ1a1⋯ψnanλb1⋯λbr\int_{\overline{\mathcal M}_{g,n}}\psi_1^{a_1}\cdots\psi_n^{a_n}\lambda_{b_1}\cdots\lambda_{b_r}, where 2g−2+n>02g-2+n>0, ai,bj≥0a_i,b_j\ge 0, and ∑iai+∑jbj=3g−3+n\sum_i a_i+\sum_j b_j=3g-3+n. The conjecture asserts that Dg=T2gD_g=T_{2g}, where T2gT_{2g} is the 2g2g-th Minkowski number.

Equivalent formulations
  1. For every g≥0g\ge 0, the least common denominator DgD_g of all genus-gg Hodge integrals satisfies Dg=(2g+1)!PD_g=(2g+1)!_{\mathbb P}, the Bhargava factorial associated to the set of primes P\mathbb P.

    source: Daniel Holmes, Denominators of Hodge integrals and Minkowski numbers

References

Progress summary

Refreshed
Open · solution reported

A 2026 preprint claims to settle the conjecture by giving the exact denominators of all Hodge integrals, but the result has not been independently checked.

The Liu–Xu conjecture concerns exact universal denominators for genus-gg Hodge integrals and the resulting multiplicities of primes in descendant-integral denominators. The latest source attributes the claimed result to Daniel Holmes.

October 21, 2026 claimed proof

Holmes claims that the least common denominator of all genus-gg Hodge integrals is the Minkowski number T2gT_{2g}, equivalently the Bhargava factorial (2g+1)!P(2g+1)!_{\mathbb{P}}. He states that this proves the Liu–Xu conjecture on the multiplicity of 33 in specified descendant denominators and yields related automorphism-divisibility results. The preprint is unrefereed, and no independent mathematical assessment was retrieved.

Current status (as of October 2026): Holmes's preprint claims the denominator formula and the Liu–Xu conjecture are solved, but the claim remains unverified; no separate classical partial results are recorded here.

Sources

Solutions 0

No solutions have been posted yet.