Beluhov’s Conjecture on the Tchoukaillon Array

Let Ti,jT_{i,j} denote the entry in row ii and column jj of the Tchoukaillon array. Then, as i+ji+j\to\infty with i,j0i,j\ge 0,

Ti,j(πi+2j)24π.T_{i,j}\sim \frac{(\pi i+2j)^2}{4\pi}.

Equivalent formulations 1

Other statements of this same problem, merged from separate entries. Each is equivalent to the statement above — proving any one settles them all.

  1. Square-root form of Beluhov's conjecture

    Equivalently, as i+ji+j\to\infty with i,j0i,j\ge 0,

    Ti,jπ2i+1πj.\sqrt{T_{i,j}}\sim \frac{\sqrt{\pi}}{2}i+\frac{1}{\sqrt{\pi}}j.

    source: Li, Shisheng, “Asymptotics of the Tchoukaillon array and a conjecture of Beluhov”

Sources & referencesView supporting material

Primary source

arXiv

Additional references

Progress summary

Refreshed
Claimed solved

A new preprint claims to prove the conjecture, but the proof has not yet been independently checked.

Beluhov’s conjecture describes the quadratic growth of the Tchoukaillon array from its recursion; it was relayed by Knuth and arose from numerical evidence. The new claim proves the asymptotic uniformly in both indices.

Known results

  • On the row edge, Erdős–Jabotinsky proved an O(j4/3)O(j^{4/3}) error bound and conjectured O(j)O(j).
  • Broline–Loeb later proved the edge estimate s(n)=n2/π+O(n)s(n)=n^2/\pi+O(n).
  • The new preprint says these edge results fit the broader two-dimensional asymptotic.

August 2026 claimed proof

Li and Shisheng claim, uniformly for i,j0i,j\ge 0 with N=i+j+1N=i+j+1\to\infty, that Чi,j=(πi+2j+2)24π+O(N4/3)\mathrm{Ч}_{i,j}=\frac{(\pi i+2j+2)^2}{4\pi}+O(N^{4/3}), implying Beluhov’s conjectured quadratic asymptotic. The paper also derives geometric and algorithmic corollaries, while the sharper O(N)O(N) error term remains open. ChatGPT and Claude Code assisted with exposition, experiments, and discussion; they are not credited with independently producing the proof.

Current status (as of August 2026): The conjecture has a new claimed uniform proof with an O(N4/3)O(N^{4/3}) error term, but remains unverified; the sharper O(N)O(N) estimate is open.

Sources

Solutions 0

No solutions have been posted yet.