← All problems

Velani’s conjecture on dyadic approximation in the middle-third Cantor set

Let K={∑j=1∞aj3−j:aj∈{0,2}}K=\left\{\sum_{j=1}^{\infty}a_j3^{-j}:a_j\in\{0,2\}\right\} be the middle-third Cantor set, let μ\mu be its Cantor probability measure, and put γ=log⁡2/log⁡3\gamma=\log 2/\log 3. For an approximation function ψ:N→[0,∞)\psi:\mathbb{N}\to[0,\infty) and y∈Ry\in\mathbb{R}, define W2(ψ,y)W_2(\psi,y) by W2(ψ,y)={x∈K:∣x−p+y2n∣<ψ(2n) for infinitely many (p,n)∈Z×N}W_2(\psi,y)=\left\{x\in K:\left|x-\frac{p+y}{2^n}\right|<\psi(2^n)\text{ for infinitely many }(p,n)\in\mathbb{Z}\times\mathbb{N}\right\}. Velani's conjecture asserts the zero-one law

μ(W2(ψ,y))={0,∑n=1∞2nγψ(2n)γ<∞,1,∑n=1∞2nγψ(2n)γ=∞,\mu\bigl(W_2(\psi,y)\bigr)=\begin{cases}0,&\displaystyle\sum_{n=1}^{\infty}2^{n\gamma}\psi(2^n)^{\gamma}<\infty,\\1,&\displaystyle\sum_{n=1}^{\infty}2^{n\gamma}\psi(2^n)^{\gamma}=\infty,\end{cases}

with the original homogeneous formulation corresponding to y=0y=0; the conjecture concerns arbitrary approximation functions, with the appropriate monotonicity restriction in the original formulation.

References

Source: arXiv

Progress summary

Refreshed
Open · progress reported

The full conjecture remains open, while recent papers have substantially narrowed the gap between known bounds and the predicted threshold.

Velani’s conjecture predicts a zero-one law for dyadic approximation to points in the middle-third Cantor set, for arbitrary approximation functions. The general inhomogeneous form concerns the measure of W2(ψ,y)W_2(\psi,y) and extends Velani’s original monotone, homogeneous formulation.

Known results

  • For ψ(n)=n−τ\psi(n)=n^{-\tau}, convergence was proved above an earlier threshold, with full measure known only for 0<τ≤0.010<\tau\leq0.01 (Baker, 2025; earlier work by Allen, Chow, and Yu, 2023).
  • A 2022 result obtained measure zero for τ>1/γ−0.03\tau>1/\gamma-0.03, where γ=log⁡2/log⁡3\gamma=\log 2/\log 3.
  • Quantitative counting results are known for the special scale n−0.012−nn^{-0.01}2^{-n}, but do not establish the all-function law.

2026 partial bounds

Two 2026 preprints improve the power-law ranges: one proves measure zero for τ>1/γ−(1−γ)/(3−γ)≈1.429\tau>1/\gamma-(1-\gamma)/(3-\gamma)\approx1.429 and full measure for 0<τ<γ/12≈0.0520<\tau<\gamma/12\approx0.052; another obtains zero for τ>2−γ\tau>2-\gamma and full measure for 0<τ<(1−γ)/2≈0.1845350<\tau<(1-\gamma)/2\approx0.184535. Neither reaches the conjectured threshold τ=1\tau=1.

Current status (as of October 2026): The all-function conjecture remains open; only partial power-law results are established, and a reported complete-solution claim lacks independent mathematical assessment.

Sources

Solutions 0

No solutions have been posted yet.