The odd-part ratio under slow variation: a note on the printed Hilberdink problem
Read the paper (PDF) · Download PDF · Supporting ZIP · Zenodo record
Plain-English explanation · Read on this page · Collaboration
Abstract
For a nonzero nonnegative multiplicative function whose summatory function is slowly varying, the ratio of its odd-part sum to its full sum converges to the reciprocal of the unweighted local Euler sum at 2. A short monotonicity argument covers both finite and infinite local sums. This answers the limit-existence question as printed in Hilberdink's Problem 3 in Oberwolfach Report 51/2017. The report's subsequent numerical prediction is incompatible with its printed hypothesis. We exhibit the discrepancy and distinguish the printed question from a possible index-one reformulation, which is not addressed here.
The idea in plain English
Give each positive integer a nonnegative weight, and add the weights up to a cutoff. Now keep only the odd integers. This note shows that the fraction of weight left over has a limit whenever doubling the cutoff barely changes the total weight in the long run.
The limit depends only on the weights assigned to powers of two. The proof also covers an infinite local sum, when the limiting fraction is zero, and a bounded total sum.
The precise result
Let f be a nonzero nonnegative multiplicative function, let F(x) = Σₙ≤ₓ f(n), and let F₂(x) be the same sum restricted to odd n. If F(2x)/F(x) tends to 1, then
F₂(x)/F(x) → 1/A, where A = Σₖ≥₀ f(2ᵏ) and 1/∞ = 0.
For a nondecreasing summatory function, this doubling condition is equivalent to slow variation: F(λx)/F(x) tends to 1 for every fixed λ > 0.
Why the result holds
Every positive integer has a unique form 2ᵏm with m odd. Multiplicativity therefore splits the full sum into copies of the odd sum. Positivity and monotonicity give matching bounds after any fixed number of powers of two is included.
The argument takes limits in finite bounds. It does not require the total sum to diverge or exchange an infinite sum with a limit.
Scope and limitations
The result concerns the slow-variation hypothesis printed in Hilberdink’s Problem 3. The same source predicts a weighted constant that conflicts with that hypothesis: f(n) = 1/n gives the actual limit 1/2, whereas the printed prediction gives 3/4.
A possible intended formulation would instead assume F(λx)/F(x) → λ. That is a different question and is not resolved here; the source’s intended wording has not been established.
Background and prior work
- Hilberdink’s Problem 3, Oberwolfach Report 51/2017 — The printed existence question and accompanying prediction appear on pp. 3065–3066.
- Yeats, Asymptotic Density in Combined Number Systems (2002) — Related factor-ratio results, including Proposition 43, under hypotheses on the factor sums.
- Yeats, A Multiplicative Analogue of Schur’s Tauberian Theorem (2003) — Related Dirichlet-convolution Tauberian results; their hypotheses differ from the formulation proved here.
Status and AI use
This is an unrefereed preprint. AI tools, including OpenAI Codex, assisted with the research, calculations, source comparisons and writing. AI review is not conventional human peer review. Alec Kriebel is responsible for the contents.
Read the paper
Supporting materials
Download unchanged native support ZIP · Zenodo original ZIP
The archive includes the editable manuscript and supporting materials.
Collaborate on this result
Useful contributions could include independent mathematical verification, a substantive extension, or work toward a journal submission. Authorship and responsibilities would reflect the contributions actually made.
Cite this paper
Alec Kriebel. (2026). The odd-part ratio under slow variation: a note on the printed Hilberdink problem. Version 1.0.1 [Preprint]. Zenodo. 10.5281/zenodo.22983161.
Licensing
The manuscript, authored audit notes, website text and metadata are licensed under CC BY 4.0. The Python scripts and website code use the MIT License, as specified in LICENSES.md.
Third-party publications retain their own rights.