← All problems

Steinerberger’s finite-range frequency-function conjecture

For a bounded real-valued function ff for which the supremum below is attained, define Arf(x)=12r∫x−rx+rf(t) dtA_r f(x)=\frac{1}{2r}\int_{x-r}^{x+r}f(t)\,dt for r>0r>0, A0f(x)=f(x)A_0f(x)=f(x), and rf(x)=min⁡{r≥0:Arf(x)=sup⁡s≥0Asf(x)}r_f(x)=\min\{r\geq 0:A_rf(x)=\sup_{s\geq 0}A_sf(x)\}. Steinerberger's conjecture asks whether range⁡(rf)\operatorname{range}(r_f) being finite implies that ff is a single sinusoidal function, namely whether there exist constants a,b,c,d∈Ra,b,c,d\in\mathbb{R} such that f(x)=a+bsin⁡(cx+d)f(x)=a+b\sin(cx+d) for all x∈Rx\in\mathbb{R}.

References

Progress summary

Refreshed
Open · solution reported

A new preprint claims the conjecture is false in general, while proving a narrower version for especially regular periodic functions, but neither claim has been independently checked.

Steinerberger’s conjecture says that a bounded real-valued function whose frequency function takes only finitely many values must be a sine-type function. A 2014 paper established related rigidity results but explicitly left this broader statement open.

Known results

  • Periodic f∈Cα(R,R)f\in C^\alpha(\mathbb{R},\mathbb{R}) with α>1/2\alpha>1/2 and a fixed critical averaging radius is sine-type (Steinerberger, 2014).
  • Under the same regularity, at most two possible maximal-interval radii also force f(x)=a+bsin⁡(cx+d)f(x)=a+b\sin(cx+d) (Steinerberger, 2014).
  • The finite-range statement for bounded f∈L∞(R)f\in L^\infty(\mathbb{R}) was explicitly left open (Steinerberger, 2014).

October 2026 preprint

Carlos Garzón and José Madrid claim a counterexample outside quasianalytic classes, thereby refuting the general characterization, and claim the characterization under a stronger quasianalytic hypothesis for periodic functions. The preprint is unrefereed and has no independent confirmation in the retrieved material.

Current status (as of October 2026): The general conjecture is claimed false, and the stronger quasianalytic periodic case is claimed proved, but both claims remain unverified.

Sources

Solutions 0

No solutions have been posted yet.