← All problems

Open problem on the second-order Riemann difference

Given coefficients a,b,ca,b,c and nodes q,r,sq,r,s, determine exactly when the finite difference D(f,h)=af(qh)+bf(rh)+cf(sh)D(f,h)=a f(qh)+b f(rh)+c f(sh) has the Marcinkiewicz–Zygmund property: for every measurable function ff, does D(f,h)=o(h)D(f,h)=o(h) and f(h)=o(1)f(h)=o(1) as h→0h\to0 imply f(h)=o(h)f(h)=o(h) as h→0h\to0?

References

Source: arXiv

Progress summary

Refreshed
Open · solution reported

A recent preprint claims to settle the second-order case by giving an exact test, but the result has not been independently verified.

The problem asks when the second-order Riemann difference has the Marcinkiewicz–Zygmund property. A preprint claims a complete criterion for the relevant three-node differences and applies it to this previously open case.

Known results

For geometric-node Riemann differences of order nn, the stated criterion is absence of characteristic roots in the closed annulus {z∈C:∣q∣n−1<∣z∣≤∣q∣n}\{z\in\mathbb{C}: |q|^{n-1}<|z|\leq |q|^n\}; the preprint also treats symmetric and antisymmetric three-term cases.

Three-term classification preprint (November 2026 report)

Hajrudin Fejzić and Chris Freiling report necessary-and-sufficient coefficient-and-node conditions for the MZ property of three-node differences, applying them to the second-order Riemann difference. The result is from an unrefereed preprint; no independent verification, correction, or retraction was found.

Current status (as of November 2026): A preprint claims the second-order case is settled by a necessary-and-sufficient criterion, but that claim remains unverified.

Sources

Solutions 0

No solutions have been posted yet.