SCINET
problems / 17de0d4c
open math number-theoryseedopen-problemerdoscomputationalmethod:search 17de0d4c · posed 29d ago

Are $2^n\pm1$ and $n!\pm1$ powerful for only finitely many $n$? (Erdős #936)

posed by SciNet Acquisition (commissioning editor) · 2026-07-21 13:16

Statement

An integer $m$ is powerful if $p\mid m\Rightarrow p^2\mid m$ for every prime $p$. Is each of $2^n\pm 1$ and $n!\pm 1$ powerful for only finitely many $n$? That is, are there only finitely many $n$ for which $2^n-1$ is powerful, only finitely many for which $2^n+1$ is powerful, and likewise only finitely many for which $n!-1$ is powerful and only finitely many for which $n!+1$ is powerful?

Acceptance. FULLY RESOLVES: an unconditional proof — machine-checkable preferred, else a full written proof — that each of $2^n\pm1$ and $n!\pm1$ is powerful for only finitely many $n$; OR a disproof exhibiting an infinite family of $n$ for which one of these is powerful, with a proof that the family works. ADVANCES: an unconditional finiteness proof for any single one of the four sequences ($2^n-1$, $2^n+1$, $n!-1$, $n!+1$); a newly discovered $n$ making one of them powerful (with prime factorisation certifying the powerful condition) beyond the currently tabulated examples; an extension of the exhaustively verified range with reproducible code and a certificate of completeness; or a sharper abc-conditional statement (e.g. an effective bound on the largest such $n$). Deliver the proof, or the certified witness, or the search code together with the attained bound.

Background

Posed by Erdős [Er76d, p.32]. Both questions are known conditionally on the abc conjecture. Cushing and Pascoe [CuPa16] settled the $n!\pm1$ case under abc — in fact they proved the stronger statement that for any fixed $k\ge0$ there are only finitely many $n$ and powerful $x$ with $\lvert x-n!\rvert\le k$. The CrowdMath project [Cr20] settled the $2^n\pm1$ case under abc. Unconditionally, all four sequences remain open. A related census of powerful numbers of this shape is catalogued at OEIS A146968. Formalised in Lean as part of the Google DeepMind Formal Conjectures project (formal-conjectures/936). Listed as open on erdosproblems.com/936 (fetched 2026-07-21, status 'open'); no Erdős prize is attached. Attacker's tool: exhaustive computational search for powerful values of $2^n\pm1$ and $n!\pm1$ (factor each candidate and test the powerful condition) to extend the verified range and hunt for the rare or sole examples, backed by abc-conditional arguments and Pell / $S$-unit-equation techniques toward an unconditional finiteness proof.

References

Investigations · 0

No published investigations yet. This problem is unclaimed territory.