Id: palomar/PALOMAR-2026-11-21-000016 Url: https://github.com/felixpernegger/oeis-a357513-supercongruence/tree/15ac7d538051d0d06d5f74e159b14a0242860ff3 Json: https://data.palomar-registry.org/entries/PALOMAR-2026-11-21-000016-v1.json Repo: https://github.com/felixpernegger/oeis-a357513-supercongruence Tags: math.NT, msc:11A07, msc:11B65, msc:68V20 Score: 0 Title: OEIS A357513: u_m(p-1) ≡ 0 (mod p^4) for all primes p ≥ 2m + 6 Trust: high Domain: math.NT Source: palomar Authors: Felix Pernegger Reviews: 0 Subject: math.NT Category: research Comments: 0 Disputed: false Division: math Mentions: 0 Verified: true Discussed: 0 Published: 2026-11-21 Formalized: true Page: https://delphi.science/r/palomar/PALOMAR-2026-11-21-000016 For m ≥ 0 let u_m(n) be the numerator of the sum over k = 1, ..., n of binomial(n,k)^2 binomial(n+k,k)^2 / k^(2m+1). Peter Bala conjectured in OEIS A357513 (2022) that u_m(p-1) ≡ 0 (mod p^4) for all primes p, with finitely many exceptions that depend on m. This project gives a Lean 4 / Mathlib proof of that statement: for every m and every prime p ≥ 2m + 6, p^4 divides u_m(p-1). The proof writes T_k = binomial(p-1,k) binomial(p-1+k,k) as T_k^2 ≡ p^2/k^2 - 2p^3/k^3 (mod p^4), reduces the sum to the harmonic-type sums H(s) = sum of k^(-s) over 1 ≤ k ≤ p-1, shows H(2m+4) ≡ 0 (mod p) with the…