NEWS · 2026-10-06

P60, P61, P63, P65, P67: 17 instances proved optimal

zzzcy emailed computer-assisted proofs for five families, written with AI assistance as stated. The site read only the certificate data, never ran the attached scripts, wrote its own checker for each family and replayed every certificate. The rows close with their record holders unchanged; zzzcy earns one +2 proof award per family.

P67 variable circles in a fixed-perimeter rectangle: n = 3–9

A complete interval branch-and-exclude over every contact order leaves a single contact structure, fixed by interval Newton, which gives the continuous optimum; n = 3, 4, 5, 7 and 8 have closed forms in radicals. A branch tree over integer offsets shows the current records are the largest nine-decimal sums of radii.

Read the proof →

P63 optimal quadrature on the torus: n = 8, 13, 21, 34

An auxiliary function with nonnegative coefficients gives the continuous lower bound, attained by a Fibonacci-type lattice. At n = 8 the lattice lies on the grid; for the other three it does not, and localisation plus a complete integer enumeration show the current records are the grid minima.

Read the proof →

P65 erasure-robust measurement directions: four rows

For d = 3 the icosahedron’s and dodecahedron’s diagonals are optimal at n = 6 and 10; at d = 4, n = 8 the optimum is the square of a root of an octic. These three are continuous optima. For d = 3, n = 8 the proof is about the nine-decimal grid: the current record is the largest score.

Read the proof →

P60 sixteen lines in five dimensions

Rankin’s bound gives coherence at least 1/√5, which real configurations attain; the proof shows rational coordinates cannot, so the current record is the optimal grid score.

Read the proof →

P61 five complex lines in three dimensions

Five complex lines have coherence at least (√13 − 1)/6, attained by the canonical code in the Game of Sloanes archive; the core of the proof is a two-million-node Bernstein certificate covering the whole phase domain.

Read the proof →

Submitting a proof

Post a proof in the problem’s discussion, or email a larger computer-assisted one. Each adopted new argument earns its contributor one +2 proof award.

zzzcy #308 · All news · Scoring