Ruzsa’s genus-one problem

witness #1765

N
14,637
|A|
62
exponent log |A| / log N
0.4303
status
current record for this modulus
submitted by
Yongxi (Aaron) Lin
submitted at
2026-08-08 14:43:20 UTC

Elements (62)

0, 1, 5, 14, 18, 19, 39, 84, 104, 105, 109, 118, 122, 123, 143, 188, 520, 521, 525, 534, 538, 539, 559, 604, 624, 625, 629, 638, 642, 643, 663, 708, 2912, 2913, 2917, 2926, 2930, 2931, 2951, 2996, 3120, 3121, 3125, 3134, 3138, 3139, 3159, 3204, 3432, 3433, 3437, 3446, 3450, 3451, 3471, 3516, 3640, 3641, 3645, 3654, 3658, 3659

Commentary

Deletion compression from witness #1728 (64 points at N=15001). At N=14637 the exact conflict-hitting problem on the original 64-point seed has optimum 2 deletions; removing 3679 and 3724 leaves this valid 62-point set. No single deletion can make the 64-point seed valid at any modulus between 3725 and the zero-deletion endpoint 14664. This 62-set is also insertion-maximal among all residues mod 14637.

last edited by Yongxi (Aaron) Lin at 2026-08-08 14:43:20 UTC · history

Log in to add commentary.