Ruzsa’s genus-one problem

witness #1646

N
3,675
|A|
49
exponent log |A| / log N
0.4741
status
current record for this modulus
submitted by
Yongxi (Aaron) Lin
submitted at
2026-08-08 01:48:19 UTC

Elements (49)

0, 20, 25, 296, 301, 308, 321, 326, 440, 446, 450, 465, 471, 475, 1028, 1053, 1171, 1196, 1475, 1500, 1761, 1768, 1786, 1793, 1906, 1910, 1931, 1935, 2503, 2528, 2651, 2676, 2935, 2945, 2960, 2970, 3231, 3233, 3241, 3256, 3258, 3266, 3381, 3385, 3390, 3406, 3410, 3415, 3670

Commentary

Deformation plus conflict repair from the 50-point N=3725 paired-fiber record. Affine-normalize the N=3725 set (multiplier 1191), write fiber bases as m*149+r, and deform 149 -> 147. The 50-point candidate has 10 ordered conflicts. Deleting 1063 reduces this to two conflict classes, both through 1798; deleting 1798 gives a saturated valid 48-set. A 3-for-3 local exchange then yields a 49-set: additionally remove {1038,1773} and insert {308,2503,2528}. The final 49-set is valid modulo 3675 with score 49/sqrt(3675) ≈ 0.80829038.

last edited by Yongxi (Aaron) Lin at 2026-08-08 01:48:19 UTC · history

Log in to add commentary.