-
2026-08-26 18:01
Aayush Rajasekaran
extended
{0, 1, 5, 7}
— length 156 (was 155)
-
2026-08-26 17:24
David Renshaw
commented on length 6874 over {0, 1, 4, 10, 12, 17}
Generalized the public lockstep sibling-recycling population search to the sparse alphabet {0,1,4,10,12,17}. Starting from word 310 (length 71), a short benchmark reached 1000, followed by a 15-minute adaptive campaign of 18 forward/reversed population branches (W=4096 and 8192, …
-
2026-08-26 17:24
David Renshaw
extended
{0, 1, 4, 10, 12, 17}
— length 6874 (was 71)
-
2026-08-26 17:20
David Renshaw
commented on length 3257 over {0, 1, 2, 3, 4, 5, 6, 7, 8, 9}
Length 3257 (+599 over word 306). Independent mobility-guided checkpoint/reversal search reached 3164, randomized exact bidirectional suffix repair reached 3227, and chained exact end extensions/repairs reached 3257. Final full-core randomized search for 3258 found none within th…
-
2026-08-26 17:20
David Renshaw
extended
{0, 1, 2, 3, 4, 5, 6, 7, 8, 9}
— length 3257 (was 2658)
-
2026-08-26 17:05
David Renshaw
commented on length 7212 over {0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11}
Population search reconstructed from the public W-word lockstep/sibling-recycling method. Starting from word 258 (length 7205), an 8192-member beta=2.5 population with 250-600-symbol reseeds found 7206; that checkpoint was immediately reseeded into an 8192-member beta=1.5 populat…
-
2026-08-26 17:05
David Renshaw
extended
{0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11}
— length 7212 (was 7205)
-
2026-08-26 16:49
Bryna Kra
opened
{0, 3, 6, 71, 74, 77}
— length 4889
-
2026-08-26 16:35
Bryna Kra
opened
{0, 1, 4, 10, 12, 17}
— length 71
-
2026-08-26 16:22
Bryna Kra
commented on length 36025 over {0, 1, 2, 211, 212, 213}
This is generated from a substitution, but it really comes from a two dimensional example which we know exists. Not a good candidate.
-
2026-08-26 16:22
Bryna Kra
opened
{0, 1, 2, 211, 212, 213}
— length 36025
-
2026-08-26 16:13
Bryna Kra
commented on length 27646 over {0, 1, 2, 184, 185, 186}
this is from a substitution
-
2026-08-26 16:12
Bryna Kra
opened
{0, 1, 2, 184, 185, 186}
— length 27646
-
2026-08-26 16:06
Bryna Kra
opened
{0, 1, 2, 70, 71, 72}
— length 4950
-
2026-08-26 16:02
David Renshaw
commented on length 2658 over {0, 1, 2, 3, 4, 5, 6, 7, 8, 9}
Length 2658 (+2). Found by exact bidirectional suffix repair from the reversed/complemented 2656 record, followed by an exact one-symbol extension. One-hour 16-core follow-up used mobility-guided one- and two-ended tail repair, checkpoint/reversal restarts, beam search, exact suf…
-
2026-08-26 16:02
David Renshaw
extended
{0, 1, 2, 3, 4, 5, 6, 7, 8, 9}
— length 2658 (was 2656)
-
2026-08-26 16:00
Aayush Rajasekaran
commented on length 12187 over {0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13}
Iterated local search: repeatedly restart from a log-uniform random truncation of the best word, then extend with a randomized DFS whose letter choice is weighted by a 1-step lookahead (how many letters stay legal next), with bounded backtracking plus random kicks; several cooper…
-
2026-08-26 16:00
Aayush Rajasekaran
extended
{0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13}
— length 12187 (was 12181)
-
2026-08-26 16:00
Aayush Rajasekaran
commented on length 10776 over {0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12}
Iterated local search: repeatedly restart from a log-uniform random truncation of the best word, then extend with a randomized DFS whose letter choice is weighted by a 1-step lookahead (how many letters stay legal next), with bounded backtracking plus random kicks; several cooper…
-
2026-08-26 16:00
Aayush Rajasekaran
extended
{0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12}
— length 10776 (was 10718)
-
2026-08-26 15:53
Aayush Rajasekaran
commented on length 12880 over {0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14}
Iterated local search, continued run: restart from a log-uniform random truncation of the best word, randomized DFS with 1-step-lookahead letter choice weighted by remaining free letters, bounded backtracking with kicks, cooperative workers sharing a best-word file.
-
2026-08-26 15:53
Aayush Rajasekaran
extended
{0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14}
— length 12880 (was 12831)
-
2026-08-26 15:45
Bryna Kra
extended
{0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15}
— length 19200 (was 19112)
-
2026-08-26 15:42
Aayush Rajasekaran
commented on length 12831 over {0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14}
Iterated local search: restart from a log-uniform random truncation of the best word, then randomized DFS with 1-step-lookahead letter choice (maximise remaining free letters), bounded backtracking with kicks. Cooperative workers sharing a best-word file.
-
2026-08-26 15:42
Aayush Rajasekaran
extended
{0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14}
— length 12831 (was 12784)
-
2026-08-26 15:10
Aayush Rajasekaran
commented on length 1156 over {0, 1, 2, 3, 4, 5, 6, 7}
Iterated tail-release with reversal, mobility-guided, following the method David Renshaw described on the {0..6} and {0,1,2,3,4,5,7} words. Each round: restore the best word, reverse it with probability 1/2 (additive-square-freeness is reversal-invariant, so this turns the untouc…
-
2026-08-26 15:10
Aayush Rajasekaran
extended
{0, 1, 2, 3, 4, 5, 6, 7}
— length 1156 (was 1117)
-
2026-08-26 15:07
David Renshaw
commented on length 1117 over {0, 1, 2, 3, 4, 5, 6, 7}
Fifteen-minute search campaign for alphabet {0,...,7}. I lifted the newly verified 1,102-symbol {0,1,2,3,4,5,7} construction by retaining its longest prefix accepting 6, producing a valid exact-alphabet seed of length 1,094. Ten parallel mobility-guided randomized DFS workers the…
-
2026-08-26 15:07
David Renshaw
extended
{0, 1, 2, 3, 4, 5, 6, 7}
— length 1117 (was 1081)
-
2026-08-26 14:56
Bryna Kra
commented on length 6042 over {0, 1, 2, 97, 98, 99}
This is generated by a substitution