Recent activity

2026-08-28 14:56 David Renshaw commented on length 26654 over {0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20}
Two-hour session of variable-length annealing (anneal2: sum-preserving interior transfer moves with sawtooth reheat, wide leash), two fleet generations seeded from the previous record and its reversal/complement symmetries, plus exact DFS tail polish. 26581 -> 26654. Verified eve…
2026-08-28 14:56 David Renshaw extended {0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20} — length 26654 (was 26581)
2026-08-28 13:05 David Renshaw commented on length 3805 over {0, 1, 2, 3, 4, 5, 6, 7, 8, 9}
Two-hour follow-up annealing session (two wide-leash fleet generations plus exact DFS tail polish to depth 70). 3797 -> 3805. This basin is now heavily optimized; gains are getting expensive.
2026-08-28 13:05 David Renshaw extended {0, 1, 2, 3, 4, 5, 6, 7, 8, 9} — length 3805 (was 3797)
2026-08-28 12:36 David Renshaw commented on length 14401 over {0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14}
One-hour hybrid population-reseed, exact suffix-release, and joint tail-repair campaign from the 14,300 record on {0,...,14}. Ten parallel searches across both orientations, adaptive reseeding, endpoint mutation probes, diverse letter and mobility biases, and tail ranges from 15 …
2026-08-28 12:36 David Renshaw extended {0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14} — length 14401 (was 14300)
2026-08-28 11:34 David Renshaw commented on length 2380 over {0, 1, 2, 3, 4, 5, 6, 7}
Seven-hour improvement campaign from the 2,245-symbol record on {0,...,7}. Guide-preserving rollout beam search, alternating endpoint reversal, exact suffix releases through depth 200, focused iterated mobility DFS, and adaptive fine-resolution endpoint beams moved the independen…
2026-08-28 11:34 David Renshaw extended {0, 1, 2, 3, 4, 5, 6, 7} — length 2380 (was 2245)
2026-08-28 09:51 David Renshaw commented on length 3347 over {0, 1, 2, 3, 4, 5, 6, 7, 8}
Final result of an 8-hour overnight annealing session: eight anneal2 fleet generations, each reseeded from the previous generation's best word and its reversal/complement symmetries, with exact DFS tail polish between generations. Wide-leash (ANNEAL_SLACK 96-192) and high-MU sett…
2026-08-28 09:51 David Renshaw extended {0, 1, 2, 3, 4, 5, 6, 7, 8} — length 3347 (was 3223)
2026-08-28 07:51 David Renshaw commented on length 3223 over {0, 1, 2, 3, 4, 5, 6, 7, 8}
Overnight annealing run, sixth fleet generation (wide leash, high MU). 3108 -> 3223.
2026-08-28 07:51 David Renshaw extended {0, 1, 2, 3, 4, 5, 6, 7, 8} — length 3223 (was 3108)
2026-08-28 06:44 David Renshaw commented on length 3108 over {0, 1, 2, 3, 4, 5, 6, 7, 8}
Overnight annealing run, fifth fleet generation (wide-leash exploration). 3049 -> 3108.
2026-08-28 06:44 David Renshaw extended {0, 1, 2, 3, 4, 5, 6, 7, 8} — length 3108 (was 3049)
2026-08-28 04:56 David Renshaw commented on length 3049 over {0, 1, 2, 3, 4, 5, 6, 7, 8}
Overnight annealing run, third fleet generation. 2941 -> 3049.
2026-08-28 04:56 David Renshaw extended {0, 1, 2, 3, 4, 5, 6, 7, 8} — length 3049 (was 2941)
2026-08-28 03:09 David Renshaw commented on length 2941 over {0, 1, 2, 3, 4, 5, 6, 7, 8}
Variable-length annealing (anneal2 sawtooth fleet) seeded from the previous record and its reversal/complement symmetries, plus exact DFS tail polish. First generation of an overnight run. 2886 -> 2941.
2026-08-28 03:09 David Renshaw extended {0, 1, 2, 3, 4, 5, 6, 7, 8} — length 2941 (was 2886)
2026-08-28 01:39 David Renshaw commented on length 3797 over {0, 1, 2, 3, 4, 5, 6, 7, 8, 9}
One more 30-minute annealing generation seeded from the 3781 word, plus exact DFS tail polish. Closes out the 4-hour session: 3262 -> 3797 overall.
2026-08-28 01:39 David Renshaw extended {0, 1, 2, 3, 4, 5, 6, 7, 8, 9} — length 3797 (was 3781)
2026-08-28 01:09 David Renshaw commented on length 3781 over {0, 1, 2, 3, 4, 5, 6, 7, 8, 9}
Final result of a 4-hour annealing session: three generations of anneal2 fleets (sum-preserving transfer moves, sawtooth reheat), each reseeded from the previous generation's best word and its reversal/complement symmetries, with exhaustive DFS tail polish (depth 70, both orienta…
2026-08-28 01:09 David Renshaw extended {0, 1, 2, 3, 4, 5, 6, 7, 8, 9} — length 3781 (was 3660)
2026-08-27 23:58 David Renshaw commented on length 3660 over {0, 1, 2, 3, 4, 5, 6, 7, 8, 9}
Continued annealing run, second reseed generation. Interim result. 3443 -> 3660.
2026-08-27 23:58 David Renshaw extended {0, 1, 2, 3, 4, 5, 6, 7, 8, 9} — length 3660 (was 3443)
2026-08-27 22:50 David Renshaw commented on length 3443 over {0, 1, 2, 3, 4, 5, 6, 7, 8, 9}
Variable-length annealing (anneal2, sum-preserving transfer moves, sawtooth reheat) seeded from the previous record and its reversal/complement symmetries; exact DFS tail polish. Interim result of an ongoing run. 3262 -> 3443.
2026-08-27 22:50 David Renshaw extended {0, 1, 2, 3, 4, 5, 6, 7, 8, 9} — length 3443 (was 3262)
2026-08-27 21:49 David Renshaw commented on length 6316 over {0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10}
Variable-length annealing (anneal2, sum-preserving transfer moves with sawtooth reheat), seeded from the previous record and its reversal/complement symmetries, followed by exhaustive DFS tail polish to depth 70 in both orientations. 5875 -> 6316.
2026-08-27 21:49 David Renshaw extended {0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10} — length 6316 (was 5875)
2026-08-27 19:19 David Renshaw commented on length 2886 over {0, 1, 2, 3, 4, 5, 6, 7, 8}
Rollout-scored beam search from scratch (beam over prefixes, each scored by the max length of randomized mobility-DFS rollouts) reached 2532; extending the word's front via search_iterated on the reversed word, then a bank of search_iterated polishers, reached 2886.
2026-08-27 19:19 David Renshaw extended {0, 1, 2, 3, 4, 5, 6, 7, 8} — length 2886 (was 2678)