| Finished | ||||||
|---|---|---|---|---|---|---|
| Styxdo | Reckless | material-output-buckets | diff | N=25000 | LLR: 2.93 (-2.25, 2.89) [0.00, 4.00] Games: 7374 W: 2492 L: 2281 D: 2601 Ptnml(0-2): 234, 786, 1469, 931, 267 | S2 Early ply skipping to 28, passed STC on Slim, not resetting optimizer this time. |
| World | Reckless | killer-stage-2 | diff | 8.0+0.08 | LLR: 2.92 (-2.25, 2.89) [0.00, 5.00] Games: 30934 W: 7547 L: 7324 D: 16063 Ptnml(0-2): 125, 3640, 7715, 3861, 126 | extract killer moves into its own stage, speedup? |
| World | Reckless | slim-early-ply-28 | diff | 8.0+0.08 | LLR: 2.90 (-2.25, 2.89) [0.00, 5.00] Games: 17016 W: 4194 L: 4021 D: 8801 Ptnml(0-2): 62, 2062, 4116, 2177, 91 | slim early ply filtering experiment |
| Peregr | Reckless | fds-hindsight | diff | 8.0+0.08 | LLR: 2.92 (-2.25, 2.89) [0.00, 4.00] Games: 29814 W: 7353 L: 7130 D: 15331 Ptnml(0-2): 112, 3508, 7470, 3679, 138 | big brain reduction - fixed (missed the 1024 * ) |
| Peregr | Reckless | fix_noisy | diff | 8.0+0.08 | LLR: 2.93 (-2.25, 2.89) [-4.00, 0.00] Games: 12032 W: 2933 L: 2837 D: 6262 Ptnml(0-2): 54, 1351, 3111, 1445, 55 | |
| World | Reckless | ghi | diff | 8.0+0.08 | LLR: 2.89 (-2.25, 2.89) [0.00, 4.00] Games: 3348 W: 684 L: 591 D: 2073 Ptnml(0-2): 1, 147, 1285, 240, 1 | GHI on fortress book (verification) |
| World | Reckless | ghi | diff | 8.0+0.08 | LLR: 2.97 (-2.94, 2.94) [-4.00, 0.00] Games: 8312 W: 2035 L: 1925 D: 4352 Ptnml(0-2): 32, 946, 2097, 1042, 39 | non-regression on UHO [0.05, 0.05] |
| World | Reckless | ghi | diff | 8.0+0.08 | LLR: 2.94 (-2.25, 2.89) [0.00, 4.00] Games: 3496 W: 728 L: 630 D: 2138 Ptnml(0-2): 5, 150, 1339, 250, 4 | GHI on fortress book |
| World | Reckless | blend_both_s1_s2 | diff | N=25000 | Elo: 1.36 +- 2.42 (95%) [N=40000] Games: 40028 W: 13231 L: 13074 D: 13723 Ptnml(0-2): 1454, 4313, 8268, 4580, 1399 | Blend 640 and 1536 data in both stages |
| Peregr | Reckless | fix_nmp_hindsight | diff | 8.0+0.08 | LLR: 2.92 (-2.25, 2.89) [-4.00, 0.00] Games: 37120 W: 9008 L: 8987 D: 19125 Ptnml(0-2): 165, 4506, 9209, 4503, 177 | |
| World | Reckless | bnp-pv | diff | 8.0+0.08 | LLR: 2.94 (-2.25, 2.89) [-4.00, 0.00] Games: 8106 W: 2027 L: 1916 D: 4163 Ptnml(0-2): 40, 934, 1992, 1049, 38 | |
| World | Reckless | bnp-mc | diff | 8.0+0.08 | LLR: 2.91 (-2.25, 2.89) [0.00, 4.00] Games: 26844 W: 6614 L: 6399 D: 13831 Ptnml(0-2): 122, 3166, 6614, 3415, 105 | |
| Peregr | Reckless | probcut-hindsight | diff | 8.0+0.08 | LLR: 2.89 (-2.25, 2.89) [0.00, 4.00] Games: 60510 W: 14833 L: 14518 D: 31159 Ptnml(0-2): 245, 7164, 15169, 7385, 292 | fix? |
| World | Reckless | retroactive-excluded | diff | 8.0+0.08 | LLR: 3.06 (-2.25, 2.89) [-4.00, 0.00] Games: 47482 W: 11494 L: 11498 D: 24490 Ptnml(0-2): 210, 5754, 11839, 5706, 232 | hindsight lmr is applied twice in SE, fix it |
| World | Reckless | rfp-noisy | diff | 8.0+0.08 | LLR: 2.96 (-2.25, 2.89) [0.00, 4.00] Games: 16384 W: 4134 L: 3949 D: 8301 Ptnml(0-2): 73, 1875, 4117, 2048, 79 | |
| World | Reckless | ghi | diff | 8.0+0.08 | LLR: 3.09 (-2.25, 2.89) [0.00, 4.00] Games: 26042 W: 4995 L: 4838 D: 16209 Ptnml(0-2): 5, 1738, 9380, 1891, 7 | endgame book |
| World | Reckless | spsa-2-200k | diff | 40.0+0.40 | LLR: 2.91 (-2.25, 2.89) [0.00, 4.00] Games: 4336 W: 1068 L: 931 D: 2337 Ptnml(0-2): 2, 448, 1137, 573, 8 | [LTC] final values |
| World | Reckless | pgo | diff | 8.0+0.08 | LLR: 3.03 (-2.25, 2.89) [0.00, 4.00] Games: 4930 W: 1253 L: 1111 D: 2566 Ptnml(0-2): 11, 469, 1364, 609, 12 | profile-guided optimization |
| World | Reckless | v19-219f55d4 | diff | 40.0+0.40 | LLR: 2.98 (-2.25, 2.89) [0.00, 4.00] Games: 19328 W: 4616 L: 4431 D: 10281 Ptnml(0-2): 25, 2224, 4985, 2401, 29 | [LTC] increased WDL in S1 and S2 |
| World | Reckless | v19-219f55d4 | diff | 8.0+0.08 | LLR: 2.93 (-2.25, 2.89) [0.00, 4.00] Games: 7268 W: 1882 L: 1722 D: 3664 Ptnml(0-2): 27, 849, 1745, 963, 50 | [STC] increased WDL in S1 and S2 |
| World | Reckless | spsa-2-sanity-31k | diff | 40.0+0.40 | LLR: 2.90 (-2.25, 2.89) [0.00, 4.00] Games: 7808 W: 1847 L: 1700 D: 4261 Ptnml(0-2): 5, 858, 2036, 995, 10 | spsa values after 31k games |
| World | Reckless | spsa-2-90k | diff | N=12500 | LLR: 4.37 (-2.25, 2.89) [0.00, 4.00] Games: 1780 W: 784 L: 480 D: 516 Ptnml(0-2): 39, 134, 343, 232, 142 | |
| Peregr | Reckless | lmr_stuff1 | diff | 8.0+0.08 | LLR: 2.93 (-2.25, 2.89) [0.00, 4.00] Games: 79640 W: 19446 L: 19071 D: 41123 Ptnml(0-2): 311, 9521, 19851, 9756, 381 | |
| World | Reckless | s1_upper_lr_bound_increase | diff | 8.0+0.08 | LLR: 3.08 (-2.25, 2.89) [0.00, 4.00] Games: 11404 W: 2986 L: 2808 D: 5610 Ptnml(0-2): 60, 1291, 2824, 1465, 62 | Increase S1 training LR Upper bound to initial_lr: 0.002 (up from 0.0015) |
| World | Reckless | fix-more-history-updates | diff | 8.0+0.08 | LLR: 3.08 (-2.25, 2.89) [-4.00, 0.00] Games: 11382 W: 2780 L: 2675 D: 5927 Ptnml(0-2): 40, 1307, 2915, 1366, 63 | as it turns out, we had another case where the history value got beyond the limit |