| Finished | ||||||
|---|---|---|---|---|---|---|
| Warreg | Pea | lazy-acc-updates | diff | 4.0+0.04 | LLR: 2.96 (-2.94, 2.94) [0.00, 5.00] Games: 1282 W: 382 L: 249 D: 651 Ptnml(0-2): 12, 105, 298, 190, 36 | fixed bug with null moves in NMP |
| Warreg | Pea | lazy-acc-updates | diff | 4.0+0.04 | LLR: -3.06 (-2.94, 2.94) [0.00, 5.00] Games: 724 W: 117 L: 258 D: 349 Ptnml(0-2): 23, 146, 146, 43, 4 | Implemented lazy accumulator updates. |
| Warreg | Pea | fused-acc-updates | diff | 4.0+0.04 | LLR: 2.97 (-2.94, 2.94) [0.00, 5.00] Games: 1716 W: 493 L: 360 D: 863 Ptnml(0-2): 18, 154, 404, 241, 41 | Implemented fuse updates, see https://www.chessprogramming.org/NNUE#Fused_Updates |
| Warreg | Pea | simd-in-evaluate-nnue | diff | 4.0+0.04 | LLR: 2.96 (-2.94, 2.94) [0.00, 5.00] Games: 5592 W: 1410 L: 1264 D: 2918 Ptnml(0-2): 64, 636, 1271, 740, 85 | simd for evaluate function as for scelu function; expected speedup is around 10% in nps |
| Warreg | Pea | pruning-conditions-cutnode | diff | 4.0+0.04 | LLR: 3.03 (-2.94, 2.94) [0.00, 5.00] Games: 8850 W: 2290 L: 2119 D: 4441 Ptnml(0-2): 149, 1015, 1953, 1132, 176 | some lmr tuned |
| Warreg | Pea | pruning-conditions-cutnode | diff | 4.0+0.04 | LLR: -3.06 (-2.94, 2.94) [0.00, 5.00] Games: 19808 W: 4877 L: 4920 D: 10011 Ptnml(0-2): 338, 2405, 4470, 2344, 347 | cutnode condition to nmp,pc |
| Warreg | Pea | pruning-conditions-cutnode | diff | 4.0+0.04 | LLR: -3.02 (-2.94, 2.94) [0.00, 5.00] Games: 8646 W: 2090 L: 2182 D: 4374 Ptnml(0-2): 178, 1045, 1933, 1025, 142 | cutnode condition to nmp,pc and iir |
| Warreg | Pea | lmr-more-factors | diff | 4.0+0.04 | LLR: 1.16 (-2.94, 2.94) [0.00, 5.00] Games: 9630 W: 2437 L: 2344 D: 4849 Ptnml(0-2): 162, 1157, 2087, 1244, 165 | added various terms to LMR, like improving,cutnode and correction scaling Will need to be tuned |
| Warreg | Pea | reduction-if-improving | diff | 4.0+0.04 | LLR: -0.18 (-2.94, 2.94) [0.00, 5.00] Games: 7500 W: 1873 L: 1847 D: 3780 Ptnml(0-2): 151, 872, 1664, 926, 137 | reduce more if not improving 2 |
| Warreg | Pea | check-evasions-qs | diff | 4.0+0.04 | LLR: -0.50 (-2.94, 2.94) [0.00, 5.00] Games: 15338 W: 3810 L: 3763 D: 7765 Ptnml(0-2): 250, 1901, 3338, 1912, 268 | generate all moves when in check in qs |
| Warreg | Pea | big-tune * | diff | 10.0+0.10 | Tuning 51 Parameters 1847/2000 Iterations 59116/64000 Games Played | big tune go brr big flop ha ha |
| Warreg | Pea | big-tune | diff | 4.0+0.04 | LLR: -2.96 (-2.94, 2.94) [-5.00, 0.00] Games: 11550 W: 2712 L: 2886 D: 5952 Ptnml(0-2): 198, 1432, 2654, 1328, 163 | added stop to nmp |
| Warreg | Pea | big-tune-tuned | diff | 8.0+0.08 | LLR: -2.97 (-2.94, 2.94) [0.00, 5.00] Games: 1216 W: 238 L: 359 D: 619 Ptnml(0-2): 31, 181, 285, 100, 11 | test after 942/2000 iteration |
| Warreg | Pea | nmp-verification-tune | diff | 8.0+0.08 | LLR: -0.57 (-2.94, 2.94) [0.00, 5.00] Games: 6882 W: 1605 L: 1599 D: 3678 Ptnml(0-2): 87, 853, 1569, 831, 101 | params from tune |
| Warreg | Pea | nmp-verification-tune | diff | 8.0+0.08 | LLR: 1.78 (-2.94, 2.94) [0.00, 5.00] Games: 10180 W: 2415 L: 2299 D: 5466 Ptnml(0-2): 126, 1189, 2360, 1273, 142 | test if tuned is better than not tuned |
| Warreg | Pea | nmp-verification | diff | 8.0+0.08 | Tuning 7 Parameters 2000/2000 Iterations 32006/32000 Games Played | NMP params tune |
| Warreg | Pea | probcut-improving-margin | diff | 8.0+0.08 | LLR: -1.83 (-2.94, 2.94) [0.00, 5.00] Games: 31988 W: 7511 L: 7451 D: 17026 Ptnml(0-2): 395, 3826, 7479, 3912, 382 | added improving margin |
| Warreg | Pea | probcut-see | diff | 8.0+0.08 | Tuning 5 Parameters 854/2000 Iterations 13672/32000 Games Played | Spsa tune for some extra probcut parameters added. See threshold and probcut improving margin |
| Warreg | Pea | probcut-tuned | diff | 8.0+0.08 | LLR: -1.12 (-2.94, 2.94) [0.00, 5.00] Games: 3230 W: 730 L: 762 D: 1738 Ptnml(0-2): 40, 403, 768, 357, 47 | probcut tuned for 4k iterations |
| Warreg | Pea | probcut-initial | diff | 8.0+0.08 | LLR: -3.03 (-2.94, 2.94) [0.00, 5.00] Games: 2458 W: 516 L: 631 D: 1311 Ptnml(0-2): 41, 341, 567, 252, 28 | added params to probcut, ready for spsa |
| Warreg | Pea | probcut-initial | diff | 8.0+0.08 | Tuning 3 Parameters 4043/5000 Iterations 64688/80000 Games Played | Pc tune |
| Warreg | Pea | ob-spsa-compatibility | diff | 4.0+0.04 | LLR: 0.40 (-2.94, 2.94) [-5.00, 0.00] Games: 1788 W: 436 L: 427 D: 925 Ptnml(0-2): 24, 212, 413, 221, 24 | exposed options to change search parameters for spsa |
| Warreg | Pea | probcut-initial | diff | 8.0+0.08 | LLR: -1.16 (-2.94, 2.94) [0.00, 5.00] Games: 17340 W: 4083 L: 4057 D: 9200 Ptnml(0-2): 237, 2097, 3962, 2151, 223 | fiddle with params try 9 |
| Warreg | Pea | probcut-initial | diff | 8.0+0.08 | LLR: -2.24 (-2.94, 2.94) [0.00, 5.00] Games: 2434 W: 543 L: 628 D: 1263 Ptnml(0-2): 39, 352, 513, 281, 32 | fiddle with params try 8 |
| Warreg | Pea | probcut-initial | diff | 8.0+0.08 | LLR: -0.23 (-2.94, 2.94) [0.00, 5.00] Games: 330 W: 65 L: 73 D: 192 Ptnml(0-2): 7, 40, 74, 42, 2 | fiddle with params try 7 |