| Finished | ||||||
|---|---|---|---|---|---|---|
| Peregr | Reckless | pere-big-brain-weight-adj | diff | 8.0+0.08 | LLR: -1.66 (-2.25, 2.89) [0.00, 4.00] Games: 20494 W: 5207 L: 5221 D: 10066 Ptnml(0-2): 108, 2454, 5109, 2496, 80 | I don't agree with the naming but let's test stc as well |
| Styxdo | Reckless | pere-big-brain-weight-adj | diff | N=25000 | LLR: -2.27 (-2.25, 2.89) [0.00, 4.00] Games: 5170 W: 1674 L: 1791 D: 1705 Ptnml(0-2): 178, 608, 1127, 497, 175 | FN : Pere weight adj idea |
| World | Reckless | bnfp-depth | diff | 8.0+0.08 | LLR: -2.27 (-2.25, 2.89) [0.00, 4.00] Games: 33640 W: 8285 L: 8287 D: 17068 Ptnml(0-2): 103, 4009, 8576, 4051, 81 | replace lmr_depth with depth |
| World | Reckless | beta1_change | diff | 8.0+0.08 | LLR: 0.40 (-2.25, 2.89) [0.00, 4.00] Games: 25590 W: 6501 L: 6406 D: 12683 Ptnml(0-2): 85, 3081, 6371, 3170, 88 | beta1 value adjustment |
| World | Reckless | adv-s1 | diff | 8.0+0.08 | Elo: -45.77 +- 7.26 (95%) [N=40000] Games: 2512 W: 518 L: 847 D: 1147 Ptnml(0-2): 30, 456, 588, 177, 5 | S1 vs. S2 |
| World | Reckless | adv-s1 | diff | N=25000 | Elo: -27.74 +- 5.71 (95%) [N=40000] Games: 7254 W: 2178 L: 2756 D: 2320 Ptnml(0-2): 357, 892, 1542, 644, 192 | S1 vs. S2 |
| World | Reckless | adv-02-halved | diff | 8.0+0.08 | LLR: -2.28 (-2.25, 2.89) [0.00, 4.00] Games: 4108 W: 948 L: 1038 D: 2122 Ptnml(0-2): 18, 529, 1042, 455, 10 | instead of rescoring opponent move with our static eval, remove them |
| World | Reckless | adv-02-halved | diff | N=25000 | LLR: -2.35 (-2.25, 2.89) [0.00, 4.00] Games: 9882 W: 3183 L: 3288 D: 3411 Ptnml(0-2): 359, 1147, 2003, 1104, 328 | instead of rescoring opponent move with our static eval, remove them |
| Styxdo | Reckless | beta1_change | diff | N=25000 | LLR: 3.20 (-2.25, 2.89) [0.00, 4.00] Games: 23482 W: 7999 L: 7705 D: 7778 Ptnml(0-2): 789, 2548, 4864, 2660, 880 | beta1 value adjustment |
| Styxdo | Reckless | big-brain-filtering | diff | N=25000 | LLR: -2.56 (-2.25, 2.89) [0.00, 4.00] Games: 974 W: 260 L: 425 D: 289 Ptnml(0-2): 66, 140, 194, 67, 20 | Big brain Filtering |
| Peregr | Reckless | pc-ttpv | diff | 8.0+0.08 | LLR: -2.27 (-2.25, 2.89) [0.00, 4.00] Games: 27900 W: 6936 L: 6955 D: 14009 Ptnml(0-2): 86, 3418, 6950, 3421, 75 | prevent probcut at tt_pv |
| Peregr | Reckless | lmr-hs-maybe | diff | 8.0+0.08 | LLR: -1.18 (-2.25, 2.89) [0.00, 4.00] Games: 2712 W: 647 L: 692 D: 1373 Ptnml(0-2): 9, 359, 660, 324, 4 | if this works then we don't need the fractional thing anymore for hs |
| World | Reckless | alex | diff | N=25000 | Elo: 12.60 +- 4.68 (95%) [N=40000] Games: 11314 W: 4001 L: 3591 D: 3722 Ptnml(0-2): 401, 1126, 2285, 1352, 493 | Alex's 1536 |
| World | Reckless | alex | diff | 8.0+0.08 | Elo: -9.76 +- 9.23 (95%) [N=40000] Games: 1638 W: 406 L: 452 D: 780 Ptnml(0-2): 9, 233, 381, 187, 9 | Alex's 1536 |
| World | Reckless | step-wdl-s2 | diff | 40.0+0.40 | LLR: 0.36 (-2.25, 2.89) [0.00, 4.00] Games: 39602 W: 9476 L: 9349 D: 20777 Ptnml(0-2): 14, 4621, 10407, 4742, 17 | wdl scales, right... |
| Peregr | Reckless | capthist-pc | diff | 8.0+0.08 | LLR: -2.30 (-2.25, 2.89) [0.00, 4.00] Games: 15822 W: 3876 L: 3931 D: 8015 Ptnml(0-2): 43, 1905, 4060, 1870, 33 | Apply noisy hist updates in probcut |
| World | Reckless | s1_linear_wdl_fixed | diff | 8.0+0.08 | LLR: -2.29 (-2.25, 2.89) [0.00, 4.00] Games: 10354 W: 2611 L: 2684 D: 5059 Ptnml(0-2): 47, 1271, 2613, 1200, 46 | |
| World | Reckless | see-apply-history-once | diff | 8.0+0.08 | LLR: -2.27 (-2.25, 2.89) [0.00, 4.00] Games: 15096 W: 3622 L: 3678 D: 7796 Ptnml(0-2): 47, 1801, 3901, 1759, 40 | |
| Peregr | Reckless | margin-lmp | diff | 8.0+0.08 | LLR: -2.27 (-2.25, 2.89) [0.00, 4.00] Games: 17000 W: 4175 L: 4226 D: 8599 Ptnml(0-2): 48, 2089, 4279, 2034, 50 | |
| Styxdo | Reckless | s1_linear_wdl_fixed | diff | 40.0+0.40 | LLR: -0.36 (-2.25, 2.89) [0.00, 4.00] Games: 412 W: 78 L: 92 D: 242 Ptnml(0-2): 0, 55, 111, 39, 1 | fix should benefit ltc, so testing |
| Styxdo | Reckless | s1_linear_wdl_fixed | diff | N=25000 | LLR: 0.03 (-2.25, 2.89) [0.00, 4.00] Games: 22984 W: 7755 L: 7659 D: 7570 Ptnml(0-2): 809, 2489, 4827, 2531, 836 | Fix Mistake |
| Peregr | Reckless | lmr-check-dive | diff | 8.0+0.08 | LLR: -2.28 (-2.25, 2.89) [0.00, 4.00] Games: 41792 W: 10297 L: 10275 D: 21220 Ptnml(0-2): 123, 5104, 10405, 5156, 108 | prevent diving into qsearch from LMR for positions in check since we break for quiet evasions there |
| Styxdo | Reckless | s1_adv_fixed | diff | N=25000 | LLR: -1.06 (-2.25, 2.89) [0.00, 4.00] Games: 14796 W: 4870 L: 4875 D: 5051 Ptnml(0-2): 491, 1682, 3060, 1671, 494 | Mistake fixed |
| Styxdo | Reckless | adv_02_s1 | diff | 40.0+0.40 | LLR: 0.26 (-2.25, 2.89) [0.00, 4.00] Games: 27440 W: 6669 L: 6580 D: 14191 Ptnml(0-2): 22, 3206, 7164, 3317, 11 | LTC: Replace 1536 Data in S1 with Adv-01+Adv-02. |
| Styxdo | Reckless | s1_wdl_linear | diff | 8.0+0.08 | LLR: -1.22 (-2.25, 2.89) [0.00, 4.00] Games: 11470 W: 2899 L: 2920 D: 5651 Ptnml(0-2): 49, 1389, 2890, 1348, 59 | Linear WDL in S1, increasing a little bit since v27 over the Stage. |