Stockfish Testing Queue

Finished - 19466 tests

13-04-13 jk no_evasion_pruning diff
LLR: -2.94 (-2.94,2.94)
Total: 9206 W: 1769 L: 1850 D: 5587
sprt @ 15+0.05 th 1 Do not prune evasions
13-04-13 jk qsearch_fut_down diff
LLR: 3.00 (-2.94,2.94)
Total: 26681 W: 5277 L: 5097 D: 16307
sprt @ 15+0.05 th 1 Drop futility margin in qsearch to 96.
13-04-13 gl pinned_null diff
LLR: -2.94 (-2.94,2.94)
Total: 8111 W: 1578 L: 1662 D: 4871
sprt @ 15+0.05 th 1 SPRT retest
14-04-13 lb simplify_eval diff
ELO: -3.65 +-3.6 (95%) LOS: 2.2%
Total: 16000 W: 3412 L: 3580 D: 9008
16000 @ 5+0.05 th 1 reduce number of parameters to pawn structure eval. (candidate from slim_master, expected 0 ELO)
13-04-13 rt BadBishop_C1 diff
LLR: -2.92 (-2.94,2.94)
Total: 14723 W: 2889 L: 2954 D: 8880
sprt @ 15+0.05 th 1 Bad bishop tuning and rewriting
12-04-13 gl gl_kingmob diff
LLR: -2.89 (-2.94,2.94)
Total: 15864 W: 3106 L: 3167 D: 9591
sprt @ 15+0.05 th 1 Penalize king in endgame for being trapped (seen this pattern in two lost games recently!)
13-04-13 jk always_prune_equal_see diff
LLR: 3.01 (-2.94,2.94)
Total: 11985 W: 2369 L: 2229 D: 7387
sprt @ 15+0.05 th 1 Prune equal SEE moves also at depth zero
12-04-13 rt RT_Rook_A1 diff
LLR: -3.06 (-2.94,2.94)
Total: 7720 W: 1456 L: 1545 D: 4719
sprt @ 15+0.05 th 1 Lowers punishment for a rook trapped inside by a castlable king
12-04-13 rt RT_KingSaftey_B1 diff
LLR: -2.94 (-2.94,2.94)
Total: 4006 W: 781 L: 877 D: 2348
sprt @ 15+0.05 th 1 Smoothes out the king safety formula
12-04-13 rt RT_Outpost_B1 diff
LLR: -2.90 (-2.94,2.94)
Total: 6038 W: 1134 L: 1222 D: 3682
sprt @ 15+0.05 th 1 Pushes outposted piece when the enemy has a knight.
12-04-13 rt RT_PassedPawns_C1 diff
LLR: -2.98 (-2.94,2.94)
Total: 16613 W: 3251 L: 3313 D: 10049
sprt @ 15+0.05 th 1 Punish passed pawn supporters that are blocked by the enemy, possibly preventing progression
09-04-13 jk no_mate_distance_prunin diff
LLR: -2.92 (-2.94,2.94)
Total: 19965 W: 3869 L: 3920 D: 12176
sprt @ 15+0.05 th 1 Retire mate distance pruning
12-04-13 rt RT_PassedPawns_B1 diff
LLR: -2.93 (-2.94,2.94)
Total: 3525 W: 688 L: 785 D: 2052
sprt @ 15+0.05 th 1 Deduces bonus of passed pawns in the endgame with enemy bishop pair.
11-04-13 rt RT_DrawishPawns_A1 diff
LLR: -2.93 (-2.94,2.94)
Total: 2500 W: 428 L: 526 D: 1546
sprt @ 15+0.05 th 1 Symmetrical pawns tend to lead to draws
11-04-13 rt RT_Span_A1 diff
LLR: -2.97 (-2.94,2.94)
Total: 7060 W: 1366 L: 1454 D: 4240
sprt @ 15+0.05 th 1 Attempts to reward bishops and knights depending on the pawn span
11-04-13 jk neg_see_5 diff
LLR: -2.98 (-2.94,2.94)
Total: 17865 W: 3095 L: 3111 D: 11659
sprt @ 60+0.05 th 1 Stage II: Prune neg. SEE moves when depth < 5 plies
11-04-13 rt RT_Bishop_B1 diff
LLR: -1.35 (-2.94,2.94)
Total: 15391 W: 2679 L: 2688 D: 10024
sprt @ 60+0.05 th 1 Bad bishop module did well in fast TC.
11-04-13 ju Knight_outpost_test diff
LLR: -2.89 (-2.94,2.94)
Total: 16571 W: 3245 L: 3304 D: 10022
sprt @ 15+0.05 th 1 Looking to see if it is worth experimenting with outpost bonuses
11-04-13 rt RT_Rook_D1 diff
LLR: 3.01 (-2.94,2.94)
Total: 20450 W: 4091 L: 3927 D: 12432
sprt @ 15+0.05 th 1 Rook on pawn and 7th rank macro tuning.
11-04-13 rt RT_Rook_C1 diff
LLR: -2.90 (-2.94,2.94)
Total: 5592 W: 1030 L: 1119 D: 3443
sprt @ 15+0.05 th 1 Reward majors keeping king on the edge instead of just rank 7.
11-04-13 jk qsearch_always_check diff
LLR: -2.92 (-2.94,2.94)
Total: 7926 W: 1554 L: 1638 D: 4734
sprt @ 15+0.05 th 1 Always generate checks in qsearch, but also prune very easily
10-04-13 gl passed_pawns diff
LLR: -2.89 (-2.94,2.94)
Total: 67974 W: 13496 L: 13414 D: 41064
sprt @ 15+0.05 th 1 A variation on James passed pawn patch
10-04-13 gl rook_pin diff
LLR: -3.07 (-2.94,2.94)
Total: 9780 W: 1908 L: 1992 D: 5880
sprt @ 15+0.05 th 1 Tweak rook_pin weight down a bit
09-04-13 jk ortho_exclusion diff
LLR: -2.86 (-2.94,2.94)
Total: 7760 W: 1481 L: 1563 D: 4716
sprt @ 15+0.05 th 1 Use orthodox zobrist key in exclusion search
10-04-13 mc master diff
ELO: 24.13 +-2.8 (95%) LOS: 100.0%
Total: 20000 W: 4178 L: 2791 D: 13031
20000 @ 60+0.05 th 1 Regression test for candidate to TCEC Stage 4. Included always check for repetition and asymm see.
05-04-13 jk bad_evasion_prune_depth diff
LLR: -3.01 (-2.94,2.94)
Total: 57219 W: 10585 L: 10543 D: 36091
sprt @ 15+0.05 th 1 SPRT Retest: Prune bad evasions only at very low depths
10-04-13 gl double_gain diff
LLR: -2.99 (-2.94,2.94)
Total: 8838 W: 1502 L: 1563 D: 5773
sprt @ 60+0.05 th 1 Re-test using SPRT and new book
10-04-13 gl all_cut_iid diff
LLR: -2.95 (-2.94,2.94)
Total: 6622 W: 1232 L: 1320 D: 4070
sprt @ 15+0.05 th 1 Try all/cut IID with the new IID tweaks
10-04-13 gl singular_tweak5 diff
LLR: -2.99 (-2.94,2.94)
Total: 4278 W: 838 L: 935 D: 2505
sprt @ 15+0.05 th 1 More aggressive rBeta in non-PV nodes
10-04-13 gl king_safety2 diff
LLR: -2.99 (-2.94,2.94)
Total: 1833 W: 384 L: 492 D: 957
sprt @ 15+0.05 th 1 EG kingsafety = MG kingsafety / 4
10-04-13 rt RT_Scaling_B2 diff
LLR: -2.95 (-2.94,2.94)
Total: 2304 W: 408 L: 508 D: 1388
sprt @ 15+0.05 th 1 Fourth try at 50rule scaling.
11-04-13 rt RT_Bishop_B1 diff
LLR: 2.80 (-2.94,2.94)
Total: 10894 W: 2220 L: 2089 D: 6585
sprt @ 15+0.05 th 1 Tuned version of bad bishop module.
10-04-13 rt RT_Bishop_B1 diff
LLR: 2.13 (-2.94,2.94)
Total: 9492 W: 1933 L: 1830 D: 5729
sprt @ 15+0.05 th 1 Macro tuning of bad bishops.
09-04-13 gl bishop_pawns diff
LLR: -3.05 (-2.94,2.94)
Total: 15702 W: 3072 L: 3139 D: 9491
sprt @ 15+0.05 th 1 Adjust bishop value by pawns
09-04-13 jk probcut_tweak diff
LLR: -3.06 (-2.94,2.94)
Total: 34159 W: 6855 L: 6871 D: 20433
sprt @ 15+0.05 th 1 Probcut: use rbeta - staticEval as threshold
09-04-13 jk always_check_repetition diff
LLR: 3.05 (-2.94,2.94)
Total: 38986 W: 6845 L: 6547 D: 25594
sprt @ 60+0.05 th 1 Retest: always check repetition
09-04-13 jk neg_see_5 diff
ELO: 1.70 +-1.9 (95%) LOS: 96.1%
Total: 51647 W: 10335 L: 10083 D: 31229
sprt @ 15+0.05 th 1 Prune neg. SEE moves when depth < 5
09-04-13 jk nonpv_check_low diff
ELO: -0.87 +-3.0 (95%) LOS: 28.5%
Total: 20483 W: 4018 L: 4069 D: 12396
sprt @ 15+0.05 th 1 Extend checks a full ply in low depths.
09-04-13 jk null_pv2 diff
ELO: -0.68 +-2.8 (95%) LOS: 31.9%
Total: 23055 W: 4569 L: 4614 D: 13872
sprt @ 15+0.05 th 1 SPRT retry: null move in PV
09-04-13 rt RT_Scaling_B1 diff
ELO: -10.53 +-7.7 (95%) LOS: 0.4%
Total: 3168 W: 591 L: 687 D: 1890
sprt @ 15+0.05 th 1 Fifth attempt at refining scaling out drawish positions.
09-04-13 rt RT_KingExp_A1 diff
Pending...
sprt @ 15+0.05 th 1 Attempts to understand the flaw in averaging king safety by testing the "Mindreading Effect".
08-04-13 rt Ryan_Bishop_A1 diff
ELO: 0.24 +-5.6 (95%) LOS: 53.3%
Total: 5883 W: 1169 L: 1165 D: 3549
sprt @ 15+0.05 th 1 Ryan Takker's bad bishop module
05-04-13 jk static_null5b diff
ELO: -0.16 +-2.2 (95%) LOS: 44.3%
Total: 33712 W: 6167 L: 6183 D: 21362
sprt @ 15+0.05 th 1 SPRT retest: static null move pruning at depth = 5 * ONE_PLY. Previous run was inconclusive.
05-04-13 jk prune_under_promo diff
ELO: -0.44 +-2.5 (95%) LOS: 36.5%
Total: 27383 W: 5122 L: 5157 D: 17104
sprt @ 15+0.05 th 1 SPRT retest: prune_under_promo. Previously scored (0 +- 3), so result was "unclear".
05-04-13 jk asymm_see2 diff
ELO: 3.53 +-2.8 (95%) LOS: 99.3%
Total: 18794 W: 3098 L: 2907 D: 12789
sprt @ 60+0.05 th 1 SPRT: Final judgement for asymm_see2
07-04-13 mc master diff
ELO: 19.02 +-2.8 (95%) LOS: 100.0%
Total: 20000 W: 4042 L: 2948 D: 13010
20000 @ 60+0.05 th 1 Regression test after amazing JP patch on UCI parameters tweaks.
07-04-13 jk king_safety2 diff
ELO: -26.15 +-15.0 (95%) LOS: 0.0%
Total: 892 W: 159 L: 226 D: 507
sprt @ 15+0.05 th 1 EG kingsafety = MG kingsafety / 4
07-04-13 gl King_Saftey_A2 diff
ELO: -2.33 +-4.0 (95%) LOS: 13.0%
Total: 11199 W: 2183 L: 2258 D: 6758
sprt @ 15+0.05 th 1 RyanTakker: Tweak take 2
06-04-13 jk nonpv_checks diff
ELO: 3.07 +-3.3 (95%) LOS: 96.7%
Total: 17071 W: 3435 L: 3284 D: 10352
sprt @ 15+0.05 th 1 More complex extension system for non-pv check
07-04-13 lb lucas_qsearch_pruning diff
ELO: 6.94 +-5.4 (95%) LOS: 99.4%
Total: 6061 W: 1210 L: 1089 D: 3762
sprt @ 15+0.05 th 1 remove PV extensions. again with check extension at ONE_PLY/2, following Joona's recommendation.