Stockfish Testing Queue

Finished - 24165 tests

24-05-13 mc master diff
Pending...
64000 @ 5+0.05 th 1 CLOP debug
24-05-13 gl probcut diff
LLR: -2.96 (-2.94,2.94)
Total: 2867 W: 545 L: 645 D: 1677
sprt @ 15+0.05 th 1 Tuned probcut based on analysis of scores from different depths
23-05-13 ur psq_divide diff
LLR: -2.97 (-2.94,2.94)
Total: 24116 W: 4996 L: 5036 D: 14084
sprt @ 15+0.05 th 1 change the middle game value of knight and bishop piece square table to 75% of their value(I think that with better evaluation it is better to have smaller piece square table values).
23-05-13 rt RandomSimple_A1 diff
LLR: -2.97 (-2.94,2.94)
Total: 11511 W: 2428 L: 2504 D: 6579
sprt @ 15+0.05 th 1 Simplifies King Safety Pawn shelter
20-05-13 jk TimeMan_MoveHorizon64 diff
LLR: -2.95 (-2.94,2.94)
Total: 15475 W: 2389 L: 2420 D: 10666
sprt @ 150+1.5 th 1 TimeMan: Movehorizon 64 (Longer time control with a typical increment
23-05-13 gl multiple_hanging diff
ELO: 46.60 +-35.5 (95%) LOS: 99.6%
Total: 135 W: 34 L: 16 D: 85
64000 @ 15+0.05 th 1 CLOP test with multiple hanging pieces
22-05-13 rt GeneralEval_A1 diff
LLR: -2.96 (-2.94,2.94)
Total: 68097 W: 14039 L: 13955 D: 40103
sprt @ 15+0.05 th 1 Lowers PSQT influence on queen side minor pieces.
21-05-13 jk space_eval diff
LLR: -2.96 (-2.94,2.94)
Total: 24450 W: 5018 L: 5057 D: 14375
sprt @ 15+0.05 th 1 Rescheduled: Changed Space evaluation
23-05-13 ra Bishop_Pinning_B1 diff
LLR: -2.96 (-2.94,2.94)
Total: 18252 W: 3822 L: 3878 D: 10552
sprt @ 15+0.05 th 1 Don't give bonus for bishop pinning enemy pawn
21-05-13 mc master^ diff
LLR: -2.97 (-2.94,2.94)
Total: 35023 W: 6408 L: 6333 D: 22282
sprt @ 60+0.05 th 1 Long TC: Delayed killers[] init: reverse regression test (we expect to fail)
23-05-13 tv pawn_ending_null_moves diff
LLR: -2.97 (-2.94,2.94)
Total: 11131 W: 2263 L: 2340 D: 6528
sprt @ 15+0.05 th 1 Try verified null moves in pawn endings
22-05-13 rt GeneralEval_C1 diff
LLR: -2.95 (-2.94,2.94)
Total: 17187 W: 3550 L: 3609 D: 10028
sprt @ 15+0.05 th 1 Reduces rook-file-based PSQT entries
22-05-13 ra Rook_Pinning diff
LLR: -2.96 (-2.94,2.94)
Total: 42652 W: 8850 L: 8837 D: 24965
sprt @ 15+0.05 th 1 Bonus for rook pinning piece
22-05-13 ur union_changes diff
LLR: 2.95 (-2.94,2.94)
Total: 64133 W: 11968 L: 11532 D: 40633
sprt @ 60+0.05 th 1
21-05-13 ur union_changes diff
LLR: 2.96 (-2.94,2.94)
Total: 4429 W: 964 L: 844 D: 2621
sprt @ 15+0.05 th 1 I test union of 3 changes that did not padd but scored more than 50% without adding code(only changing parameters or deleting a line)
21-05-13 dk space_pawn_advande diff
LLR: -2.96 (-2.94,2.94)
Total: 22392 W: 4745 L: 4789 D: 12858
sprt @ 15+0.05 th 1 space eval, pawn advance bonus
21-05-13 jk noRedRef diff
LLR: -2.96 (-2.94,2.94)
Total: 11381 W: 2072 L: 2118 D: 7191
sprt @ 60+0.05 th 1 Long TC: No LMR for countermoves (Try this one more time)
21-05-13 gl multiple_hanging diff
LLR: -2.97 (-2.94,2.94)
Total: 17318 W: 3145 L: 3161 D: 11012
sprt @ 60+0.05 th 1 Long TC: Penalize multiple hanging pieces in eval
21-05-13 gl multiple_hanging diff
LLR: 2.96 (-2.94,2.94)
Total: 11507 W: 2466 L: 2325 D: 6716
sprt @ 15+0.05 th 1 Penalize multiple hanging pieces in eval
21-05-13 gl stand_pat_penalty diff
LLR: -2.96 (-2.94,2.94)
Total: 28720 W: 6026 L: 6052 D: 16642
sprt @ 15+0.05 th 1 Idea from Lucas, penalty for standing pat with multiple hanging pieces
20-05-13 jk noRedRef diff
LLR: 2.95 (-2.94,2.94)
Total: 30450 W: 6464 L: 6269 D: 17717
sprt @ 15+0.05 th 1 No LMR for countermoves (Try this one more time)
20-05-13 dk space_eval diff
LLR: -0.13 (-2.94,2.94)
Total: 249 W: 45 L: 49 D: 155
sprt @ 15+0.05 th 1 Changed Space evaluation
17-05-13 gl hanging diff
LLR: 2.95 (-2.94,2.94)
Total: 81438 W: 16975 L: 16637 D: 47826
sprt @ 15+0.05 th 1 Better definition of weak pieces, only try one good saving move
19-05-13 mc killer_lmr diff
LLR: -2.95 (-2.94,2.94)
Total: 4423 W: 863 L: 958 D: 2602
sprt @ 15+0.05 th 1 Reduce killer moves (but less). Same idea that proved good with countermoves.
19-05-13 mc master^ diff
LLR: -2.95 (-2.94,2.94)
Total: 6670 W: 1309 L: 1379 D: 3982
sprt @ 15+0.05 th 1 Delayed killers[] init: reverse regression test (we expect to fail)
19-05-13 jk TimeMan_MoveHorizon64 diff
LLR: 2.96 (-2.94,2.94)
Total: 22361 W: 4123 L: 3905 D: 14333
sprt @ 60+0.05 th 1 TimeMan: Movehorizon 64 (Even if this passes, do not commit yet, we need more tests)
19-05-13 jk master diff
ELO: 132.40 +-98.4 (95%) LOS: 99.9%
Total: 22 W: 9 L: 1 D: 12
64000 @ 5+0.05 th 1 CLOP sanity check. Just a test
19-05-13 jk NoOppColorBishopScaling diff
LLR: -2.96 (-2.94,2.94)
Total: 23124 W: 4667 L: 4710 D: 13747
sprt @ 15+0.05 th 1 No scaling for opposite coloured bishops
19-05-13 jk TimeMan_MoveHorizon64 diff
LLR: 2.96 (-2.94,2.94)
Total: 18322 W: 3873 L: 3713 D: 10736
sprt @ 15+0.05 th 1 TimeMan: Movehorizon 64
18-05-13 mc countermove_lmr diff
LLR: 2.95 (-2.94,2.94)
Total: 22672 W: 4232 L: 4011 D: 14429
sprt @ 60+0.05 th 1 Reschedule at long TC: Take 2 at LMR of countermoves, this time reduce less instead of not reducing at all.
18-05-13 jk DoubleCounterMove diff
LLR: -2.96 (-2.94,2.94)
Total: 20576 W: 3678 L: 3678 D: 13220
sprt @ 60+0.05 th 1 LONG TC: Countermoves^2
18-05-13 jk DoubleCounterMove diff
LLR: 2.95 (-2.94,2.94)
Total: 7255 W: 1524 L: 1397 D: 4334
sprt @ 15+0.05 th 1 Countermoves^2
17-05-13 do fixed_contempt diff
LLR: -2.97 (-2.94,2.94)
Total: 15545 W: 3388 L: 3452 D: 8705
sprt @ 15+0.05 th 1 Contempt 16 cp without scaling. In my tests on 5+0.05 seems to be stronger than default.
16-05-13 mc master diff
ELO: 22.11 +-3.0 (95%) LOS: 100.0%
Total: 20000 W: 4483 L: 3212 D: 12305
20000 @ 60+0.05 th 1 Init regression test vs new sf_3 base
18-05-13 mc countermove_lmr diff
LLR: 2.95 (-2.94,2.94)
Total: 17724 W: 3756 L: 3598 D: 10370
sprt @ 15+0.05 th 1 Take 2 at LMR of countermoves, this time reduce less instead of not reducing at all.
17-05-13 gl countermove_lmr diff
LLR: 0.15 (-2.94,2.94)
Total: 26920 W: 4979 L: 4835 D: 17106
sprt @ 60+0.05 th 1 Long TC: Don't reduce countermoves as we do for killers
17-05-13 mc no_countermove_after_nu diff
LLR: -1.16 (-2.94,2.94)
Total: 48000 W: 9919 L: 9826 D: 28255
sprt @ 15+0.05 th 1 Don't use countermoves if previous move is null
17-05-13 rt TimeMan_D1 diff
LLR: -2.97 (-2.94,2.94)
Total: 2291 W: 426 L: 528 D: 1337
sprt @ 15+0.05 th 1 Replaces moveimportance array with higher steal ratio.
17-05-13 mc countermove_lmr diff
LLR: 2.97 (-2.94,2.94)
Total: 4330 W: 966 L: 845 D: 2519
sprt @ 15+0.05 th 1 Don't reduce countermoves as we do for killers
16-05-13 gl hanging diff
LLR: 0.35 (-2.94,2.94)
Total: 22634 W: 4105 L: 3978 D: 14551
sprt @ 60+0.05 th 1 Long TC: only try one move of weak piece first
16-05-13 gl hanging diff
LLR: -2.94 (-2.94,2.94)
Total: 3343 W: 643 L: 741 D: 1959
sprt @ 15+0.05 th 1 Penalize not saving weak pieces instead of bonuses
16-05-13 gl hanging diff
LLR: -2.96 (-2.94,2.94)
Total: 23927 W: 4972 L: 5012 D: 13943
sprt @ 15+0.05 th 1 Better definition of weak pieces
16-05-13 jk FromStats diff
LLR: -2.97 (-2.94,2.94)
Total: 12509 W: 2537 L: 2610 D: 7362
sprt @ 15+0.05 th 1 Pt-From-To stats
15-05-13 rt LowCap_A1 diff
LLR: -2.96 (-2.94,2.94)
Total: 59671 W: 12503 L: 12441 D: 34727
sprt @ 15+0.05 th 1 Long shot test. Lowers the cap on attack-units.
16-05-13 gl hanging diff
LLR: 0.12 (-2.94,2.94)
Total: 114 W: 24 L: 19 D: 71
sprt @ 60+0.05 th 1 Long TC: Try only one move for saving weak piece
16-05-13 gl hanging diff
LLR: 2.97 (-2.94,2.94)
Total: 11440 W: 2405 L: 2265 D: 6770
sprt @ 15+0.05 th 1 Try only one move for saving weak piece
16-05-13 rt TimeMan_A1 diff
ELO: -17.00 +-3.1 (95%) LOS: 0.0%
Total: 20000 W: 3680 L: 4658 D: 11662
20000 @ 15+0.05 th 1 Worked fairly well in small local tests. Removes helpful, but inherently flawed Time Management strategy
16-05-13 rt AsemEval_B2 diff
LLR: -2.95 (-2.94,2.94)
Total: 4802 W: 943 L: 1037 D: 2822
sprt @ 15+0.05 th 1 Much more reasonable asem space eval.
15-05-13 jk Counter2 diff
LLR: 2.95 (-2.94,2.94)
Total: 5954 W: 1151 L: 1016 D: 3787
sprt @ 60+0.05 th 1 LONG TC: Two countermoves
15-05-13 rt AsemEval_B1 diff
LLR: -2.97 (-2.94,2.94)
Total: 7144 W: 1429 L: 1517 D: 4198
sprt @ 15+0.05 th 1 Asem eval for space. *should* work :P