Stockfish Testing Queue

Pending - 0 tests 0.0 hrs

None

Active - 1 tests

28-04-17 pb mate_in_qsearch diff
LLR: -0.83 (-2.94,2.94) [-3.00,1.00]
Total: 84908 W: 10770 L: 10920 D: 63218
sprt @ 60+0.6 th 1 LTC: Simplification for mate detection in qsearch (bench changes at higher depths)

Finished - 379 tests

28-04-17 pb mate_in_qsearch diff
LLR: 2.96 (-2.94,2.94) [-3.00,1.00]
Total: 19067 W: 3451 L: 3327 D: 12289
sprt @ 10+0.1 th 1 Simplification for mate detection in qsearch (bench changes at higher depths)
28-04-17 pb lmr_full diff
LLR: -2.97 (-2.94,2.94) [0.00,5.00]
Total: 15297 W: 2696 L: 2757 D: 9844
sprt @ 10+0.1 th 1 Don't research on LMR high-fail when searched just one ply behind
27-04-17 pb root_move_lmr diff
LLR: -2.98 (-2.94,2.94) [0.00,5.00]
Total: 50163 W: 6482 L: 6439 D: 37242
sprt @ 60+0.6 th 1 LTC As queue rather empty, testing a weird root-move-lmr idea.
27-04-17 pb root_move_lmr diff
LLR: 2.95 (-2.94,2.94) [0.00,5.00]
Total: 18516 W: 3448 L: 3244 D: 11824
sprt @ 10+0.1 th 1 As queue rather empty, testing a weird root-move-lmr idea.
24-04-17 pb smp_skip_move diff
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 19375 W: 3069 L: 3116 D: 13190
sprt @ 5+0.05 th 7 On helpers skip root move, if it has poor chances to raise alpha. Turns out when another thread with higher completed depth reports given move as best, but it's score is already overtrumped.
24-04-17 pb smp_skip_move diff
LLR: -2.97 (-2.94,2.94) [0.00,5.00]
Total: 32240 W: 5231 L: 5228 D: 21781
sprt @ 5+0.05 th 5 Take 2
21-04-17 pb smp_ttbest_killer diff
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 15350 W: 2477 L: 2539 D: 10334
sprt @ 5+0.05 th 5 SMP: use root-ttMove as killer, if it differs from thread's best-move (bugfix)
21-04-17 pb smp_ttbest_second diff
LLR: -2.97 (-2.94,2.94) [0.00,5.00]
Total: 7924 W: 1229 L: 1321 D: 5374
sprt @ 5+0.05 th 5 Take 2: use root-ttMove as second move, if it differs from thread's best-move. Regardless if it is quiet or not. (Other moves get deferred by one until gap get closed)
21-04-17 pb smp_ttbest_killer diff
LLR: 0.02 (-2.94,2.94) [0.00,5.00]
Total: 29 W: 5 L: 4 D: 20
sprt @ 5+0.05 th 7 SMP: use root-ttMove as killer, if it differs from thread's best-move
19-04-17 pb escape_reduct diff
LLR: -2.94 (-2.94,2.94) [0.00,5.00]
Total: 14449 W: 2618 L: 2681 D: 9150
sprt @ 10+0.1 th 1 LMR: Once a capture-escape-move is detected, reduce all remaining (not escaping) moves by one ply, since we can assume them weak.
18-04-17 pb rootMoveOrder diff
LLR: -2.96 (-2.94,2.94) [0.00,5.00]
Total: 6853 W: 1014 L: 1110 D: 4729
sprt @ 5+0.05 th 7 Synchronized rootMove-search among main- and helper-threads. Does it increase the communication over TT?
18-04-17 pb rootMoveOrder diff
LLR: -2.96 (-2.94,2.94) [0.00,5.00]
Total: 17934 W: 3240 L: 3289 D: 11405
sprt @ 10+0.1 th 1 Take 2: At high depths, process rootMoves in order of rootMoves vector
18-04-17 pb rootMoveOrder diff
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 8281 W: 1470 L: 1560 D: 5251
sprt @ 10+0.1 th 1 Process rootMoves in order of rootMoves vector. This way former best-moves still will be returned among first moves.
12-04-17 pb verification_search8 diff
LLR: -2.97 (-2.94,2.94) [0.00,5.00]
Total: 24596 W: 4430 L: 4451 D: 15715
sprt @ 10+0.1 th 1 Zugzwang: what if we remove all precautions and let verification search detect it?
11-04-17 pb formula2 diff
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 20692 W: 3723 L: 3760 D: 13209
sprt @ 10+0.1 th 1 Increase stats values near the leaves a bit
11-04-17 pb skipQuiets_smp diff
LLR: -2.97 (-2.94,2.94) [0.00,5.00]
Total: 2568 W: 362 L: 475 D: 1731
sprt @ 5+0.05 th 7 Try more aggressive skipQuiets for helper threads
07-04-17 pb formula2 diff
LLR: -2.96 (-2.94,2.94) [0.00,5.00]
Total: 29064 W: 4309 L: 4324 D: 20431
sprt @ 30+0.3 th 1 Take 2: Trying the opposite, lowering the linear component (and compensate in the exponent to reach the same ceiling). Since this is unlikely to work with very short tc, I use STC x 3 (standard throughput / 3 )
06-04-17 pb checkExtension1 diff
LLR: -2.96 (-2.94,2.94) [0.00,5.00]
Total: 72483 W: 13220 L: 13036 D: 46227
sprt @ 10+0.1 th 1 Take 2: more moveCountPruning-based check extensions
07-04-17 pb formula2 diff
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 8869 W: 1092 L: 1181 D: 6596
sprt @ 60+0.6 th 1 LTC: Another formula tweak attempt: higher linear coefficient and higher denominators to possibly avoid to fast saturation (=clumping to maxvalue)
07-04-17 pb formula2 diff
LLR: 2.95 (-2.94,2.94) [0.00,5.00]
Total: 25557 W: 4698 L: 4465 D: 16394
sprt @ 10+0.1 th 1 Another formula tweak attempt: higher linear coefficient and higher denominators to possibly avoid to fast saturation (=clumping to maxvalue)
07-04-17 pb checkExtension1 diff
LLR: -2.94 (-2.94,2.94) [0.00,5.00]
Total: 15012 W: 2670 L: 2731 D: 9611
sprt @ 10+0.1 th 1 Take 4
06-04-17 pb checkExtension1 diff
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 18981 W: 3432 L: 3476 D: 12073
sprt @ 10+0.1 th 1 Take 3: allow even more moveCountPruning-based check extensions
06-04-17 pb no_escape_detection diff
LLR: -2.95 (-2.94,2.94) [-3.00,1.00]
Total: 21113 W: 3630 L: 3821 D: 13662
sprt @ 10+0.1 th 1 Does LMR capture escape detection justify/outweigh the induced complication in see_ge?
06-04-17 pb checkExtension1 diff
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 12900 W: 2300 L: 2370 D: 8230
sprt @ 10+0.1 th 1 Less moveCountPruning-based check extensions
04-04-17 pb pv_draw_by_rule diff
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 14640 W: 2586 L: 2649 D: 9405
sprt @ 10+0.1 th 1 Take 3: Increment bestMoveChanges when PV leads into a draw by rule (with fix suggested by sg)
31-03-17 pb pv_draw_by_rule diff
LLR: -2.96 (-2.94,2.94) [0.00,5.00]
Total: 12627 W: 2206 L: 2278 D: 8143
sprt @ 10+0.1 th 1 Take 2: iterate deeper when standard logic would stop but PV leads into a draw by rule an (do it max. 8 times per game) bugfix
31-03-17 pb pv_draw_by_rule diff
LLR: -0.06 (-2.94,2.94) [0.00,5.00]
Total: 73 W: 11 L: 13 D: 49
sprt @ 10+0.1 th 1 Take 2: iterate deeper when standard logic would stop but PV leads into a draw by rule an (do it max. 8 times per game)
29-03-17 pb no_iid_for_allnodes diff
LLR: -2.92 (-2.94,2.94) [0.00,5.00]
Total: 38556 W: 6892 L: 6853 D: 24811
sprt @ 10+0.1 th 1 Does internal iterative deepening make sense on predicted all-nodes?
29-03-17 pb pv_draw_by_rule diff
LLR: -2.96 (-2.94,2.94) [0.00,5.00]
Total: 26156 W: 4709 L: 4723 D: 16724
sprt @ 10+0.1 th 1 Based on a suggestion by Brian: lengthen search time when PV leads into a draw (3fold or 50-moves) by incrementing bestMoveChanges
28-03-17 pb experiment1 diff
LLR: -2.96 (-2.94,2.94) [0.00,5.00]
Total: 30892 W: 5571 L: 5565 D: 19756
sprt @ 10+0.1 th 1 Hash pressure experiment: at predicted cutNodes use ttMove also in case of non TT-hit, if it is legal.
28-03-17 pb checkExt3 diff
LLR: -2.96 (-2.94,2.94) [0.00,5.00]
Total: 38576 W: 6854 L: 6817 D: 24905
sprt @ 10+0.1 th 1 Take 6: with mc>=5: since queue is rather empty, give a last shot at this to understand better if Take 4 (mc >=4) was merely an unlucky run.
28-03-17 pb separateTT diff
LLR: -2.96 (-2.94,2.94) [0.00,5.00]
Total: 5993 W: 1039 L: 1139 D: 3815
sprt @ 10+0.1 th 1 Take 3 and last attempt: Use simple formula (excludedMove ^2) for calculating key switch
28-03-17 pb separateTT diff
LLR: -2.96 (-2.94,2.94) [0.00,5.00]
Total: 6673 W: 1128 L: 1225 D: 4320
sprt @ 10+0.1 th 1 Take 2: use 0,5% size of main TT and include excluded-move into key-generation.
28-03-17 pb separateTT diff
LLR: -2.97 (-2.94,2.94) [0.00,5.00]
Total: 6546 W: 1146 L: 1244 D: 4156
sprt @ 10+0.1 th 1 Trying a dedicated T-Table (1% size of main TT) for excluded move positions. Test against VoyagerOne's passed excludedTT patch. Bugfix
28-03-17 pb separateTT diff
LLR: -0.45 (-2.94,2.94) [0.00,5.00]
Total: 995 W: 169 L: 184 D: 642
sprt @ 10+0.1 th 1 Trying a dedicated T-Table (1% size of main TT) for excluded move positions. Test against VoyagerOne's passed excludedTT patch.
27-03-17 pb checkExt2 diff
LLR: -2.97 (-2.94,2.94) [0.00,5.00]
Total: 16144 W: 2058 L: 2124 D: 11962
sprt @ 60+0.6 th 1 Take 2: Retire non-singular check extension on late evasion moves (movecount >=3)
27-03-17 pb checkExt2 diff
LLR: 2.95 (-2.94,2.94) [0.00,5.00]
Total: 93199 W: 16999 L: 16479 D: 59721
sprt @ 10+0.1 th 1 Take 2: Retire non-singular check extension on late evasion moves (movecount >=3) (fixed base-bench)
27-03-17 pb checkExt3 diff
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 13017 W: 2280 L: 2350 D: 8387
sprt @ 10+0.1 th 1 Take 5: Retire non-singular check extension on late evasion moves (movecount >=2)
27-03-17 pb checkExt3 diff
LLR: -2.96 (-2.94,2.94) [0.00,5.00]
Total: 22061 W: 3941 L: 3973 D: 14147
sprt @ 10+0.1 th 1 Take 4: Retire non-singular check extension on late evasion moves (movecount >=4)
27-03-17 pb extension2 diff
LLR: -2.97 (-2.94,2.94) [0.00,5.00]
Total: 26552 W: 4847 L: 4859 D: 16846
sprt @ 10+0.1 th 1 Take 3: Retire all check extensions on late evasion moves (movecount >=3)
24-03-17 pb checkExt diff
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 36845 W: 6553 L: 6523 D: 23769
sprt @ 10+0.1 th 1 Retire check extension on late evasion move when checking move was late too (bugfix). Test on top of passed extT.
24-03-17 pb checkExt diff
LLR: -0.19 (-2.94,2.94) [0.00,5.00]
Total: 513 W: 97 L: 103 D: 313
sprt @ 10+0.1 th 1 Retire check extension on late evasion move when checking move was late too. Test on top of passed extT
23-03-17 pb singular_checking diff
LLR: -2.96 (-2.94,2.94) [0.00,5.00]
Total: 44595 W: 7999 L: 7936 D: 28660
sprt @ 10+0.1 th 1 Singular Extension search: search one ply more when checking. Test on top of passed extT
22-03-17 pb killerCapture diff
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 37244 W: 6635 L: 6603 D: 24006
sprt @ 10+0.1 th 1 Sophisticated capturekiller logic showing good local results.
20-03-17 pb advanced_pawn_push diff
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 27201 W: 4832 L: 4842 D: 17527
sprt @ 10+0.1 th 1 Refine definition of advanced pawn push
18-03-17 pb verification_search8 diff
LLR: -2.96 (-2.94,2.94) [0.00,5.00]
Total: 18657 W: 1946 L: 2010 D: 14701
sprt @ 30+0.3 th 1 Does the zugzang-awareness allow us to null-move-prune also in pv-nodes? STC with 3 x times larger TC (1/3 throughput) to hit modified verification search often enough.
16-03-17 pb verification_search8 diff
LLR: -2.97 (-2.94,2.94) [0.00,5.00]
Total: 33809 W: 3574 L: 3597 D: 26638
sprt @ 30+0.3 th 1 Trying yet another variant: STC with 3 times larger TC (1/3 throughput) to trigger modified verification search enough often to measure any difference.
15-03-17 pb verification_search4'' diff
LLR: -2.96 (-2.94,2.94) [0.00,5.00]
Total: 45367 W: 4088 L: 4094 D: 37185
sprt @ 60+0.6 th 1 LTC: My try on make SF more aware of zugzwang, take 2: now verification triggered again not before depth 12. (8 moves book)
15-03-17 pb verification_search4'' diff
ELO: 2.35 +-2.1 (95%) LOS: 98.8%
Total: 19660 W: 1824 L: 1691 D: 16145
25000 @ 60+0.6 th 1 My try on make SF more aware of zugzwang, take 2: now with triggered again not before depth 12. (lower throughput and 8 moves book)
14-03-17 pb verification_search4' diff
ELO: 1.51 +-2.1 (95%) LOS: 92.3%
Total: 19561 W: 1825 L: 1740 D: 15996
20000 @ 60+0.6 th 1 My try on make SF more aware of zugzwang based on verification search. On pos FEN: 1k3b1q/pP2p1p1/P1K1P1Pp/7P/2B5/8/8/8 w - - 0 1 it finds Bb5 in 11 secs. (lower throughput.)