Stockfish Testing Queue

Finished - 40562 tests

16-10-19 Voy sc4 diff
LLR: 2.95 (-2.94,2.94) [0.00,5.00]
Total: 216596 W: 28157 L: 27343 D: 161096
sprt @ 60+0.6 th 1 LTC: Take 4
16-10-22 aji tune_fmc diff
24043/25000 iterations
48894/50000 games played
50000 @ 20+0.2 th 1 tune fmc low depth vaues separately as these are elo sensitive
16-10-23 sg kingtropism diff
LLR: -2.96 (-2.94,2.94) [0.00,5.00]
Total: 36708 W: 6520 L: 6491 D: 23697
sprt @ 10+0.1 th 1 add bonus only if no queen already is in the king tropism zone
16-10-23 vdv pick diff
LLR: 2.95 (-2.94,2.94) [-3.00,1.00]
Total: 47940 W: 6205 L: 6120 D: 35615
sprt @ 60+0.6 th 1 LTC: simplify next_move, local elo gain.
16-10-23 sg kingtropism diff
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 38639 W: 6951 L: 6912 D: 24776
sprt @ 10+0.1 th 1 Additional bonus for save queen moves. Disallow squares occupied by own pieces. Take 2
16-10-23 Dis rook_seventh_pawns_2 diff
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 21211 W: 3782 L: 3817 D: 13612
sprt @ 10+0.1 th 1 Last try, now adjusting score with number of pawns
16-10-22 Voy combo diff
LLR: 2.95 (-2.94,2.94) [0.00,5.00]
Total: 45121 W: 8156 L: 7842 D: 29123
sprt @ 10+0.1 th 1 stc
16-10-23 sg kingtropism diff
LLR: -2.97 (-2.94,2.94) [0.00,5.00]
Total: 7914 W: 1332 L: 1424 D: 5158
sprt @ 10+0.1 th 1 Additional bonus for save queen moves
16-10-23 Dis rook_seventh_pawns_2 diff
LLR: -2.96 (-2.94,2.94) [0.00,5.00]
Total: 16681 W: 2941 L: 2996 D: 10744
sprt @ 10+0.1 th 1 Second try, with tweaks to the score and small improvement thanks to lucabrivio
16-10-23 Fis mgeginit diff
LLR: -2.96 (-2.94,2.94) [0.00,5.00]
Total: 14728 W: 2608 L: 2671 D: 9449
sprt @ 10+0.1 th 1 Use both mg and eg in initiative.
16-10-22 vdv pick diff
LLR: 2.95 (-2.94,2.94) [-3.00,1.00]
Total: 29539 W: 5291 L: 5185 D: 19063
sprt @ 10+0.1 th 1 simplify next_move, local elo gain.
16-10-22 fau idea diff
LLR: -2.96 (-2.94,2.94) [0.00,5.00]
Total: 31440 W: 5590 L: 5583 D: 20267
sprt @ 10+0.1 th 1 My try on the Bonus for connected rooks in case of open files.
16-10-22 Fis rankSpan diff
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 15040 W: 2683 L: 2744 D: 9613
sprt @ 10+0.1 th 1 Include the Rank span of pawns in initiative.
16-10-22 cib cutNode2 diff
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 10081 W: 1246 L: 1331 D: 7504
sprt @ 60+0.6 th 1 LTC, adjust cutNode reduction by move count.
16-10-22 cib cutNode2 diff
LLR: 2.96 (-2.94,2.94) [0.00,5.00]
Total: 93220 W: 16828 L: 16311 D: 60081
sprt @ 10+0.1 th 1 STC, adjust cutNode reduction by move count.
16-10-22 sni heavy diff
LLR: -2.97 (-2.94,2.94) [0.00,5.00]
Total: 19322 W: 3418 L: 3462 D: 12442
sprt @ 10+0.1 th 1 Bonus for heavy pieces batteries on open file
16-10-22 aji improving2 diff
LLR: -2.96 (-2.94,2.94) [0.00,5.00]
Total: 3162 W: 517 L: 629 D: 2016
sprt @ 10+0.1 th 1 fixed bench: STC
16-10-22 sni heavy diff
LLR: -2.96 (-2.94,2.94) [0.00,5.00]
Total: 6881 W: 1188 L: 1284 D: 4409
sprt @ 10+0.1 th 1 Bonus for connected rooks in case of open files, as suggested by lilmafya in the forum
16-10-22 sg rooks diff
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 8475 W: 1448 L: 1537 D: 5490
sprt @ 10+0.1 th 1 Rook penalty in endgame if no open files exists
16-10-21 sni pawnless_flank diff
LLR: 2.95 (-2.94,2.94) [0.00,5.00]
Total: 71808 W: 13038 L: 12610 D: 46160
sprt @ 10+0.1 th 1 A variant of the pawnless flank idea. Malus = S(20,80)
16-10-21 fau PLF diff
LLR: -2.96 (-2.94,2.94) [0.00,5.00]
Total: 12331 W: 2185 L: 2258 D: 7888
sprt @ 10+0.1 th 1 PLF with SPSA tuned values.
16-10-21 fau tune diff
9881/10000 iterations
20000/20000 games played
20000 @ 10+0.1 th 1 Fast tune of pawnless flank idea.
16-10-21 aji improving diff
LLR: -2.96 (-2.94,2.94) [0.00,5.00]
Total: 38333 W: 6891 L: 6854 D: 24588
sprt @ 10+0.1 th 1 Improving tweak: STC
16-10-21 pb0 rHist2 diff
LLR: -2.96 (-2.94,2.94) [0.00,5.00]
Total: 4309 W: 718 L: 825 D: 2766
sprt @ 10+0.1 th 1 Continuously tracking of history-mean-value to better calculate rHist, fixed logic
16-10-21 Voy nullStat diff
LLR: -2.97 (-2.94,2.94) [0.00,5.00]
Total: 17861 W: 3078 L: 3129 D: 11654
sprt @ 10+0.1 th 1 stc
16-10-21 Voy nullStat2 diff
LLR: -1.28 (-2.94,2.94) [0.00,5.00]
Total: 2385 W: 402 L: 446 D: 1537
sprt @ 10+0.1 th 1 Consider cmh null as well.
16-10-20 atu probcut diff
LLR: -2.94 (-2.94,2.94) [-3.00,1.00]
Total: 115545 W: 20257 L: 20606 D: 74682
sprt @ 10+0.1 th 1 Simplify probcut: don't add 1 to threshold
16-10-21 sni pawn_chains6' diff
LLR: -2.94 (-2.94,2.94) [0.00,5.00]
Total: 23682 W: 4265 L: 4289 D: 15128
sprt @ 10+0.1 th 1 Take 2 (only midgame correction for psqt)
16-10-21 sni pawn_chains6' diff
LLR: -2.96 (-2.94,2.94) [0.00,5.00]
Total: 21190 W: 3730 L: 3766 D: 13694
sprt @ 10+0.1 th 1 Pawn chains in the center (with psqt compensation)
16-10-21 Elb knight_end diff
LLR: -2.97 (-2.94,2.94) [0.00,5.00]
Total: 20423 W: 3649 L: 3688 D: 13086
sprt @ 10+0.1 th 1 Endings with a lot of pawns and B vs N are drawish (see https://chessprogramming.wikispaces.com/Bishop+versus+Knight)
16-10-21 pb0 rHist2 diff
LLR: -0.79 (-2.94,2.94) [0.00,5.00]
Total: 6750 W: 1197 L: 1202 D: 4351
sprt @ 10+0.1 th 1 Continuously tracking of history-mean-value to better calculate rHist
16-10-20 Dis drawish_pawns diff
LLR: -2.96 (-2.94,2.94) [0.00,5.00]
Total: 2983 W: 484 L: 597 D: 1902
sprt @ 10+0.1 th 1 Reduce the score of pawns when they are too close to the opposing pawns.
16-10-20 sni futility2 diff
LLR: -2.95 (-2.94,2.94) [0.00,4.00]
Total: 52471 W: 9441 L: 9419 D: 33611
sprt @ 10+0.1 th 1 Use futility depth = 5 and parent futility depth = 5
16-10-20 Fis TTeval diff
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 28334 W: 5067 L: 5072 D: 18195
sprt @ 10+0.1 th 1 Use value16 in TT replacement policy. 1MB high pressure.
16-10-20 vdv undefined diff
LLR: -2.96 (-2.94,2.94) [-3.00,1.00]
Total: 16480 W: 2790 L: 2973 D: 10717
sprt @ 10+0.1 th 1 Fixes undefined behaviour.
16-10-20 fau KDFauzi2 diff
LLR: -2.95 (-2.94,2.94) [0.00,4.00]
Total: 37642 W: 6768 L: 6796 D: 24078
sprt @ 10+0.1 th 1 Take 2
16-10-20 fau elberto diff
LLR: -2.94 (-2.94,2.94) [0.00,5.00]
Total: 39688 W: 7039 L: 6997 D: 25652
sprt @ 10+0.1 th 1 Expand also to middlegame Elberto's idea of giving penalty to a king in a pawnless flank
16-10-20 fau KDFauzi diff
LLR: -2.96 (-2.94,2.94) [0.00,4.00]
Total: 20277 W: 3570 L: 3658 D: 13049
sprt @ 10+0.1 th 1 King Danger Tune
16-10-20 Voy rHist diff
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 16441 W: 2938 L: 2993 D: 10510
sprt @ 10+0.1 th 1 rHist might be a better metric to use...
16-10-20 Voy scoreLMR diff
LLR: -2.97 (-2.94,2.94) [0.00,5.00]
Total: 13396 W: 2343 L: 2412 D: 8641
sprt @ 10+0.1 th 1 stc
16-10-20 cib cutNodeTweak diff
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 7077 W: 1179 L: 1274 D: 4624
sprt @ 10+0.1 th 1 CutNode logic tweak.
16-10-19 Elb trapped_knight diff
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 16777 W: 2952 L: 3006 D: 10819
sprt @ 10+0.1 th 1 Take 2
16-10-17 vdv valgrind diff
LLR: 2.95 (-2.94,2.94) [-3.00,1.00]
Total: 187869 W: 33303 L: 33463 D: 121103
sprt @ 10+0.1 th 1 simplification & valgrind error fix. See also https://github.com/official-stockfish/Stockfish/pull/829
16-10-19 Voy sc4' diff
LLR: -2.25 (-2.94,2.94) [0.00,5.00]
Total: 20758 W: 3741 L: 3748 D: 13269
sprt @ 10+0.1 th 1 Take 3
16-10-18 Elb futility diff
LLR: -2.95 (-2.94,2.94) [0.00,4.00]
Total: 115103 W: 20542 L: 20311 D: 74250
sprt @ 10+0.1 th 1 Try futility depth = 4
16-10-19 fau threat2 diff
LLR: -2.96 (-2.94,2.94) [0.00,4.00]
Total: 43461 W: 7725 L: 7735 D: 28001
sprt @ 10+0.1 th 1 Take 2 on Threats
16-10-19 aji threatlimit3 diff
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 33192 W: 5843 L: 5829 D: 21520
sprt @ 10+0.1 th 1 Threat limit take 3: STC
16-10-19 Voy sc4 diff
LLR: 2.96 (-2.94,2.94) [0.00,5.00]
Total: 27754 W: 5066 L: 4824 D: 17864
sprt @ 10+0.1 th 1 Take 4
16-10-19 Elb trapped_knight diff
LLR: -2.97 (-2.94,2.94) [0.00,5.00]
Total: 8864 W: 1544 L: 1632 D: 5688
sprt @ 10+0.1 th 1 Give penalty for trapped knights at the corners of the board
16-10-19 fau SimplifyB diff
LLR: 0.64 (-2.94,2.94) [-3.00,1.00]
Total: 7608 W: 1368 L: 1347 D: 4893
sprt @ 10+0.1 th 1 Just curious to see if we get any Elo by removing this. (If this illogical just do not approve it).