Stockfish Testing Queue

Finished - 33316 tests

14-04-07 gli queen_mob diff
LLR: 2.95 (-2.94,2.94) [0.00,6.00]
Total: 6336 W: 1010 L: 882 D: 4444
sprt @ 60+0.05 th 1 LTC: Restrict queen mobility to safe squares, but still allow unsafe queen moves to count as king attacks
14-04-08 lbr imbalance_test diff
LLR: -1.88 (-2.94,2.94) [0.00,4.00]
Total: 46387 W: 9398 L: 9328 D: 27661
sprt @ 6+0.03 th 1 see if we quadratic minorDiff gains something
14-04-08 inf storm_tune diff
ELO: -10.48 +-3.8 (95%) LOS: 0.0%
Total: 14684 W: 3080 L: 3523 D: 8081
20000 @ 5+0.05 th 1 Take 3
14-04-08 inf storm_tune diff
ELO: -12.63 +-6.4 (95%) LOS: 0.0%
Total: 5286 W: 1135 L: 1327 D: 2824
20000 @ 5+0.05 th 1 Simplification and tuning: Take 1
14-04-08 inf storm_tune diff
ELO: -17.02 +-3.9 (95%) LOS: 0.0%
Total: 14831 W: 3214 L: 3940 D: 7677
20000 @ 5+0.05 th 1 Take 4: Just one more quick one.
14-04-08 sni pawns_on_semi_open_file diff
LLR: -2.95 (-2.94,2.94) [-1.50,4.50]
Total: 5385 W: 977 L: 1068 D: 3340
sprt @ 15+0.05 th 1 Add penalty for backward unsupported pawns on semi-open files
14-04-08 sni backward diff
LLR: -2.96 (-2.94,2.94) [-1.50,4.50]
Total: 27269 W: 5024 L: 5058 D: 17187
sprt @ 15+0.05 th 1 Pawns which are blocked and at the base of a chain are considered backwards
14-04-08 lbr imbalance_test diff
LLR: -2.96 (-2.94,2.94) [0.00,4.00]
Total: 17760 W: 3570 L: 3665 D: 10525
sprt @ 6+0.03 th 1 bump up bishop pair
14-04-08 sni backward diff
LLR: -2.96 (-2.94,2.94) [-1.50,4.50]
Total: 22645 W: 4206 L: 4252 D: 14187
sprt @ 15+0.05 th 1 Blocked bases of chains are considered backward, and malus depends on rank
14-04-08 sni backward diff
LLR: -0.13 (-2.94,2.94) [-1.50,4.50]
Total: 2235 W: 403 L: 402 D: 1430
sprt @ 15+0.05 th 1 Isolated pawns can never be backward
14-04-08 sni backward diff
LLR: -2.97 (-2.94,2.94) [-1.50,4.50]
Total: 11675 W: 2081 L: 2156 D: 7438
sprt @ 15+0.05 th 1 Same as master, except that the backward malus depends on rank, not on file.
14-04-08 Fis optimize_score diff
LLR: -2.96 (-2.94,2.94) [-1.50,4.50]
Total: 23787 W: 4253 L: 4297 D: 15237
sprt @ 15+0.05 th 1 Optimize mg_value() and apply_weight() to use shift instead of divide. See commit for more info.
14-04-08 lbr imbalance diff
LLR: -2.95 (-2.94,2.94) [-4.00,0.00]
Total: 9322 W: 1594 L: 1782 D: 5946
sprt @ 15+0.05 th 1 simplified imbalance: last attempt.
14-04-08 gli low_mob diff
LLR: -0.08 (-2.94,2.94) [-1.50,4.50]
Total: 65232 W: 11773 L: 11610 D: 41849
sprt @ 15+0.05 th 1 Penalty for pieces with low mobility on edge
14-04-08 sni backward diff
LLR: -2.97 (-2.94,2.94) [-1.50,4.50]
Total: 10082 W: 1766 L: 1845 D: 6471
sprt @ 15+0.05 th 1 Isolated pawns can never be backward
14-04-08 inf storm_tune diff
ELO: -10.54 +-3.7 (95%) LOS: 0.0%
Total: 14572 W: 2970 L: 3412 D: 8190
20000 @ 5+0.05 th 1 Take 5: The last simplification idea was bad, so this time I don't fiddle around experimenting. These are the original values obtained from CLOP.
14-04-08 inf storm_tune diff
ELO: -6.91 +-3.8 (95%) LOS: 0.0%
Total: 14021 W: 2985 L: 3264 D: 7772
20000 @ 5+0.05 th 1 Take 6: After the clop tuning, stockfish seems to be able to solve this position without having to resort to the hacky adhoc rule: r2qk2r/ppp2p2/2npbn2/2b1p3/2P1P1P1/2NB1PPp/PPNP3K/R1BQ1R2 b kq - 0 13
14-04-08 gli low_mob diff
LLR: -0.03 (-2.94,2.94) [-1.50,4.50]
Total: 37 W: 7 L: 8 D: 22
sprt @ 15+0.05 th 1 Penalize pieces with low mobility on the edge of the board (fixed EdgeBB - thanks Joerg!)
14-04-08 gli low_mob diff
LLR: 2.96 (-2.94,2.94) [-1.50,4.50]
Total: 81857 W: 14652 L: 14342 D: 52863
sprt @ 15+0.05 th 1 Penalize pieces with low mobility on the edge of the board (fixed EdgeBB - thanks Joerg!) (also sync master to other queen mobility change)
14-04-08 gli master diff
ELO: 47.70 +-2.1 (95%) LOS: 100.0%
Total: 40000 W: 10146 L: 4689 D: 25165
40000 @ 60+0.05 th 1 Regression test after Queen mobility change
14-04-08 Fis optimize_score diff
LLR: -2.96 (-2.94,2.94) [-1.50,4.50]
Total: 13156 W: 2272 L: 2343 D: 8541
sprt @ 15+0.05 th 1 Optimize mg_value() and apply_weight() to use shift instead of divide. See commit for more info. Take 2
14-04-08 jos q_imbalance2 diff
LLR: -2.95 (-2.94,2.94) [0.00,6.00]
Total: 26061 W: 3960 L: 3945 D: 18156
sprt @ 60+0.05 th 1 LTC: Higher penalty. Take 3.
14-04-09 jki rmob diff
LLR: -2.97 (-2.94,2.94) [-1.50,4.50]
Total: 3857 W: 651 L: 746 D: 2460
sprt @ 15+0.05 th 1 rook mob, exclude minor attacks
14-04-09 gli low_mob diff
LLR: 2.97 (-2.94,2.94) [0.00,6.00]
Total: 45400 W: 6999 L: 6697 D: 31704
sprt @ 60+0.05 th 1 LTC: Penalize pieces with low mobility on the edge of the board (fixed EdgeBB - thanks Joerg!) (also sync master to other queen mobility change)
14-04-10 sni type_of_score2 diff
LLR: -2.97 (-2.94,2.94) [-1.50,4.50]
Total: 10562 W: 1829 L: 1907 D: 6826
sprt @ 15+0.05 th 1 Experimental : use type punning and 64 bits registers for score. This time this is a no functional change (for the number of nodes).
14-04-10 sto rook_mobility diff
LLR: -2.95 (-2.94,2.94) [-1.50,4.50]
Total: 2960 W: 495 L: 592 D: 1873
sprt @ 15+0.05 th 1 Apply the recent Gary's idea about queen mobility to rooks too.
14-04-10 sni r_bnn_imbalance3 diff
LLR: -2.96 (-2.94,2.94) [-1.50,4.50]
Total: 85515 W: 15197 L: 15086 D: 55232
sprt @ 15+0.05 th 1 Change imbalance correction: take 1
14-04-10 sni r_bnn_imbalance3 diff
LLR: -2.97 (-2.94,2.94) [-1.50,4.50]
Total: 8910 W: 1564 L: 1646 D: 5700
sprt @ 15+0.05 th 1 Change imbalance correction: take 2
14-04-10 sni r_bnn_imbalance3 diff
LLR: -2.97 (-2.94,2.94) [-1.50,4.50]
Total: 8966 W: 1560 L: 1642 D: 5764
sprt @ 15+0.05 th 1 Change imbalance correction: take 3
14-04-10 Fis piece_mob diff
LLR: -2.95 (-2.94,2.94) [-1.50,4.50]
Total: 1697 W: 262 L: 362 D: 1073
sprt @ 15+0.05 th 1 Extend Gary's queen mobility idea to other pieces.
14-04-10 Fis mob_area diff
LLR: -2.96 (-2.94,2.94) [-1.50,4.50]
Total: 3812 W: 648 L: 743 D: 2421
sprt @ 15+0.05 th 1 Remove all our pieces from mobility area not just pawns and king. Is this a bug fix?
14-04-10 jki qimb2 diff
ELO: 0.78 +-1.9 (95%) LOS: 78.5%
Total: 50000 W: 10166 L: 10054 D: 29780
50000 @ 5+0.05 th 1 queen imbalance, 96k iterations
14-04-10 Fis mob_area diff
LLR: -2.97 (-2.94,2.94) [-1.50,4.50]
Total: 3934 W: 665 L: 760 D: 2509
sprt @ 15+0.05 th 1 Don't exclude our pawns from mobility area.
14-04-10 jki qimb2 diff
LLR: -3.08 (-2.94,2.94) [0.00,4.00]
Total: 116257 W: 20586 L: 20360 D: 75311
sprt @ 15+0.05 th 1 queen imbalance, 96k iterations, sprt [0.0, 4.0], fast, prior -1.
14-04-10 Fis piece_mob2 diff
LLR: -2.96 (-2.94,2.94) [-1.50,4.50]
Total: 1423 W: 212 L: 313 D: 898
sprt @ 15+0.05 th 1 Factor out EvalInfo calculations from evaluate_pieces() so that we are now call order independent. This opens the door for excluding attacked & undefended squares from mobility area. Thanks to Gary & joergoster!
14-04-10 Fis piece_mob2 diff
LLR: -2.96 (-2.94,2.94) [-1.50,4.50]
Total: 3455 W: 550 L: 645 D: 2260
sprt @ 15+0.05 th 1 Factor out EvalInfo calculations from evaluate_pieces() so that we are now call order independent. This opens the door for excluding attacked & undefended squares from mobility area. Thanks to Gary & joergoster! Take 2
14-04-10 jos nomob_queen diff
LLR: -0.77 (-2.94,2.94) [0.00,6.00]
Total: 50032 W: 7379 L: 7190 D: 35463
sprt @ 60+0.05 th 1 LTC: Penalty for queen with low mobility. Take 7.
14-04-10 Fis mob_area diff
LLR: -2.95 (-2.94,2.94) [-1.50,4.50]
Total: 9108 W: 1599 L: 1680 D: 5829
sprt @ 15+0.05 th 1 Don't exclude our king from mobility area.
14-04-10 jos nomob_rook^ diff
LLR: -2.96 (-2.94,2.94) [-1.50,4.50]
Total: 41081 W: 7361 L: 7361 D: 26359
sprt @ 15+0.05 th 1 Penalize rooks with low mobility in the middle of the board. Take 1.
14-04-10 jos nomob_rook diff
LLR: -2.97 (-2.94,2.94) [-1.50,4.50]
Total: 5468 W: 939 L: 1030 D: 3499
sprt @ 15+0.05 th 1 Penalize rooks with low mobility in the middle of the board. Take 2.
14-04-10 Fis mob_area diff
LLR: -2.94 (-2.94,2.94) [-1.50,4.50]
Total: 4258 W: 714 L: 807 D: 2737
sprt @ 15+0.05 th 1 Also exclude our queen from mobility area. Final try.
14-04-11 sni keep_material2 diff
LLR: -2.96 (-2.94,2.94) [-1.50,4.50]
Total: 1255 W: 185 L: 287 D: 783
sprt @ 15+0.05 th 1 Keep material when ahead
14-04-11 sni keep_material2 diff
LLR: -2.96 (-2.94,2.94) [-1.50,4.50]
Total: 2248 W: 363 L: 462 D: 1423
sprt @ 15+0.05 th 1 Swap white and black flexibilities
14-04-11 jos nomob_rook^ diff
LLR: -2.95 (-2.94,2.94) [-1.50,4.50]
Total: 29448 W: 5161 L: 5191 D: 19096
sprt @ 15+0.05 th 1 Penalize rooks with low mobility in the middle of the board. Take 3.
14-04-11 jos nomob_rook diff
LLR: -2.97 (-2.94,2.94) [-1.50,4.50]
Total: 7723 W: 1319 L: 1404 D: 5000
sprt @ 15+0.05 th 1 Penalize rooks with low mobility in the middle of the board. Take 4.
14-04-11 jos nomob_queen diff
LLR: 2.95 (-2.94,2.94) [-1.50,4.50]
Total: 42878 W: 7718 L: 7506 D: 27654
sprt @ 15+0.05 th 1 Penalty for queen with low mobility. Take 8. (Final try)
14-04-11 sni keep_material2 diff
LLR: -2.97 (-2.94,2.94) [-1.50,4.50]
Total: 9377 W: 1664 L: 1745 D: 5968
sprt @ 15+0.05 th 1 Take 3
14-04-11 sni keep_material2 diff
LLR: -2.97 (-2.94,2.94) [-1.50,4.50]
Total: 28392 W: 5025 L: 5058 D: 18309
sprt @ 15+0.05 th 1 Same flexibility formula for both colors
14-04-11 gli doubled diff
LLR: 2.96 (-2.94,2.94) [-1.50,4.50]
Total: 38339 W: 6849 L: 6649 D: 24841
sprt @ 15+0.05 th 1 Reduce penalty for doubled pawns far away from each other
14-04-11 gli lowmob diff
LLR: 2.95 (-2.94,2.94) [-3.00,1.00]
Total: 36705 W: 6537 L: 6443 D: 23725
sprt @ 15+0.05 th 1 Remove LowMobPenalty, and move into psq/mobility tables. Run as simplification