Stockfish Testing Queue

Finished - 22407 tests

17-09-14 lb test diff
LLR: -2.95 (-2.94,2.94) [-1.00,4.00]
Total: 17033 W: 3494 L: 3575 D: 9964
sprt @ 15+0.05 th 1 reduction and razoring not null search
17-09-14 lb test diff
LLR: -2.96 (-2.94,2.94) [-1.00,4.00]
Total: 13944 W: 2851 L: 2941 D: 8152
sprt @ 15+0.05 th 1 idem for razoring
16-09-14 ee ext_verify_g diff
ELO: 1.38 +-2.1 (95%) LOS: 90.5%
Total: 30000 W: 4172 L: 4053 D: 21775
30000 @ 60+0.05 th 1 Measure only the ProbCut like change LTC
17-09-14 lb test diff
LLR: -2.95 (-2.94,2.94) [-1.00,4.00]
Total: 22220 W: 4628 L: 4694 D: 12898
sprt @ 15+0.05 th 1 idem for null search
17-09-14 lb test diff
LLR: -2.95 (-2.94,2.94) [-1.00,4.00]
Total: 17011 W: 3513 L: 3594 D: 9904
sprt @ 15+0.05 th 1 allow a reduced search to dive into a deep QS (possibly skipping the quiet checks first ply and going directly into captures only, or even recaptures only)
17-09-14 lb master diff
ELO: -0.44 +-2.6 (95%) LOS: 36.9%
Total: 29420 W: 6125 L: 6162 D: 17133
30000 @ 15+0.05 th 1 do we really need 30ms/move?
17-09-14 lb master diff
LLR: -0.69 (-2.94,2.94) [-1.50,4.50]
Total: 719 W: 129 L: 152 D: 438
sprt @ 15+0.05 th 1 do we really need 30ms/move? try without the quotes
16-09-14 ur piece_advantage diff
LLR: -2.95 (-2.94,2.94) [-1.50,4.50]
Total: 11068 W: 1771 L: 1847 D: 7450
sprt @ 15+0.05 th 1 fix wrong evaluation of 6q1/8/8/8/3k4/8/8/N1R2K1R b - - 0 1 types.h and of cases when a player has piece advantage with 3 pieces against 2 pieces(usually enough to win)
14-09-14 lb master diff
ELO: 160.64 +-2.8 (95%) LOS: 100.0%
Total: 40000 W: 21520 L: 4240 D: 14240
40000 @ 15+0.05 th 1 master vs. SF 4: 2moves_v1.pgn
15-09-14 jh null_verification_fix diff
LLR: 2.95 (-2.94,2.94) [-3.00,1.00]
Total: 37498 W: 4726 L: 4627 D: 28145
sprt @ 120+0.05 th 1 Test null verification fix at VLTC.
16-09-14 ee ext_verify_g diff
ELO: 0.94 +-2.3 (95%) LOS: 79.2%
Total: 30000 W: 4986 L: 4905 D: 20109
30000 @ 15+0.05 th 1 Measure only the ProbCut like change
14-09-14 jk kinga diff
ELO: 0.57 +-2.1 (95%) LOS: 70.1%
Total: 30000 W: 4330 L: 4281 D: 21389
30000 @ 60+0.05 th 1 Test value of Reduced king safety without queens. (LTC)
15-09-14 Fi time diff
ELO: 0.09 +-2.8 (95%) LOS: 52.5%
Total: 20000 W: 3307 L: 3302 D: 13391
20000 @ 15+0.05 th 1 Check if tuning time usage values is converging to something better or just noise.
15-09-14 aj knight_closed_C diff
LLR: -2.96 (-2.94,2.94) [-1.50,4.50]
Total: 8643 W: 1396 L: 1478 D: 5769
sprt @ 15+0.05 th 1 bonus for k,nights in closed positions: last attempt: STC
15-09-14 jo clop-matimb3 diff
ELO: 1.47 +-2.7 (95%) LOS: 86.1%
Total: 30000 W: 6944 L: 6817 D: 16239
30000 @ 7+0.05 th 1 2nd quick check with values after 750k games.
14-09-14 lb master diff
ELO: 141.22 +-2.5 (95%) LOS: 100.0%
Total: 40000 W: 18808 L: 3389 D: 17803
40000 @ 15+0.05 th 1 master vs. SF 4: 8moves_v3.pgn
15-09-14 Fi time_tuning diff
24403/25000 iterations
50000/50000 games played
50000 @ 15+0.05 th 1 Tune time management parameters.
14-09-14 jk timesimp diff
LLR: 2.96 (-2.94,2.94) [-4.00,0.00]
Total: 15777 W: 2403 L: 2298 D: 11076
sprt @ 30+0.05 th 1 Verify that time management simplification is not a regression (Should be no functional change), retry.
15-09-14 ee ext_verify_f diff
ELO: -0.88 +-2.3 (95%) LOS: 22.4%
Total: 30000 W: 4980 L: 5056 D: 19964
30000 @ 15+0.05 th 1 Measure independant R for verification search plus a ProbCut like change
13-09-14 jk master diff
ELO: 15.90 +-1.8 (95%) LOS: 100.0%
Total: 40000 W: 6815 L: 4986 D: 28199
40000 @ 60+0.05 th 1 Regression test, STC [f5622cd5ec7836e8 vs. sf5]
14-09-14 jh nullmove diff
LLR: -2.95 (-2.94,2.94) [0.00,4.00]
Total: 40692 W: 5510 L: 5548 D: 29634
sprt @ 60+0.05 th 1 Tune null move pruning (cap R).
14-09-14 gl cap_null diff
LLR: -2.97 (-2.94,2.94) [-1.50,4.50]
Total: 21693 W: 3603 L: 3654 D: 14436
sprt @ 15+0.05 th 1 Cap value reduction in null move to depth/4
14-09-14 gl king_danger diff
LLR: -2.97 (-2.94,2.94) [-1.50,4.50]
Total: 5329 W: 894 L: 985 D: 3450
sprt @ 15+0.05 th 1 Increase king danger threshold to 25.
14-09-14 aj knight_closed_B diff
LLR: -2.99 (-2.94,2.94) [-1.50,4.50]
Total: 4064 W: 670 L: 765 D: 2629
sprt @ 15+0.05 th 1 Bonus for knights in closed positions(when there are more than 4 blocked pawns) : STC
14-09-14 hw fm3s diff
LLR: -2.95 (-2.94,2.94) [-1.50,4.50]
Total: 5949 W: 938 L: 1026 D: 3985
sprt @ 15+0.05 th 1 Try a new futility margin
14-09-14 jo shelter diff
LLR: -2.97 (-2.94,2.94) [-1.50,4.50]
Total: 13544 W: 2757 L: 2827 D: 7960
sprt @ 15+0.05 th 1 If no enemy queen, halve shelter-storm penalty.
14-09-14 aj full_threats_A diff
LLR: -2.97 (-2.94,2.94) [-1.50,4.50]
Total: 19913 W: 3344 L: 3399 D: 13170
sprt @ 15+0.05 th 1 Evaluate all threats(increase ThreatenedByPawn) : STC
14-09-14 aj knight_closed_A diff
ELO: -7.53 +-4.9 (95%) LOS: 0.1%
Total: 6181 W: 938 L: 1072 D: 4171
20000 @ 15+0.05 th 1 Bonus for knights in closed positions: Tune 1
13-09-14 lb test diff
LLR: -2.97 (-2.94,2.94) [0.00,5.00]
Total: 19607 W: 3455 L: 3498 D: 12654
sprt @ 60+0.05 th 1 try an LTC anyway, as the framework is empty. this patch aims to help in some zugzwang situations. it's likely to be one of those "zero plus" elo patches that will fail, though...
13-09-14 jk timesimp diff
LLR: -2.96 (-2.94,2.94) [-4.00,0.00]
Total: 23138 W: 3406 L: 3626 D: 16106
sprt @ 30+0.05 th 1 Verify that time management simplification is not a regression (Should be no functional change)
13-09-14 Fi time_bias diff
LLR: -2.97 (-2.94,2.94) [-1.50,4.50]
Total: 10020 W: 1647 L: 1726 D: 6647
sprt @ 15+0.05 th 1 Use more time on positions that have more moves to choose from.
13-09-14 gl king_danger diff
LLR: -2.95 (-2.94,2.94) [-1.50,4.50]
Total: 2967 W: 474 L: 570 D: 1923
sprt @ 15+0.05 th 1 Slight increase in king danger threshold (10 -> 15)
13-09-14 gl king_danger diff
LLR: -2.94 (-2.94,2.94) [-1.50,4.50]
Total: 2395 W: 410 L: 509 D: 1476
sprt @ 15+0.05 th 1 Only apply R reduction if king not in danger.
12-09-14 ur king_safety_eval_test diff
ELO: -0.89 +-2.6 (95%) LOS: 25.2%
Total: 20000 W: 2881 L: 2932 D: 14187
20000 @ 60+0.05 th 1 measuring the value of avoiding king safety evaluation when the attacker has no queen.
13-09-14 jk kinga diff
ELO: 2.08 +-2.3 (95%) LOS: 96.1%
Total: 30000 W: 5322 L: 5142 D: 19536
30000 @ 15+0.05 th 1 Test value of Reduced king safety without queens.
13-09-14 jo clop-matimb3 diff
ELO: -1.17 +-2.4 (95%) LOS: 17.2%
Total: 30000 W: 5659 L: 5760 D: 18581
30000 @ 7+0.05 th 1 Quick check of the first set of new values. I started a last tuning session with CLOP, with a small change. Now all values seem to be converging much better, hopefully.
13-09-14 lb test diff
LLR: -2.97 (-2.94,2.94) [-1.00,4.00]
Total: 49683 W: 10301 L: 10290 D: 29092
sprt @ 15+0.05 th 1 skip step 13 when the *opponent* has no pieces (fixes previous patch)
13-09-14 lb safety diff
LLR: -2.95 (-2.94,2.94) [-3.00,1.00]
Total: 8422 W: 1771 L: 1952 D: 4699
sprt @ 15+0.05 th 1 always evaluate king safety (failed because I inverted the signatures)
13-09-14 lb test diff
LLR: -2.96 (-2.94,2.94) [-1.00,4.00]
Total: 19232 W: 3920 L: 3995 D: 11317
sprt @ 15+0.05 th 1 skip step 13 in pawn endgames
13-09-14 lb qsnott diff
LLR: -2.96 (-2.94,2.94) [-4.00,0.00]
Total: 1113 W: 162 L: 329 D: 622
sprt @ 15+0.05 th 1 no TT in QS (realistic hash pressure)
12-09-14 gl king_danger diff
LLR: -2.97 (-2.94,2.94) [-1.50,4.50]
Total: 2111 W: 322 L: 421 D: 1368
sprt @ 15+0.05 th 1 Tweak search when king is in danger
12-09-14 My king_en diff
LLR: -2.96 (-2.94,2.94) [-1.50,4.50]
Total: 61233 W: 10346 L: 10301 D: 40586
sprt @ 15+0.05 th 1 Add enemies defended by king to protected enemies. Low priority.
12-09-14 ur king_safety_eval_test diff
ELO: -0.56 +-2.9 (95%) LOS: 35.3%
Total: 20000 W: 3560 L: 3592 D: 12848
20000 @ 15+0.05 th 1 measuring the value of avoiding king safety evaluation when the attacker has no queen.
12-09-14 My bckwrd_pwn diff
LLR: -2.96 (-2.94,2.94) [-1.50,4.50]
Total: 2571 W: 413 L: 511 D: 1647
sprt @ 15+0.05 th 1 Tweak backward pawn definition
12-09-14 Fi easymove diff
LLR: -2.96 (-2.94,2.94) [-1.50,4.50]
Total: 7927 W: 1293 L: 1377 D: 5257
sprt @ 15+0.05 th 1 move quicker for easy moves
11-09-14 jo clop-matimb2 diff
ELO: -2.95 +-1.9 (95%) LOS: 0.1%
Total: 50000 W: 9541 L: 9965 D: 30494
50000 @ 7+0.05 th 1 Check current values.
11-09-14 rt TM_C1 diff
LLR: 2.95 (-2.94,2.94) [-1.50,4.50]
Total: 12956 W: 2050 L: 1920 D: 8986
sprt @ 30+0.05 th 1 Add a pawn push metric into the experimental time management system.
11-09-14 rt TM_C2 diff
LLR: -2.95 (-2.94,2.94) [-1.50,4.50]
Total: 5491 W: 753 L: 841 D: 3897
sprt @ 30+0.05 th 1 Pars the updated version against the main branch.
09-09-14 sg ks_simplification diff
LLR: -2.96 (-2.94,2.94) [-3.00,1.00]
Total: 27564 W: 4480 L: 4678 D: 18406
sprt @ 15+0.05 th 1 king safety: set endgame value to zero
09-09-14 ee attackedGuards_1 diff
LLR: -2.95 (-2.94,2.94) [-1.50,4.50]
Total: 9391 W: 1539 L: 1619 D: 6233
sprt @ 15+0.05 th 1 Count attacked pieces in King zone, test 2