Stockfish Testing Queue

Pending - 2 tests 0.6 hrs

12-12-17 sg update_stats2 diff
LLR: -2.40 (-2.94,2.94) [0.00,5.00]
Total: 51347 W: 9408 L: 9290 D: 32649
sprt @ 10+0.1 th 1 More stats bonus if best move count > 15
16-12-17 sg update_stats_exclude diff
Pending... (0.2 hrs)
sprt @ 10+0.1 th 1 Exclude tt move from negative stats update

Active - 2 tests

13-12-17 sg update_stats diff
LLR: -2.53 (-2.94,2.94) [0.00,5.00]
Total: 65069 W: 8321 L: 8214 D: 48534
sprt @ 60+0.6 th 1 LTC: More stats bonus if best move found at expected ALL node.
03-12-17 sg master diff
ELO: 37.68 +-1.5 (95%) LOS: 100.0%
Total: 38649 W: 5965 L: 1790 D: 30894
40000 @ 180+1.8 th 1 VLTC now with higher hash 256 MB (see https://groups.google.com/forum/#!topic/fishcooking/A78xxpyU8dI). Regression test until "Use constexpr when makes sense" (even lower throughput)

Finished - 1612 tests

16-12-17 sg update_stats6 diff
LLR: -2.96 (-2.94,2.94) [0.00,5.00]
Total: 15729 W: 2865 L: 2923 D: 9941
sprt @ 10+0.1 th 1 Last test was yellow at LTC, so try now even 50% more update bonus if no tt hit
15-12-17 sg update_stats_b diff
LLR: -2.96 (-2.94,2.94) [0.00,5.00]
Total: 61836 W: 7910 L: 7830 D: 46096
sprt @ 60+0.6 th 1 LTC: Use higher update bonus only for quiet moves
15-12-17 sg update_stats6 diff
LLR: -2.96 (-2.94,2.94) [0.00,5.00]
Total: 82100 W: 14979 L: 14754 D: 52367
sprt @ 10+0.1 th 1 25% greater stats update weight (with cap) for quiet move if no ttHit
14-12-17 sg update_stats_b diff
LLR: 2.97 (-2.94,2.94) [0.00,5.00]
Total: 122168 W: 22521 L: 21872 D: 77775
sprt @ 10+0.1 th 1 Use higher update bonus only for quiet moves
14-12-17 sg update_stats_a diff
LLR: -2.96 (-2.94,2.94) [0.00,5.00]
Total: 7910 W: 1381 L: 1473 D: 5056
sprt @ 10+0.1 th 1 Try the opposite. Use higher update bonus only for continuation history (Park with prio -1)
14-12-17 sg update_stats_a diff
LLR: -3.31 (-2.94,2.94) [0.00,5.00]
Total: 12594 W: 2210 L: 2297 D: 8087
sprt @ 10+0.1 th 1 Use higher update bonus not for continuation history. A variant of my update_stats patch which have at LTC a positive score but probably fails.
14-12-17 sg update_stats_c diff
LLR: -2.97 (-2.94,2.94) [0.00,5.00]
Total: 14675 W: 2650 L: 2713 D: 9312
sprt @ 10+0.1 th 1 Use higher update bonus only for capture moves
14-12-17 sg update_stats4 diff
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 8215 W: 1021 L: 1112 D: 6082
sprt @ 60+0.6 th 1 LTC; Greater stats update if fail low at an expected pv or cut node for previous move
13-12-17 sg update_stats4 diff
LLR: 2.95 (-2.94,2.94) [0.00,5.00]
Total: 14007 W: 2638 L: 2453 D: 8916
sprt @ 10+0.1 th 1 Greater stats update if fail low at an expected pv or cut node for previous move
12-12-17 sg update_stats diff
LLR: 2.95 (-2.94,2.94) [0.00,5.00]
Total: 60657 W: 11130 L: 10747 D: 38780
sprt @ 10+0.1 th 1 More stats bonus if best move found at expected ALL node.
12-12-17 sg PawnDifference diff
LLR: -2.96 (-2.94,2.94) [0.00,4.00]
Total: 26562 W: 4801 L: 4867 D: 16894
sprt @ 10+0.1 th 1 Only a quarter factor
11-12-17 sg queen_mobility2 diff
LLR: -2.97 (-2.94,2.94) [0.00,5.00]
Total: 34236 W: 6267 L: 6246 D: 21723
sprt @ 10+0.1 th 1 Exclude all own pieces for diagonal/non-diagonal check.
11-12-17 sg queen_mobility diff
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 30598 W: 5598 L: 5592 D: 19408
sprt @ 10+0.1 th 1 Less queen mobility if only diagonal or non-diagonal moves exists.
11-12-17 sg queen_mobility5 diff
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 27824 W: 5134 L: 5139 D: 17551
sprt @ 10+0.1 th 1 Now decrease queen mobility if moves only for one line available
12-12-17 sg PawnDifference diff
LLR: -2.96 (-2.94,2.94) [0.00,4.00]
Total: 5494 W: 923 L: 1061 D: 3510
sprt @ 10+0.1 th 1 Tweak pawn imbalance (Inspired by Guardians tests)
03-12-17 sg master diff
ELO: 39.88 +-1.5 (95%) LOS: 100.0%
Total: 40000 W: 6377 L: 1806 D: 31817
40000 @ 180+1.8 th 1 VLTC for comparison (see https://groups.google.com/forum/#!topic/fishcooking/A78xxpyU8dI). Regression test until "Use constexpr when makes sense" (low throughput)
11-12-17 sg queen_mobility3 diff
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 13322 W: 2432 L: 2500 D: 8390
sprt @ 10+0.1 th 1 Exclude for queen mobility all own pieces
10-12-17 sg check_threats diff
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 26791 W: 4838 L: 4849 D: 17104
sprt @ 10+0.1 th 1 Try now safe knight check threats.
10-12-17 sg check_threats diff
LLR: -2.94 (-2.94,2.94) [0.00,5.00]
Total: 13787 W: 2475 L: 2541 D: 8771
sprt @ 10+0.1 th 1 Double bonus
09-12-17 sg check_threats diff
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 24714 W: 4434 L: 4454 D: 15826
sprt @ 10+0.1 th 1 Now add bonus to kingDanger (fixed bench)
09-12-17 sg check_threats diff
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 12877 W: 2298 L: 2368 D: 8211
sprt @ 10+0.1 th 1 Give bonus only if threat is supported by another piece
08-12-17 sg check_threats diff
LLR: -2.96 (-2.94,2.94) [0.00,5.00]
Total: 5734 W: 978 L: 1079 D: 3677
sprt @ 10+0.1 th 1 Now give prepare check bonus at most once
08-12-17 sg check_threats diff
LLR: -2.96 (-2.94,2.94) [0.00,5.00]
Total: 5578 W: 993 L: 1095 D: 3490
sprt @ 10+0.1 th 1 Small bonus for preparing safe queen check threat.
08-12-17 sg nullmove_mobility diff
LLR: -2.94 (-2.94,2.94) [0.00,5.00]
Total: 12891 W: 2288 L: 2358 D: 8245
sprt @ 10+0.1 th 1 No null move pruning if side to move has no piece mobility (excluding king and pawns)
03-12-17 sg master diff
ELO: 43.31 +-1.7 (95%) LOS: 100.0%
Total: 40000 W: 7367 L: 2406 D: 30227
40000 @ 60+0.6 th 1 Regression test until "Use constexpr when makes sense"
05-12-17 sg queenBetween3 diff
LLR: -2.97 (-2.94,2.94) [0.00,5.00]
Total: 33650 W: 6192 L: 6173 D: 21285
sprt @ 10+0.1 th 1 Allow now also check attacks through opponent undefended pieces (excluding pawns). Test against passed queenBetween patch.
05-12-17 sg queenBetween4 diff
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 768 W: 106 L: 232 D: 430
sprt @ 10+0.1 th 1 Allow now also check attacks through opponent pinned pieces. Test against passed queenBetween patch.
05-12-17 sg queenBetween2 diff
LLR: -2.96 (-2.94,2.94) [0.00,5.00]
Total: 18272 W: 3261 L: 3309 D: 11702
sprt @ 10+0.1 th 1 Allow now also diagonal check attacks through opponent rooks. Test against passed queenBetween patch.
05-12-17 sg queenBetween3 diff
LLR: -2.96 (-2.94,2.94) [0.00,5.00]
Total: 14915 W: 2683 L: 2745 D: 9487
sprt @ 10+0.1 th 1 Allow now also check attacks through opponent undefended pieces. Test against passed queenBetween patch.
04-12-17 sg improving2 diff
LLR: -2.95 (-2.94,2.94) [0.00,4.00]
Total: 2636 W: 404 L: 551 D: 1681
sprt @ 10+0.1 th 1 Use for FutilityMoveCount cutNode instead of improving
04-12-17 sg improving diff
LLR: -2.96 (-2.94,2.94) [0.00,5.00]
Total: 13530 W: 2390 L: 2458 D: 8682
sprt @ 10+0.1 th 1 Also set improving if eval curve is locally convex
04-12-17 sg improving diff
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 25180 W: 4523 L: 4541 D: 16116
sprt @ 10+0.1 th 1 Additionally check if opponent eval decreases
04-12-17 sg improving2 diff
LLR: -2.93 (-2.94,2.94) [0.00,5.00]
Total: 7603 W: 1356 L: 1448 D: 4799
sprt @ 10+0.1 th 1 Test now only the condition if opponent eval decreases
03-12-17 sg tuned_imbalance_score diff
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 14782 W: 2678 L: 2740 D: 9364
sprt @ 10+0.1 th 1 Test tuned imbalance scores
02-12-17 sg tune_imbalance_score diff
44977/50000 iterations
96805/100000 games played
100000 @ 10+0.1 th 1 use now scores for material imbalance tables and tune them (82 parameter). C=80 R=0.004
02-12-17 sg bishop_pair_imbalance diff
LLR: -2.96 (-2.94,2.94) [0.00,4.00]
Total: 25575 W: 4568 L: 4638 D: 16369
sprt @ 10+0.1 th 1 Same weight for ours and theirs pawn + compensation. max change 25%
30-11-17 sg phased_contempt3_no_dra diff
ELO: -2.59 +-2.2 (95%) LOS: 1.2%
Total: 20000 W: 2086 L: 2235 D: 15679
20000 @ 60+0.6 th 1 (Fixed book and adjudication) LTC: phased_contempt3 with contempt=15 against current master but with disabled draw adjudication on fishtest. I want to check that the promising results against master are not an artifact of adjudication. (see my post here https://groups.google.com/forum/?fromgroups=#!searchin/fishcooking/no$20adjudication|sort:date/fishcooking/B2psTv9FfoU/BAzTsa38CgAJ)
01-12-17 sg bishop_pair_imbalance diff
LLR: -2.95 (-2.94,2.94) [0.00,4.00]
Total: 21231 W: 3828 L: 3912 D: 13491
sprt @ 10+0.1 th 1 Tweak bishop pair value. More pawns less value. Max change of 25%
01-12-17 sg bishop_pair_imbalance diff
LLR: -2.95 (-2.94,2.94) [0.00,4.00]
Total: 17382 W: 3067 L: 3164 D: 11151
sprt @ 10+0.1 th 1 Max change of 12.5%
30-11-17 sg phased_contempt3_no_dra diff
ELO: -0.90 +-2.2 (95%) LOS: 21.2%
Total: 20000 W: 2081 L: 2133 D: 15786
20000 @ 60+0.6 th 1 For contempt=15 its seems to be a slight loss so try contempt=10.(Fixed book and adjudication) LTC: phased_contempt3 with contempt=10 against current master but with disabled draw adjudication on fishtest. I want to check that the promising results against master are not an artifact of adjudication. (see my post here https://groups.google.com/forum/?fromgroups=#!searchin/fishcooking/no$20adjudication|sort:date/fishcooking/B2psTv9FfoU/BAzTsa38CgAJ)
01-12-17 sg distant_passed_pawns diff
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 25633 W: 4613 L: 4629 D: 16391
sprt @ 10+0.1 th 1 Double bonus
01-12-17 sg distant_passed_pawns2 diff
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 9648 W: 1757 L: 1841 D: 6050
sprt @ 10+0.1 th 1 Give bonus only if opponent has only pawns
01-12-17 sg distant_passed_pawns diff
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 14110 W: 2526 L: 2591 D: 8993
sprt @ 10+0.1 th 1 Try now flat endgame bonus.
01-12-17 sg safe_attacks4 diff
LLR: -2.95 (-2.94,2.94) [0.00,4.00]
Total: 34744 W: 4461 L: 4518 D: 25765
sprt @ 60+0.6 th 1 LTC: Ok increase midgame value of ThreatByAttackOnQueen even further. Test against my passed safe_attacks2 patch.
01-12-17 sg distant_passed_pawns diff
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 13870 W: 2469 L: 2535 D: 8866
sprt @ 10+0.1 th 1 Endgame bonus if two distanct passed pawns exists.
28-11-17 sg safe_attacks4 diff
LLR: -2.44 (-2.94,2.94) [0.00,4.00]
Total: 85501 W: 15449 L: 15288 D: 54764
sprt @ 10+0.1 th 1 Test an idea of Stephane Nicolet (see repo comments) against my passed safe_attacks2 patch.
30-11-17 sg safe_attacks5 diff
LLR: -2.92 (-2.94,2.94) [0.00,4.00]
Total: 61113 W: 11190 L: 11135 D: 38788
sprt @ 10+0.1 th 1 try now a big jump and double up migdame value. Test against my passed safe_attacks2 patch
30-11-17 sg safe_attacks4 diff
LLR: 2.95 (-2.94,2.94) [0.00,4.00]
Total: 56688 W: 10486 L: 10135 D: 36067
sprt @ 10+0.1 th 1 Ok increase midgame value of ThreatByAttackOnQueen even further. Test against my passed safe_attacks2 patch.
28-11-17 sg safe_attacks5 diff
LLR: -0.88 (-2.94,2.94) [0.00,4.00]
Total: 128000 W: 23012 L: 22627 D: 82361
sprt @ 10+0.1 th 1 Last test struggles so raise midgame value a little more. Test against my passed safe_attacks2 patch.
30-11-17 sg phased_contempt3_no_adj diff
ELO: 15.34 +-29.2 (95%) LOS: 84.9%
Total: 136 W: 20 L: 14 D: 102
20000 @ 60+0.6 th 1 LTC: phased_contempt3 with contempt=15 against current master but with disabled adjudication on fishtest. I want to check that the promising results against master are not an artifact of adjudication. (see my post here https://groups.google.com/forum/?fromgroups=#!searchin/fishcooking/no$20adjudication|sort:date/fishcooking/B2psTv9FfoU/BAzTsa38CgAJ)