Variant Stockfish Testing Queue

Filter: variant = crazyhouse

Finished - 1166 tests

11-11-17 ia crazyhouse tune_zh_pv diff 
19551/20000 iterations
40000/40000 games played
 40000 @ 10+0.1 th 1 retune piece values after several months of development
12-11-17 ia crazyhouse zh_weak2 diff 
LLR: -2.96 (-2.94,2.94) [0.00,10.00]
Total: 4891 W: 2316 L: 2344 D: 231
 sprt @ 30+0.3 th 1 weak squares. Take 2
11-11-17 ia crazyhouse master diff 
ELO: 5.49 +-9.4 (95%) LOS: 87.3%
Total: 5000 W: 2446 L: 2367 D: 187
 5000 @ 10+0.1 th 1 regression test for upstream merge
11-11-17 ia crazyhouse zh_weak2 diff 
LLR: -2.95 (-2.94,2.94) [0.00,10.00]
Total: 6935 W: 3336 L: 3335 D: 264
 sprt @ 10+0.1 th 1 weak squares. Take 2
11-11-17 X crazyhouse zh_fix_drop_history2 diff 
LLR: -2.96 (-2.94,2.94) [0.00,10.00]
Total: 4739 W: 2266 L: 2297 D: 176
 sprt @ 10+0.1 th 1 Expand arrays, take 2
11-11-17 X crazyhouse zh_fix_drop_history diff 
LLR: -2.96 (-2.94,2.94) [0.00,10.00]
Total: 189 W: 44 L: 142 D: 3
 sprt @ 10+0.1 th 1 Try to fake drops in crazyhouse as being not from the same square but from a square dependent on the piece type. Quite likely broken.
11-11-17 ia crazyhouse zh_weak diff 
LLR: -2.97 (-2.94,2.94) [0.00,10.00]
Total: 304 W: 100 L: 196 D: 8
 sprt @ 10+0.1 th 1 weak squares
03-11-17 dd crazyhouse zh_non_pawn_material diff 
LLR: 2.99 (-2.94,2.94) [-10.00,5.00]
Total: 1257 W: 627 L: 570 D: 60
 sprt @ 30+0.3 th 1 Fix crazyhouse non-pawn material count and pos_is_ok #432
03-11-17 dd crazyhouse zh_non_pawn_material diff 
LLR: 2.94 (-2.94,2.94) [-10.00,5.00]
Total: 1509 W: 755 L: 700 D: 54
 sprt @ 10+0.1 th 1 Fix crazyhouse non-pawn material count and pos_is_ok #432
30-10-17 ia crazyhouse zh_asym3 diff 
LLR: -2.96 (-2.94,2.94) [0.00,10.00]
Total: 3112 W: 1448 L: 1501 D: 163
 sprt @ 30+0.3 th 1 double asymmetry. LTC
30-10-17 ia crazyhouse zh_asym3 diff 
LLR: 2.96 (-2.94,2.94) [0.00,10.00]
Total: 21922 W: 10743 L: 10329 D: 850
 sprt @ 10+0.1 th 1 double asymmetry
29-10-17 ia crazyhouse zh_asym1 diff 
LLR: 2.95 (-2.94,2.94) [0.00,10.00]
Total: 12180 W: 5956 L: 5683 D: 541
 sprt @ 30+0.3 th 1 asymmetric king safety evaluation. LTC
29-10-17 ia crazyhouse zh_asym2 diff 
LLR: -2.96 (-2.94,2.94) [0.00,10.00]
Total: 8760 W: 4220 L: 4193 D: 347
 sprt @ 10+0.1 th 1 asymmetric king safety evaluation. Take 2
29-10-17 ia crazyhouse zh_asym1 diff 
LLR: 2.97 (-2.94,2.94) [0.00,10.00]
Total: 9099 W: 4491 L: 4261 D: 347
 sprt @ 10+0.1 th 1 asymmetric king safety evaluation
29-10-17 ia crazyhouse zh_asym0 diff 
LLR: -2.96 (-2.94,2.94) [-10.00,5.00]
Total: 3090 W: 1434 L: 1522 D: 134
 sprt @ 10+0.1 th 1 check whether breaking symmetry loses elo
09-10-17 ia crazyhouse zh_lmr_simp diff 
LLR: 2.98 (-2.94,2.94) [-10.00,5.00]
Total: 14290 W: 6754 L: 6791 D: 745
 sprt @ 30+0.3 th 1 simplify away LMR tweak for checks. LTC
08-10-17 ia crazyhouse zh_lmr_simp diff 
LLR: 2.95 (-2.94,2.94) [-10.00,5.00]
Total: 5075 W: 2443 L: 2414 D: 218
 sprt @ 10+0.1 th 1 now that checks in qsearch are fixed, maybe we can remove the LMR tweak
07-10-17 ia crazyhouse fishnet-071017 diff 
ELO: 69.17 +-15.1 (95%) LOS: 100.0%
Total: 2000 W: 1142 L: 749 D: 109
 2000 @ 30+0.3 th 1 regression test
05-10-17 ia crazyhouse zh_nmp4 diff 
LLR: -2.95 (-2.94,2.94) [0.00,10.00]
Total: 4453 W: 2081 L: 2114 D: 258
 sprt @ 30+0.3 th 1 tweak NMP parameters. LTC2
05-10-17 ia crazyhouse zh_nmp4 diff 
LLR: -2.96 (-2.94,2.94) [0.00,10.00]
Total: 7354 W: 3487 L: 3479 D: 388
 sprt @ 30+0.3 th 1 tweak NMP parameters. LTC
04-10-17 dd crazyhouse zh_generate diff 
LLR: 2.96 (-2.94,2.94) [-10.00,5.00]
Total: 3042 W: 1456 L: 1413 D: 173
 sprt @ 30+0.3 th 1 Refactor crazyhouse move generation (LTC)
04-10-17 dd crazyhouse zh_generate diff 
LLR: 2.98 (-2.94,2.94) [-10.00,5.00]
Total: 20754 W: 9870 L: 9953 D: 931
 sprt @ 10+0.1 th 1 Refactor crazyhouse move generation
04-10-17 ia crazyhouse zh_nmp4 diff 
LLR: -2.96 (-2.94,2.94) [0.00,10.00]
Total: 11648 W: 5592 L: 5523 D: 533
 sprt @ 10+0.1 th 1 tweak NMP parameters
02-10-17 ia crazyhouse tune_zh_nmp diff 
9656/10000 iterations
20000/20000 games played
 20000 @ 20+0.2 th 1 tune NMP paramters
03-10-17 ia crazyhouse zh_generation1 diff 
LLR: -2.95 (-2.94,2.94) [0.00,10.00]
Total: 20259 W: 9771 L: 9578 D: 910
 sprt @ 10+0.1 th 1 generate drops after normal moves
04-10-17 dd crazyhouse zh_generate diff 
LLR: -2.95 (-2.94,2.94) [0.00,10.00]
Total: 8938 W: 4297 L: 4267 D: 374
 sprt @ 10+0.1 th 1 Refactor crazyhouse move generation
01-10-17 ia crazyhouse zh_movepick2 diff 
LLR: -2.95 (-2.94,2.94) [0.00,10.00]
Total: 7397 W: 3496 L: 3487 D: 414
 sprt @ 30+0.3 th 1 prefer non-drops in move ordering. LTC
02-10-17 ia crazyhouse zh_movepick4 diff 
LLR: -2.95 (-2.94,2.94) [0.00,10.00]
Total: 4342 W: 2059 L: 2095 D: 188
 sprt @ 10+0.1 th 1 only apply move score malus to non-attacking drops
02-10-17 dd crazyhouse crazyhouse_qsearch_drop diff 
LLR: 2.97 (-2.94,2.94) [0.00,10.00]
Total: 501 W: 290 L: 185 D: 26
 sprt @ 30+0.3 th 1 Generate piece and pawn drops in crazyhouse qsearch (LTC)
02-10-17 dd crazyhouse crazyhouse_qsearch_drop diff 
LLR: 2.96 (-2.94,2.94) [0.00,10.00]
Total: 819 W: 446 L: 336 D: 37
 sprt @ 10+0.1 th 1 Generate piece and pawn drops in crazyhouse qsearch
02-10-17 ia crazyhouse zh_movepick3 diff 
LLR: -2.97 (-2.94,2.94) [0.00,10.00]
Total: 2258 W: 1043 L: 1109 D: 106
 sprt @ 10+0.1 th 1 try Vincent's idea
01-10-17 ia crazyhouse zh_movepick2 diff 
LLR: 2.97 (-2.94,2.94) [0.00,10.00]
Total: 6010 W: 2958 L: 2773 D: 279
 sprt @ 10+0.1 th 1 prefer non-drops in move ordering
01-10-17 ia crazyhouse zh_nmp3 diff 
LLR: -2.95 (-2.94,2.94) [0.00,10.00]
Total: 2176 W: 998 L: 1064 D: 114
 sprt @ 10+0.1 th 1 NMP
01-10-17 ia crazyhouse zh_nmp2 diff 
LLR: -2.96 (-2.94,2.94) [0.00,10.00]
Total: 1139 W: 503 L: 585 D: 51
 sprt @ 10+0.1 th 1 deeper NMP verification search
30-09-17 ia crazyhouse blackloss2 diff 
ELO: -47.12 +-9.6 (95%) LOS: 0.0%
Total: 5000 W: 2125 L: 2799 D: 76
 5000 @ 180+1.8 th 1 retest with no adjudication for both sides
29-09-17 ia crazyhouse blackloss2 diff 
ELO: -46.06 +-9.6 (95%) LOS: 0.0%
Total: 5000 W: 2125 L: 2784 D: 91
 5000 @ 180+1.8 th 1 measure white score without book, bypass adjudication, move overhead 200. Hash=64
29-09-17 ia crazyhouse blackloss2 diff 
ELO: -51.30 +-9.6 (95%) LOS: 0.0%
Total: 5000 W: 2086 L: 2819 D: 95
 5000 @ 90+0.9 th 1 measure white score without book, bypass adjudication, move overhead 200. VLTC. Hash=64
29-09-17 ia crazyhouse blackloss2 diff 
ELO: -69.28 +-9.7 (95%) LOS: 0.0%
Total: 5000 W: 1965 L: 2949 D: 86
 5000 @ 60+0.6 th 1 measure white score without book, bypass adjudication.
29-09-17 ia crazyhouse no_adjudication diff 
ELO: 9.10 +-9.4 (95%) LOS: 97.1%
Total: 5000 W: 2473 L: 2342 D: 185
 5000 @ 10+0.1 th 1 test whether asymmetric adjudication causes bias
29-09-17 ia crazyhouse blackloss2 diff 
ELO: -57.51 +-23.0 (95%) LOS: 0.0%
Total: 890 W: 367 L: 513 D: 10
 5000 @ 90+0.9 th 1 measure white score without book, bypass adjudication, move overhead 200. VLTC
29-09-17 ia crazyhouse blackloss2 diff 
ELO: -99.65 +-9.9 (95%) LOS: 0.0%
Total: 5000 W: 1755 L: 3151 D: 94
 5000 @ 20+0.2 th 1 measure white score without book, bypass adjudication.
28-09-17 ia crazyhouse blackloss2 diff 
ELO: -46.41 +-23.1 (95%) LOS: 0.0%
Total: 866 W: 368 L: 483 D: 15
 5000 @ 90+0.9 th 1 measure white score without book, bypass adjudication. VLTC
28-09-17 ia crazyhouse blackloss2 diff 
ELO: -86.74 +-9.8 (95%) LOS: 0.0%
Total: 5000 W: 1835 L: 3058 D: 107
 5000 @ 30+0.3 th 1 measure white score without book, bypass adjudication. LTC
28-09-17 ia crazyhouse blackloss1 diff 
ELO: -115.46 +-10.1 (95%) LOS: 0.0%
Total: 5000 W: 1655 L: 3258 D: 87
 5000 @ 10+0.1 th 1 measure white score without book
28-09-17 ia crazyhouse blackloss2 diff 
ELO: -113.53 +-10.0 (95%) LOS: 0.0%
Total: 5000 W: 1671 L: 3249 D: 80
 5000 @ 10+0.1 th 1 measure white score without book, bypass adjudication
26-09-17 ia crazyhouse zh_probcut diff 
LLR: -2.95 (-2.94,2.94) [0.00,10.00]
Total: 2588 W: 1176 L: 1235 D: 177
 sprt @ 30+0.3 th 1 deeper probcut search. use crazyhouse-p4-f12.epd
26-09-17 ia crazyhouse zh_probcut diff 
LLR: 2.96 (-2.94,2.94) [0.00,10.00]
Total: 4144 W: 2017 L: 1861 D: 266
 sprt @ 30+0.3 th 1 deeper probcut search. use crazyhouse-p4-f12.epd
26-09-17 dd crazyhouse zh_deep_nmp diff 
LLR: -2.96 (-2.94,2.94) [0.00,10.00]
Total: 2055 W: 948 L: 1017 D: 90
 sprt @ 10+0.1 th 1 Skip crazyhouse shallow search null-move pruning
26-09-17 ia crazyhouse zh_null3 diff 
LLR: -2.97 (-2.94,2.94) [0.00,10.00]
Total: 1713 W: 779 L: 853 D: 81
 sprt @ 30+0.3 th 1 skip nmp. LTC
26-09-17 ia crazyhouse zh_probcut diff 
LLR: -2.96 (-2.94,2.94) [0.00,10.00]
Total: 1879 W: 859 L: 930 D: 90
 sprt @ 10+0.1 th 1 deeper probcut search. use crazyhouse-p4-f12.epd