Stockfish Testing Queue

Finished - 40836 tests

14-01-20 inf less_pv diff
LLR: -0.95 (-2.94,2.94) [0.00,6.00]
Total: 43657 W: 6764 L: 6601 D: 30292
sprt @ 60+0.05 th 1 LTC for glinscott: Remove most PV distinctions
14-01-20 inf material diff
LLR: -0.45 (-2.94,2.94) [-1.50,4.50]
Total: 731 W: 127 L: 141 D: 463
sprt @ 15+0.05 th 1 LTC: PawnValue - 5
14-01-20 hwi see_simp diff
ELO: -0.39 +-1.9 (95%) LOS: 34.3%
Total: 40000 W: 6190 L: 6235 D: 27575
40000 @ 60+0.05 th 1 Retest of patch to make sure it's not affecting performance.
14-01-20 inf material diff
LLR: -2.95 (-2.94,2.94) [0.00,6.00]
Total: 10336 W: 1594 L: 1648 D: 7094
sprt @ 60+0.05 th 1 LTC: PawnValue - 5
14-01-20 inf pptune diff
ELO: -16.92 +-3.1 (95%) LOS: 0.0%
Total: 20000 W: 3724 L: 4697 D: 11579
20000 @ 15+0.05 th 1 Further simplify calculation of k-factor. Based on elo estimate tests in the framework by Joachim and some local tests, this should not regress. (Hopefully!)
14-01-20 inf pptune2 diff
LLR: 2.96 (-2.94,2.94) [-1.50,4.50]
Total: 60877 W: 11420 L: 11156 D: 38301
sprt @ 15+0.05 th 1 Try a combination of 2 of Joachim's patches that almost passed.
14-01-20 sg followup_moves diff
LLR: -2.95 (-2.94,2.94) [-1.50,4.50]
Total: 13064 W: 2316 L: 2387 D: 8361
sprt @ 15+0.05 th 1 non-capture/promotion check (take 4)
14-01-20 sg followup_moves diff
LLR: -2.95 (-2.94,2.94) [-1.50,4.50]
Total: 6010 W: 1067 L: 1156 D: 3787
sprt @ 15+0.05 th 1 non-capture/promotion check (take 3). Sorry, wrong bench on first attempt.
14-01-20 pec tm_fix diff
ELO: -2.35 +-2.7 (95%) LOS: 4.3%
Total: 20000 W: 3028 L: 3163 D: 13809
20000 @ 60+0.05 th 1 Measure elo change at LTC
14-01-20 gli less_pv diff
LLR: -2.95 (-2.94,2.94) [-1.50,4.50]
Total: 18380 W: 3354 L: 3411 D: 11615
sprt @ 15+0.05 th 1 Don't prune the root node! Also, fix <= alpha case
14-01-20 joa pp_blockSq diff
LLR: 2.96 (-2.94,2.94) [-1.50,4.50]
Total: 4941 W: 945 L: 828 D: 3168
sprt @ 15+0.05 th 1 Bonus for pp eval when our king attacks the blockSq
14-01-20 inf pptune2 diff
LLR: -2.97 (-2.94,2.94) [0.00,6.00]
Total: 10727 W: 1654 L: 1707 D: 7366
sprt @ 60+0.05 th 1 LTC: Try a combination of 2 of Joachim's patches that almost passed.
14-01-20 rst more_ext diff
LLR: -2.96 (-2.94,2.94) [-1.50,4.50]
Total: 10739 W: 1906 L: 1983 D: 6850
sprt @ 15+0.05 th 1 more check extensions
14-01-20 rst less_ext diff
LLR: -2.96 (-2.94,2.94) [-1.50,4.50]
Total: 17699 W: 3279 L: 3338 D: 11082
sprt @ 15+0.05 th 1 less check extensions
14-01-21 rst lesser_PV diff
LLR: 2.96 (-2.94,2.94) [-1.50,4.50]
Total: 52077 W: 9694 L: 9454 D: 32929
sprt @ 15+0.05 th 1 even lesser PV distinctions
14-01-21 inf pp_blockSq diff
LLR: -2.96 (-2.94,2.94) [0.00,6.00]
Total: 19524 W: 3017 L: 3030 D: 13477
sprt @ 60+0.05 th 1 LTC for joachim: Bonus for pp eval when our king attacks the blockSq
14-01-21 inf pptune2 diff
LLR: -2.95 (-2.94,2.94) [-1.50,4.50]
Total: 4768 W: 867 L: 960 D: 2941
sprt @ 15+0.05 th 1 Take 2: Increasing k-factor for ebonus
14-01-21 inf pptune2 diff
LLR: -2.95 (-2.94,2.94) [-1.50,4.50]
Total: 9885 W: 1808 L: 1887 D: 6190
sprt @ 15+0.05 th 1 Take 3
14-01-21 inf pptune2 diff
LLR: -2.97 (-2.94,2.94) [-1.50,4.50]
Total: 69452 W: 13078 L: 13001 D: 43373
sprt @ 15+0.05 th 1 Final Take
14-01-21 joa pp_blockSq diff
LLR: -2.96 (-2.94,2.94) [-1.50,4.50]
Total: 41601 W: 7672 L: 7669 D: 26260
sprt @ 15+0.05 th 1 Lesser bonus for pp eval when our king attacks the blockSq
14-01-21 joa pp_blockSq diff
LLR: 0.60 (-2.94,2.94) [-1.50,4.50]
Total: 128000 W: 23809 L: 23455 D: 80736
sprt @ 15+0.05 th 1 More bonus for pp eval when our king attacks the blockSq
14-01-21 mco master diff
ELO: 32.49 +-2.0 (95%) LOS: 100.0%
Total: 40000 W: 8818 L: 5088 D: 26094
40000 @ 60+0.05 th 1 Regression test after SEE simplification (8moves_v3 book)
14-01-21 pec time_trouble diff
ELO: 0.28 +-3.1 (95%) LOS: 56.9%
Total: 20000 W: 4223 L: 4207 D: 11570
20000 @ 5+0.05 th 1 Handle time trouble. Take 1
14-01-21 dra lesser_PV diff
LLR: -2.96 (-2.94,2.94) [0.00,6.00]
Total: 9011 W: 1329 L: 1390 D: 6292
sprt @ 60+0.05 th 1 LTC for : even lesser PV distinctions
14-01-21 pec time_trouble diff
ELO: 6.53 +-3.5 (95%) LOS: 100.0%
Total: 20000 W: 5366 L: 4990 D: 9644
20000 @ 0.05+0.05 th 1 Handle time trouble. Take 1. Play on increment only.
14-01-21 pec time_trouble diff
ELO: 3.86 +-3.4 (95%) LOS: 98.8%
Total: 20000 W: 4981 L: 4759 D: 10260
20000 @ 1+0.05 th 1 Handle time trouble. Take 1. Play with disproportionately large increment
14-01-21 jki pvraz2 diff
ELO: 1.28 +-1.4 (95%) LOS: 96.4%
Total: 100000 W: 21215 L: 20846 D: 57939
100000 @ 5+0.05 th 1 Test value of enabling razoring in PV at very fast time control
14-01-21 pec time_trouble diff
ELO: -0.36 +-2.0 (95%) LOS: 36.7%
Total: 40000 W: 7239 L: 7280 D: 25481
40000 @ 15+0.05 th 1 Handle time trouble. Take 1. STC test for neutrality.
14-01-22 pec play_on_increment diff
ELO: 4.07 +-3.1 (95%) LOS: 99.5%
Total: 20000 W: 4153 L: 3919 D: 11928
20000 @ 0.25+0.25 th 1 Test play on increment patch separately. Its is equivalent to time_trouble patch in case of increment play only, it should have no effect when base/increment > 100 (i.e STC, LTC), and help when base/increment < 20. Testing with larger increment to have the same total time as STC
14-01-22 pec time_trouble diff
ELO: 18.71 +-3.4 (95%) LOS: 100.0%
Total: 20000 W: 5495 L: 4419 D: 10086
20000 @ 5 th 1 Handle time trouble. Take 1. Check that gain at no increment tc is preserved
14-01-22 pec time_trouble diff
ELO: -0.84 +-1.9 (95%) LOS: 19.2%
Total: 40000 W: 6147 L: 6244 D: 27609
40000 @ 60+0.05 th 1 Handle time trouble. Take 1. LTC test for neutrality at 60+0.05.
14-01-22 pec time_trouble diff
ELO: 4.33 +-3.0 (95%) LOS: 99.7%
Total: 20000 W: 4110 L: 3861 D: 12029
20000 @ 15 th 1 Handle time trouble. Take 1. Check if no increment gains hold for STC eqiuvalent TC
14-01-22 jki pvraz2 diff
ELO: 0.14 +-1.3 (95%) LOS: 58.3%
Total: 100000 W: 18382 L: 18342 D: 63276
100000 @ 15+0.05 th 1 Test value of enabling razoring in PV at fast time control
14-01-22 pec time_trouble_2 diff
ELO: 3.18 +-3.4 (95%) LOS: 96.6%
Total: 20000 W: 5117 L: 4934 D: 9949
20000 @ 0.05+0.05 th 1 Take 2. Keeping good results and removing bad. Verify at very short tc. That patch acts as intended
14-01-22 joa master diff
ELO: 17.88 +-14.5 (95%) LOS: 99.2%
Total: 1128 W: 318 L: 260 D: 550
40000 @ 60+0.05 th 1 Regression test after SEE simplification (chess960_book_3moves.pgn) to test for resolution of the chess960book. Low prio.
14-01-22 jos 3fold_fix^ diff
ELO: -6.66 +-2.0 (95%) LOS: 0.0%
Total: 40000 W: 6464 L: 7231 D: 26305
40000 @ 15+0.05 th 1 Fix 3-fold repetition, take 2.
14-01-22 jos 3fold_fix diff
ELO: -1.09 +-2.0 (95%) LOS: 14.9%
Total: 40000 W: 7136 L: 7261 D: 25603
40000 @ 15+0.05 th 1 Fix 3-fold repetition, take 3. Seems unlogical, but let's check it, too.
14-01-22 dor master diff
ELO: 41.01 +-2.2 (95%) LOS: 100.0%
Total: 40000 W: 10847 L: 6147 D: 23006
40000 @ 60+0.05 th 1 Regression test after SEE simplification with shallow book (2moves_v1). With 8moves_v3 book it was +32.49 ELO.
14-01-22 pec time_trouble diff
ELO: 3.29 +-2.4 (95%) LOS: 99.6%
Total: 30000 W: 5849 L: 5565 D: 18586
30000 @ 5+0.25 th 1 Handle time trouble. Take 1. STC equivalent time control for base/inc = 20
14-01-23 pec time_trouble diff
ELO: 2.25 +-1.9 (95%) LOS: 98.9%
Total: 40000 W: 6468 L: 6209 D: 27323
40000 @ 60 th 1 Handle time trouble. Take 1. LTC test with no increment 60+0
14-01-23 pec time_trouble diff
ELO: 0.96 +-2.0 (95%) LOS: 83.2%
Total: 40000 W: 6722 L: 6611 D: 26667
40000 @ 1+1 th 1 Handle time trouble. Take 1. LTC test for playing only on increment 1+1
14-01-23 pec time_trouble diff
ELO: 1.73 +-1.9 (95%) LOS: 96.0%
Total: 40000 W: 6537 L: 6338 D: 27125
40000 @ 16+0.8 th 1 Handle time trouble. Take 1. LTC test for base/inc=20 16 + 0.8
14-01-23 jos 3fold_fix diff
ELO: -0.88 +-2.1 (95%) LOS: 20.1%
Total: 40000 W: 7241 L: 7342 D: 25417
40000 @ 15+0.05 th 1 Fix 3-fold repetition, take 5. Common implementation like everybody else does.
14-01-23 pec time_trouble_3 diff
LLR: 2.96 (-2.94,2.94) [-3.00,3.00]
Total: 28831 W: 5767 L: 5660 D: 17404
sprt @ 15 th 1 Take 3. Simpler version. Check if simple version without fade down is better at no increment
14-01-23 uri fix_bug_infinite diff
ELO: 1.98 +-2.1 (95%) LOS: 97.0%
Total: 40000 W: 7485 L: 7257 D: 25258
40000 @ 15+0.05 th 1 test fixing some bug that cause the program not to solve 8/8/8/2p5/1pp5/brpp4/1pprp2P/qnkbK3 w - - 0 1 or to have wrong evaluation that is too high.
14-01-23 joa rr diff
LLR: -2.97 (-2.94,2.94) [-1.50,4.50]
Total: 2321 W: 528 L: 635 D: 1158
sprt @ 5+0.05 th 1 changing the calculation of rr to a linear function
14-01-24 pec time_trouble_3 diff
LLR: -2.95 (-2.94,2.94) [-3.00,3.00]
Total: 11305 W: 2863 L: 2978 D: 5464
sprt @ 0.05+0.05 th 1 Take 3. Simpler version. Check if no spillover of sudden death tc of simpler version is better at non-zero increment
14-01-24 pec time_trouble_3 diff
LLR: -2.95 (-2.94,2.94) [-3.00,3.00]
Total: 9996 W: 2379 L: 2492 D: 5125
sprt @ 1+0.05 th 1 Take 3. Simpler version. Check if no spillover of sudden death tc of simpler version is better at non-zero increment at base/inc = 20
14-01-24 pec time_trouble_3 diff
LLR: -1.80 (-2.94,2.94) [-3.00,3.00]
Total: 105542 W: 22012 L: 22078 D: 61452
sprt @ 5+0.05 th 1 Take 3. Simpler version. Check if no spillover of sudden death tc of simpler version is better at non-zero increment at base/inc = 100. One more quick test while other patches getting approved or crashing
14-01-24 oki 3fold_fix diff
ELO: -1.50 +-1.9 (95%) LOS: 5.8%
Total: 40000 W: 5996 L: 6169 D: 27835
40000 @ 60+0.05 th 1 LTC for jo: LTFix 3-fold repetition, take 5. Common implementation like everybody else does.