Stockfish Testing Queue

Finished - 27247 tests

09-02-16 Ro Majority2 diff
LLR: -1.52 (-2.94,2.94) [0.00,5.00]
Total: 7331 W: 1326 L: 1359 D: 4646
sprt @ 10+0.1 th 1 Trying a much larger penalty (S(0, 40) instead of S(0, 8))
09-02-16 SC qsearchVarianceRazor diff
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 12924 W: 2299 L: 2369 D: 8256
sprt @ 10+0.1 th 1 If a several quiet moves are found, make possible to lower the value returned by qsearch by a something times the standard deviation of the qsearch values, in order to account for the uncertainity in the qsearch results. Take 1: use it in razoring with sigma/3.
09-02-16 My c=15 diff
ELO: 0.37 +-5.9 (95%) LOS: 54.9%
Total: 3724 W: 528 L: 524 D: 2672
20000 @ 30+0.3 th 3 Is there really no obvious ELO change from: http://tests.stockfishchess.org/tests/view/56b95fb10ebc590247cdfcdc Test how this scales at MTC, with hardcoded change to be certain. (Low priority)
09-02-16 Vo inCheck diff
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 5388 W: 927 L: 1029 D: 3432
sprt @ 10+0.1 th 1 Don't update killers and cm if in check.
08-02-16 SC seeRefactoring diff
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 21959 W: 4089 L: 4119 D: 13751
sprt @ 10+0.1 th 1 Save some see calls, take 2.
08-02-16 Ro Majority2 diff
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 17910 W: 3293 L: 3341 D: 11276
sprt @ 10+0.1 th 1 Fixed signature. Is there a problem with fishbench ?
08-02-16 Vo keepKillers diff
LLR: -2.96 (-2.94,2.94) [0.00,5.00]
Total: 6964 W: 1239 L: 1335 D: 4390
sprt @ 10+0.1 th 1 See what happens if we don't clear killers.
08-02-16 Vo killers diff
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 36348 W: 4860 L: 4857 D: 26631
sprt @ 60+0.6 th 1 LTC: Take 2
08-02-16 My pt diff
LLR: -2.95 (-2.94,2.94) [0.00,4.00]
Total: 15309 W: 2862 L: 2966 D: 9481
sprt @ 10+0.1 th 1 Adjust passed pawn file values
08-02-16 jo matimb_prq diff
LLR: -2.94 (-2.94,2.94) [0.00,4.00]
Total: 15776 W: 2935 L: 3037 D: 9804
sprt @ 10+0.1 th 1 Try some earlier values.
08-02-16 My bp diff
LLR: -0.89 (-2.94,2.94) [0.00,5.00]
Total: 346 W: 52 L: 89 D: 205
sprt @ 10+0.1 th 1 Adjust bishop pawn penalty
08-02-16 SC historyPruning diff
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 549 W: 63 L: 190 D: 296
sprt @ 10+0.1 th 1 Prune with history also at higher depths. With normal bounds.
08-02-16 SC threadConfidence diff
LLR: -2.97 (-2.94,2.94) [0.00,5.00]
Total: 7691 W: 1171 L: 1264 D: 5256
sprt @ 10+0.1 th 3 Take 2. More depth and less idx.
08-02-16 SC futilitySee diff
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 22658 W: 4142 L: 4170 D: 14346
sprt @ 10+0.1 th 1 Tweak on futility pruning.
08-02-16 Vo killers diff
LLR: 2.95 (-2.94,2.94) [0.00,5.00]
Total: 26541 W: 4986 L: 4746 D: 16809
sprt @ 10+0.1 th 1 Take 2
08-02-16 SC threadConfidence diff
LLR: -2.96 (-2.94,2.94) [0.00,5.00]
Total: 10405 W: 1652 L: 1734 D: 7019
sprt @ 10+0.1 th 3 This one, inspired from best_of_seven, looked quite strong at 3 + 0.03 locally. As we lack data, it seems sensible to increase hash when going to more threads. Reschedule as I forgot to set Threads=3.
07-02-16 pb best_of_seven diff
LLR: -1.64 (-2.94,2.94) [0.00,5.00]
Total: 23944 W: 3848 L: 3823 D: 16273
sprt @ 3+0.03 th 23 Peter encouraged me to do this test and he will join his 3 23core machines. After his machines joined this test I will decrease throughput. After 10000 games played we will eventually reconsider if together with the old result this test is worth to go on...
08-02-16 SC threadConfidence diff
LLR: -0.00 (-2.94,2.94) [0.00,5.00]
Total: 31 W: 3 L: 3 D: 25
sprt @ 10+0.1 th 1 This one, inspired from best_of_seven, looked quite strong at 3 + 0.03 locally. As we lack data, it seems sensible to increase hash when going to more threads.
08-02-16 Ro HealthyMajority diff
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 13942 W: 2566 L: 2631 D: 8745
sprt @ 10+0.1 th 1 Yet another logical fix, some majorities were penalized too quickly Last one, this was not my day.
07-02-16 Ro HealthyMajority diff
LLR: -2.96 (-2.94,2.94) [0.00,5.00]
Total: 25677 W: 4722 L: 4737 D: 16218
sprt @ 10+0.1 th 1 Sorry for this. Retry the green STC test with a small adjustment (33% of the applicable cases were skipped by mistake) Txs to mstembera
07-02-16 Vo killers diff
LLR: -0.03 (-2.94,2.94) [0.00,5.00]
Total: 18 W: 2 L: 3 D: 13
sprt @ 10+0.1 th 1 Take 2...
07-02-16 Ro HealthyMajority diff
LLR: -2.12 (-2.94,2.94) [0.00,5.00]
Total: 24002 W: 4548 L: 4533 D: 14921
sprt @ 10+0.1 th 1 Let's try S(0, 15) penalty before going LTC
07-02-16 Ro HealthyMajority diff
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 26134 W: 4865 L: 4877 D: 16392
sprt @ 10+0.1 th 1 Try the other way around, penalize when a pawn majority is (possibly) compromised by unhealthy pawns. S(0,5) penalty
07-02-16 Vo killers diff
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 17132 W: 3181 L: 3232 D: 10719
sprt @ 10+0.1 th 1 Score main killer move in Evasions stage.
07-02-16 Vo combo diff
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 20480 W: 3756 L: 3793 D: 12931
sprt @ 10+0.1 th 1 Patch combination that came close to passing.
07-02-16 Ro HealthyMajority diff
LLR: 2.94 (-2.94,2.94) [0.00,5.00]
Total: 20753 W: 3925 L: 3710 D: 13118
sprt @ 10+0.1 th 1 S(0, 10) penalty when compromised.
07-02-16 Vo se-cm diff
ELO: -0.52 +-3.9 (95%) LOS: 39.6%
Total: 10000 W: 1609 L: 1624 D: 6767
10000 @ 10+0.1 th 3 I believe this patch may benefit smp. Try a quick measurement.
07-02-16 Vo lmrTweak3 diff
LLR: -2.95 (-2.94,2.94) [0.00,4.00]
Total: 11032 W: 1979 L: 2098 D: 6955
sprt @ 10+0.1 th 1 stc
07-02-16 Ro HealthyMajority diff
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 49262 W: 9236 L: 9147 D: 30879
sprt @ 10+0.1 th 1 Take 2: S(0, 5)
07-02-16 My sd2 diff
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 12180 W: 2184 L: 2257 D: 7739
sprt @ 10+0.1 th 1 Another scaling take - opp bishops & piece tweak
07-02-16 Vo lmrTweak2 diff
LLR: -2.96 (-2.94,2.94) [0.00,5.00]
Total: 17477 W: 3227 L: 3277 D: 10973
sprt @ 10+0.1 th 1 Another tweak...
07-02-16 Ro HealthyMajority diff
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 6890 W: 1209 L: 1305 D: 4376
sprt @ 10+0.1 th 1 Take 1 S(0,10) per majority
07-02-16 Vo lmrTweak diff
LLR: -2.96 (-2.94,2.94) [0.00,5.00]
Total: 15402 W: 2840 L: 2899 D: 9663
sprt @ 10+0.1 th 1 LMR Idea.
05-02-16 pe tm diff
LLR: -2.95 (-2.94,2.94) [-3.00,1.00]
Total: 81768 W: 11020 L: 11279 D: 59469
sprt @ 60+0.6 th 1 LTC. Tuned, after 45K games
06-02-16 Ro ChallengingControl diff
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 45584 W: 8547 L: 8474 D: 28563
sprt @ 10+0.1 th 1 Another tweak to mobility area.
06-02-16 Ro NotSoWeak diff
LLR: -2.94 (-2.94,2.94) [0.00,5.00]
Total: 12815 W: 2322 L: 2392 D: 8101
sprt @ 10+0.1 th 1 The other way around. More penalty when attacked. S(4, 4)
06-02-16 jh legality diff
ELO: -7.68 +-3.7 (95%) LOS: 0.0%
Total: 10000 W: 1374 L: 1595 D: 7031
10000 @ 40+0.4 th 1 Move legality check to beginning of search, check scaling.
06-02-16 Vo se-qs_cm diff
LLR: -2.97 (-2.94,2.94) [0.00,5.00]
Total: 10237 W: 1861 L: 1943 D: 6433
sprt @ 10+0.1 th 1 Update CM at both qs and se. (correct bench)
06-02-16 Vo alpha-cm diff
LLR: -1.87 (-2.94,2.94) [0.00,5.00]
Total: 4809 W: 878 L: 937 D: 2994
sprt @ 10+0.1 th 1 Always update cm if value > alpha
06-02-16 Ro NotSoWeak diff
LLR: -2.96 (-2.94,2.94) [0.00,5.00]
Total: 9703 W: 1816 L: 1900 D: 5987
sprt @ 10+0.1 th 1 Last try, smaller compensation S(4, 4)
06-02-16 Vo qs-cmh diff
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 8898 W: 1606 L: 1693 D: 5599
sprt @ 10+0.1 th 1 Update cmh in qs
06-02-16 Ro NotSoWeak diff
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 6436 W: 1145 L: 1243 D: 4048
sprt @ 10+0.1 th 1 Fixed version.
05-02-16 pb best_of_seven diff
ELO: 2.22 +-4.1 (95%) LOS: 85.5%
Total: 10000 W: 1857 L: 1793 D: 6350
10000 @ 3+0.03 th 11 Quick check. Since helpers with higher idx (= higher skip-sizes) are known to deliver less quality, choose best-thread among the first seven threads only.
06-02-16 Vo qs-cm diff
LLR: -2.96 (-2.94,2.94) [0.00,5.00]
Total: 47938 W: 8859 L: 8778 D: 30301
sprt @ 10+0.1 th 1 Update cm in qs.
05-02-16 jo lmrt2 diff
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 57532 W: 10766 L: 10641 D: 36125
sprt @ 10+0.1 th 1 Allow LMR to drop into qsearch.
06-02-16 Ro NotSoWeak diff
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 10566 W: 1906 L: 1986 D: 6674
sprt @ 10+0.1 th 1 Small adjustment on pawns which are weak but not attacked
05-02-16 Vo shallowPruning diff
LLR: -2.95 (-2.94,2.94) [0.00,4.00]
Total: 32140 W: 6020 L: 6065 D: 20055
sprt @ 10+0.1 th 1 Shallow Pruning Tweak
05-02-16 jh legality diff
ELO: -2.61 +-4.1 (95%) LOS: 10.8%
Total: 10000 W: 1798 L: 1873 D: 6329
10000 @ 10+0.1 th 1 Move legality check to beginning of search
05-02-16 pe tune_easy diff
22140/22500 iterations
45000/45000 games played
45000 @ 10+0.1 th 1 tune second model
05-02-16 pe easy diff
ELO: -2.85 +-2.9 (95%) LOS: 2.9%
Total: 20000 W: 3641 L: 3805 D: 12554
20000 @ 10+0.1 th 1 Measure elo after 36K games