Stockfish Testing Queue

Finished - 4710 tests

19-08-18 31m tweak_WeakUnopposed diff
LLR: -2.95 (-2.94,2.94) [0.00,4.00]
Total: 20993 W: 3422 L: 3509 D: 14062
sprt @ 60+0.6 th 1 LTC. Eliminate mg-eg gradient altogether.
19-08-18 sg tuned_king_danger diff
LLR: -1.00 (-2.94,2.94) [0.00,4.00]
Total: 11657 W: 1938 L: 1953 D: 7766
sprt @ 60+0.6 th 1 The STC struggles but seems clearly positive (after 116K games LLR around 0.7). So try a speculative LTC with low TP (set STC to pro -1). Test the values of my extended second tuning. Use same scale as master (to the tuning scale 128).
19-08-17 Viz futMarginTuned2 diff
LLR: -2.96 (-2.94,2.94) [0.00,4.00]
Total: 24148 W: 3989 L: 4066 D: 16093
sprt @ 60+0.6 th 1 Tuned values as a parameter tweak. Testing on top of my passed patch, 1/3 TP.
19-08-17 xot vary4 diff
LLR: -2.95 (-2.94,2.94) [0.50,4.50]
Total: 28759 W: 4832 L: 4870 D: 19057
sprt @ 60+0.6 th 1 Sanity check for big search tune 2, test after 45k games. Test against FutilityMargin change by Viz, low tp.
19-08-15 Ala FutilityArrayTune diff
42658/45000 iterations
88801/90000 games played
90000 @ 60+0.6 th 1 The first tune results are showing some promise. Further tuning with lower max depth and with negative range bound to allow for the improving at depth 1 case to go there.
19-08-16 Voy nmpFt6 diff
LLR: -2.95 (-2.94,2.94) [0.00,4.00]
Total: 42423 W: 6998 L: 7017 D: 28408
sprt @ 60+0.6 th 1 Try this at LTC. Since the formula difference more at higher depths. Low TP.
19-08-14 sni master diff
ELO: 32.24 +-1.9 (95%) LOS: 100.0%
Total: 40000 W: 7869 L: 4168 D: 27963
40000 @ 60+0.6 th 1 Regression/progression test against SF10 after "Tweak unsafe checks" of August 14th.
19-08-16 Elb sc_rq diff
LLR: -2.96 (-2.94,2.94) [0.00,3.50]
Total: 17961 W: 2891 L: 3017 D: 12053
sprt @ 60+0.6 th 1 Queen and Rook checks tweak (LTC)
19-08-15 31m TOP_blocked^^ diff
LLR: -2.94 (-2.94,2.94) [0.00,3.50]
Total: 28086 W: 4597 L: 4694 D: 18795
sprt @ 60+0.6 th 1 I'm confident in STC gain, due to multiple similar long yellow runs. But there's no point in continued efforts if none of this replicates at LTC--so pause further STC attempts and attempt speculative LTC for 65K yellow. Low throughput.
19-08-15 Viz futilityArrayLTC diff
LLR: 2.95 (-2.94,2.94) [0.00,3.50]
Total: 33913 W: 5800 L: 5529 D: 22584
sprt @ 60+0.6 th 1 LTC with new master and full TP now to be safe.
19-08-14 Viz futilityArray1 diff
LLR: 2.95 (-2.94,2.94) [0.50,4.50]
Total: 85676 W: 14530 L: 14034 D: 57112
sprt @ 60+0.6 th 1 This is doing reasonable at STC so try LTC for truncated array till depth 7. STC bounds for faster results, 1/3 TP.
19-08-15 xor qstatT diff
LLR: -2.95 (-2.94,2.94) [0.00,3.50]
Total: 13144 W: 2108 L: 2247 D: 8789
sprt @ 60+0.6 th 1 LTC: take 1.
19-08-15 Voy nmpFt diff
LLR: -2.95 (-2.94,2.94) [0.00,3.50]
Total: 18371 W: 3005 L: 3129 D: 12237
sprt @ 60+0.6 th 1 Curious how this yellow will fare at LTC, since nmp is quite finicky. Low TP
19-08-14 sni nmp_bound diff
ELO: -13.86 +-4.0 (95%) LOS: 0.0%
Total: 9959 W: 1508 L: 1905 D: 6546
10000 @ 60+0.6 th 1 Estimate the cost of bounding null move pruning to depth 12
19-08-12 Ala FutilityTune diff
47121/50000 iterations
98364/100000 games played
100000 @ 60+0.6 th 1 Futility margin tuning with an array.
19-08-14 Viz futilityArray2 diff
LLR: -2.95 (-2.94,2.94) [0.50,4.50]
Total: 21536 W: 3618 L: 3685 D: 14233
sprt @ 60+0.6 th 1 Test at LTC since tuning was at LTC. Full values, 1/3 TP.
19-08-11 Voy nmpT2 diff
LLR: -0.40 (-2.94,2.94) [0.00,3.50]
Total: 134901 W: 22750 L: 22395 D: 89756
sprt @ 60+0.6 th 1 See how this scale low tp...
19-08-13 Voy nmpCombo diff
LLR: 2.95 (-2.94,2.94) [0.00,3.50]
Total: 73895 W: 12496 L: 12114 D: 49285
sprt @ 60+0.6 th 1 ltc
19-08-13 sg tweak_unsafe_checks diff
LLR: 2.96 (-2.94,2.94) [0.00,3.50]
Total: 84968 W: 14499 L: 14083 D: 56386
sprt @ 60+0.6 th 1 LTC: Same bonus for all unsafe checks (equivalent to W=148).
19-08-13 31m TOP_blocked diff
LLR: 0.02 (-2.94,2.94) [0.50,4.50]
Total: 178 W: 31 L: 29 D: 118
sprt @ 60+0.6 th 1 Extra S(0, 10) for ThreatByMinor[PAWN] and ThreatByRook[PAWN] if the enemy pawn is blocked.
19-08-13 31m TOP_blocked^ diff
LLR: -0.01 (-2.94,2.94) [0.50,4.50]
Total: 79 W: 13 L: 13 D: 53
sprt @ 60+0.6 th 1 Extra S(0, 10) for ThreatByRook[PAWN] if the enemy pawn is blocked.
19-08-12 MJZ Check_Extension diff
LLR: -0.21 (-2.94,2.94) [0.00,3.50]
Total: 25992 W: 4412 L: 4351 D: 17229
sprt @ 60+0.6 th 1 Limit check extension to lower depths. d < 12. Trying Spec LTC (I will stop it if no clear improvement after 10k games).
19-08-12 xot vary2 diff
LLR: -2.95 (-2.94,2.94) [0.00,3.50]
Total: 16030 W: 2613 L: 2744 D: 10673
sprt @ 60+0.6 th 1 Vary just 9 parameters, use values from full tune. Test directly at ltc since tune was at ltc, low tp.
19-08-12 Viz futilityTweakScale4 diff
LLR: -2.95 (-2.94,2.94) [0.00,4.00]
Total: 18790 W: 3067 L: 3161 D: 12562
sprt @ 60+0.6 th 1 Try this way directly at LTC, normalized TP...
19-08-12 Viz futilityTweakScale7 diff
LLR: -2.95 (-2.94,2.94) [0.00,3.50]
Total: 88487 W: 14804 L: 14732 D: 58951
sprt @ 60+0.6 th 1 Let's see how this one fares at longer TC
19-08-12 Viz futilityTweakScale2 diff
LLR: -2.95 (-2.94,2.94) [0.00,3.50]
Total: 54080 W: 9038 L: 9062 D: 35980
sprt @ 60+0.6 th 1 LTC v3
19-08-12 xot vary2 diff
LLR: -2.95 (-2.94,2.94) [0.50,4.50]
Total: 7729 W: 1267 L: 1390 D: 5072
sprt @ 60+0.6 th 1 Vary 24 parameters, use values from full tune. Test directly at LTC since tune was at LTC, low tp.
19-08-12 Viz futilityTweakOpp1 diff
LLR: -2.95 (-2.94,2.94) [0.00,4.00]
Total: 9532 W: 1666 L: 1790 D: 6076
sprt @ 40+0.4 th 1 This futility margin and pruning tweaks passed STCs in like 5 different variations but failed LTC every single time. Maybe this indicates scaling problem so try the opposite - increase/increase for margin and depth instead of decreasing. x4 TC and 1/4 TP to see how it does there...
19-08-12 Viz futPrTweak2 diff
LLR: -1.62 (-2.94,2.94) [0.00,4.00]
Total: 15050 W: 2555 L: 2591 D: 9904
sprt @ 60+0.6 th 1 LTC
19-08-05 xot vary2tuneb diff
47952/50000 iterations
100001/100000 games played
100000 @ 60+0.6 th 1 Vary 24 params and tune the values that control the variation. Low tp since ltc tune.
19-08-12 Viz futPrTweak1 diff
LLR: -2.96 (-2.94,2.94) [0.00,4.00]
Total: 16293 W: 2716 L: 2818 D: 10759
sprt @ 60+0.6 th 1 ElTeSee
19-08-10 Voy nmpImp2 diff
LLR: 2.96 (-2.94,2.94) [-3.00,1.00]
Total: 55803 W: 8512 L: 8443 D: 38848
sprt @ 120+1.2 th 1 There are some concerns of scaling for this passed patch. Make sure it doesn't regress at VLTC. (Lower TP to not hog up resources)
19-08-05 Fis master diff
ELO: -10.11 +-2.0 (95%) LOS: 0.0%
Total: 40000 W: 6273 L: 7437 D: 26290
40000 @ 60+0.6 th 1 See how much elo contempt 50 loses compared to 0. To be compared to http://tests.stockfishchess.org/tests/view/5cf2ac170ebc5925cf087454 10% throughput.
19-08-09 sg tweak_history2 diff
LLR: -2.49 (-2.94,2.94) [0.00,4.00]
Total: 38068 W: 6348 L: 6356 D: 25364
sprt @ 60+0.6 th 1 LTC: Use as stat bonus for depth > 17 the same value as for depth 17. The max stat bonus is significant lower after recent tuning so perhaps this works now. Test directly at LTC because this change kicks only in for depth >= 18. Low TP.
19-08-09 Voy nmpConst diff
LLR: -2.95 (-2.94,2.94) [0.00,4.00]
Total: 73902 W: 12460 L: 12376 D: 49066
sprt @ 60+0.6 th 1 LTC: Is my nmp green patches simply because of this const?
19-08-08 MJZ NMP-lastMove2 diff
LLR: -2.95 (-2.94,2.94) [0.00,3.50]
Total: 137288 W: 23018 L: 22809 D: 91461
sprt @ 60+0.6 th 1 NMP if staticEval decrease in last move. Threshold 140 - LTC
19-08-07 xot nmptest2 diff
LLR: -1.95 (-2.94,2.94) [0.00,3.50]
Total: 13072 W: 1961 L: 2041 D: 9070
sprt @ 120+1.2 th 1 VLTC: This seems to scale well, did well at stc and similar to master at ltc. Investigate if scaling continues at vltc. Nmp test, -33*d + 11*rd + 57.
19-08-06 Voy nmpImp2 diff
LLR: 2.95 (-2.94,2.94) [0.00,3.50]
Total: 179493 W: 30198 L: 29521 D: 119774
sprt @ 60+0.6 th 1 LTC: I think this may fare better than the current one at LTC. (Place current at pri. -1).
19-08-07 Voy xotCapP diff
LLR: -2.95 (-2.94,2.94) [0.00,4.00]
Total: 92592 W: 15638 L: 15494 D: 61460
sprt @ 60+0.6 th 1 LTC
19-08-08 vdv NMPtweak^ diff
LLR: -2.96 (-2.94,2.94) [0.00,4.00]
Total: 11314 W: 1781 L: 1899 D: 7634
sprt @ 60+0.6 th 1 LTC: Tweak.
19-08-07 xot nmptest2 diff
LLR: -2.95 (-2.94,2.94) [0.00,3.50]
Total: 86658 W: 14477 L: 14410 D: 57771
sprt @ 60+0.6 th 1 Nmp test, 45*depth - 3*rootdepth + 299.
19-08-03 xot searchstats1 diff
ELO: 12.52 +-2.6 (95%) LOS: 100.0%
Total: 19291 W: 3119 L: 2424 D: 13748
20000 @ 180+1.8 th 1 VLTC to see how this big param tune scales after LTC (since tune was done at ltc). See discussion at https://github.com/official-stockfish/Stockfish/pull/2260 Low tp as I don't want to slow down the ltc tests that are currently running too much.
19-08-05 Voy nmpImp diff
LLR: 0.67 (-2.94,2.94) [0.00,3.50]
Total: 86558 W: 14577 L: 14295 D: 57686
sprt @ 60+0.6 th 1 LTC
19-08-06 vdv NMPtweak diff
LLR: -2.95 (-2.94,2.94) [0.00,4.00]
Total: 77238 W: 13056 L: 12961 D: 51221
sprt @ 60+0.6 th 1 LTC: Tweak. Take 4.
19-08-06 xot nmptest2 diff
LLR: -2.95 (-2.94,2.94) [0.00,3.50]
Total: 46653 W: 7774 L: 7819 D: 31060
sprt @ 60+0.6 th 1 LTC: Nmp test, -33*d + 11*rd + 57.
19-08-06 MJZ NMP-rootDepth2 diff
LLR: -0.56 (-2.94,2.94) [0.00,3.50]
Total: 6840 W: 1162 L: 1176 D: 4502
sprt @ 60+0.6 th 1 NMP formula with relative depth - 300 / 180 - limited test @LTC (I will stop it if no improvement after 10k games)
19-08-05 MJZ NMP-ply_lastMove diff
LLR: -0.61 (-2.94,2.94) [0.00,3.50]
Total: 21043 W: 3495 L: 3473 D: 14075
sprt @ 60+0.6 th 1 NMP staticEval formula : try after LTC tuning - spec LTC
19-08-05 Voy xotStats diff
LLR: -2.95 (-2.94,2.94) [0.00,4.00]
Total: 14704 W: 2382 L: 2489 D: 9833
sprt @ 60+0.6 th 1 Use original parameters for stat formula. Suggested to do Spec LTC...low TP.
19-08-05 Voy npmDepth diff
LLR: -1.16 (-2.94,2.94) [0.50,4.50]
Total: 1042 W: 137 L: 193 D: 712
sprt @ 60+0.6 th 1 Get an idea if we dont do npm at higher depth. Will need to test at LTC because of time sensitivity. Will stop if results are obvious. Low TP
19-08-05 Voy xotNMP diff
LLR: -2.96 (-2.94,2.94) [0.00,4.00]
Total: 11360 W: 1856 L: 1974 D: 7530
sprt @ 60+0.6 th 1 LTC Use original parameters for NMP.