Stockfish Testing Queue

Finished - 56674 tests

13-07-08 rap Danger_A diff
LLR: -2.96 (-2.94,2.94)
Total: 51966 W: 10755 L: 10716 D: 30495
sprt @ 15+0.05 th 1 Try to handle better X-Ray checks (Take 1)
13-07-08 rap Danger_B diff
LLR: -2.96 (-2.94,2.94)
Total: 87011 W: 18113 L: 17974 D: 50924
sprt @ 15+0.05 th 1 Try to handle better X-Ray checks (Take 2)
13-07-08 rap Danger_C diff
LLR: -2.95 (-2.94,2.94)
Total: 16132 W: 3355 L: 3417 D: 9360
sprt @ 15+0.05 th 1 Try to handle better X-Ray checks (Take 3)
13-07-09 gli Eval diff
LLR: -1.57 (-2.94,2.94)
Total: 43870 W: 8121 L: 7950 D: 27799
sprt @ 60+0.05 th 1 (tv test) Long TC: Attempt to simplify trapped rook code
13-07-10 rta master diff
LLR: -2.96 (-2.94,2.94)
Total: 6787 W: 1391 L: 1480 D: 3916
sprt @ 15+0.05 th 1 Removes discretion between PvNodes and non-PvNodes in singular node regression. Uses a reverse SPRT as a test of plausibility. (Fail = Success)
13-07-10 rta master diff
LLR: 2.96 (-2.94,2.94)
Total: 5969 W: 1330 L: 1204 D: 3435
sprt @ 15+0.05 th 1 Removes queen PSQT. Likely to fail quickly. Tested with reverse SPRT as it would make eval smoother.
13-07-10 rta master diff
LLR: 2.98 (-2.94,2.94)
Total: 20158 W: 4330 L: 4163 D: 11665
sprt @ 15+0.05 th 1 Longshot test in reverse SPRT. Removes redundant major evaluation term and replaces it with simple quadratic numerals.
13-07-10 gli master diff
ELO: 37.13 +-2.6 (95%) LOS: 100.0%
Total: 28328 W: 7388 L: 4372 D: 16568
40000 @ 60+0.05 th 1 Regression test with IID in check change
13-07-10 mco search diff
LLR: -2.94 (-2.94,2.94)
Total: 8296 W: 1675 L: 1759 D: 4862
sprt @ 15+0.05 th 1 Prune quiet discovered checks in qsearch
13-07-10 mco capKiller diff
LLR: -2.95 (-2.94,2.94)
Total: 42712 W: 8955 L: 8941 D: 24816
sprt @ 15+0.05 th 1 Capture Killer: take 1
13-07-10 mco capKiller diff
LLR: -3.05 (-2.94,2.94)
Total: 9956 W: 2022 L: 2105 D: 5829
sprt @ 15+0.05 th 1 Capture killer: take 2
13-07-10 mco capKiller diff
LLR: -2.95 (-2.94,2.94)
Total: 26021 W: 5360 L: 5394 D: 15267
sprt @ 15+0.05 th 1 Capture killer: take 3
13-07-10 mco capKiller diff
LLR: -2.96 (-2.94,2.94)
Total: 3104 W: 559 L: 657 D: 1888
sprt @ 15+0.05 th 1 Capture killer: take 4
13-07-10 rta SimpleFish_A1 diff
LLR: -2.96 (-2.94,2.94)
Total: 7174 W: 1253 L: 1321 D: 4600
sprt @ 60+0.05 th 1 (Long-TC): Removes discretion between PvNodes and non-PvNodes in singular node regression. The test worked well at fast TC so I used standard SPRT with long TC.
13-07-11 mco capKiller diff
LLR: -2.95 (-2.94,2.94)
Total: 5090 W: 992 L: 1085 D: 3013
sprt @ 15+0.05 th 1 Capture killer:take 5
13-07-11 mco capKiller diff
LLR: -2.97 (-2.94,2.94)
Total: 20933 W: 4333 L: 4382 D: 12218
sprt @ 15+0.05 th 1 Capture killer: take 6
13-07-11 mco search diff
LLR: -2.96 (-2.94,2.94)
Total: 36131 W: 7598 L: 7603 D: 20930
sprt @ 15+0.05 th 1 Don't prune capture discovered checks
13-07-11 tvi 26e573633c1907947fd0fc4 diff
LLR: 2.96 (-2.94,2.94)
Total: 30267 W: 6508 L: 6312 D: 17447
sprt @ 15+0.05 th 1 Just adjust threshold to 4
13-07-11 tvi 4b0f309cae1eff8074e7e43 diff
LLR: 2.97 (-2.94,2.94)
Total: 10514 W: 2283 L: 2144 D: 6087
sprt @ 15+0.05 th 1 Threshold 3 for trapped rook
13-07-11 tvi 4b1d6fe7d26101f4ebca94f diff
LLR: -2.96 (-2.94,2.94)
Total: 22498 W: 4716 L: 4760 D: 13022
sprt @ 15+0.05 th 1 Remove rank condition for trapped rook
13-07-11 jki razor diff
LLR: -2.97 (-2.94,2.94)
Total: 1041 W: 165 L: 270 D: 606
sprt @ 15+0.05 th 1 Try futility-like razoring formula
13-07-11 jki redDepth diff
LLR: -2.96 (-2.94,2.94)
Total: 10688 W: 2194 L: 2272 D: 6222
sprt @ 15+0.05 th 1 Reduce more aggressively at low depths
13-07-11 jki redMc diff
LLR: -2.95 (-2.94,2.94)
Total: 2493 W: 453 L: 553 D: 1487
sprt @ 15+0.05 th 1 More aggressive reductions in the beginning of the move list
13-07-11 jki redCutAll diff
LLR: -2.96 (-2.94,2.94)
Total: 12363 W: 2556 L: 2629 D: 7178
sprt @ 15+0.05 th 1 log-based cut-node reduction formula
13-07-11 rta master diff
LLR: 2.95 (-2.94,2.94)
Total: 35867 W: 8005 L: 7789 D: 20073
sprt @ 15+0.05 th 1 [Reverse SPRT]: Removes redundant major term and replaces it with same-color quadratic coefficient modification. Reverse SPRT as it removes a slightly dubious term in the material file source.
13-07-11 rta QSearch_B1 diff
LLR: -2.95 (-2.94,2.94)
Total: 3894 W: 777 L: 874 D: 2243
sprt @ 15+0.05 th 1 Attempts to improve qsearch by allowing the pruning of checks that only create one new threat.
13-07-11 rta QSearch_A1 diff
LLR: 2.94 (-2.94,2.94)
Total: 32120 W: 6849 L: 6649 D: 18622
sprt @ 15+0.05 th 1 Attempts to improve qsearch by including previously existing attacks for uncuttable checks.
13-07-12 tvi 4b0f309cae1eff8074e7e43 diff
LLR: -1.42 (-2.94,2.94)
Total: 34103 W: 6311 L: 6185 D: 21607
sprt @ 60+0.05 th 1 Long TC: Threshold 3 for trapped rook
13-07-12 rta QSearch_A1 diff
LLR: -2.97 (-2.94,2.94)
Total: 26972 W: 4898 L: 4865 D: 17209
sprt @ 60+0.05 th 1 LONG TC: Attempts to improve qsearch by including previously existing attacks for uncuttable checks.
13-07-12 tvi Eval diff
LLR: -2.97 (-2.94,2.94)
Total: 18351 W: 3935 L: 3991 D: 10425
sprt @ 15+0.05 th 1 Depth 2 threshold for trapped rook
13-07-12 tvi QSearch diff
LLR: 0.72 (-2.94,2.94)
Total: 47540 W: 10004 L: 9842 D: 27694
sprt @ 15+0.05 th 1 Enough material check only at depth 0
13-07-12 jki cut diff
LLR: -2.92 (-2.94,2.94)
Total: 14338 W: 2951 L: 3017 D: 8370
sprt @ 15+0.05 th 1 One more cut-node reduction formula try
13-07-13 tvi QSearch diff
LLR: -0.86 (-2.94,2.94)
Total: 46314 W: 9761 L: 9660 D: 26893
sprt @ 15+0.05 th 1 Try removing enough material condition for futility in QSearch
13-07-13 jki cut3 diff
LLR: -2.94 (-2.94,2.94)
Total: 32490 W: 6786 L: 6801 D: 18903
sprt @ 15+0.05 th 1 Last cut node reduction formula tweak attempt
13-07-13 mco search diff
LLR: 2.97 (-2.94,2.94)
Total: 10061 W: 1948 L: 1790 D: 6323
sprt @ 60+0.05 th 1 Bundle 2 promising Tom's patches: directly tested at LONG TC that is the most critical
13-07-13 mco master diff
ELO: 39.27 +-3.1 (95%) LOS: 100.0%
Total: 20000 W: 5303 L: 3052 D: 11645
20000 @ 60+0.05 th 1 Regression test with Tom's speed up patch in
13-07-13 rap DangerCheck diff
LLR: -2.95 (-2.94,2.94)
Total: 18192 W: 3867 L: 3923 D: 10402
sprt @ 15+0.05 th 1 Mark checks that forcing queen exchange as dangerous.
13-07-13 mco search diff
LLR: -2.95 (-2.94,2.94)
Total: 32573 W: 6981 L: 6995 D: 18597
sprt @ 15+0.05 th 1 Retire UndefendedMinor: let's see if speed improvement pays off
13-07-13 rta StableMaterial_A1 diff
ELO: -61.16 +-3.9 (95%) LOS: 0.0%
Total: 16000 W: 2860 L: 5648 D: 7492
16000 @ 15+0.05 th 1 A long shot attempt to fix the material tables, it remakes them in a more logical format. This should not be better, so it starts out with fix number games. From here, I plan to build upon the new base if it has sufficient hope.
13-07-14 mco clop diff
ELO: -1.27 +-3.1 (95%) LOS: 21.5%
Total: 20000 W: 4239 L: 4312 D: 11449
20000 @ 15+0.05 th 1 Tempo (+5, +5)
13-07-14 mco clop diff
ELO: -1.58 +-3.1 (95%) LOS: 16.1%
Total: 20000 W: 4188 L: 4279 D: 11533
20000 @ 15+0.05 th 1 Tempo (-5, -5)
13-07-14 mco clop diff
ELO: -0.52 +-3.1 (95%) LOS: 37.1%
Total: 20000 W: 4129 L: 4159 D: 11712
20000 @ 15+0.05 th 1 BishopPin (+10, 0)
13-07-14 mco clop diff
ELO: -1.21 +-2.2 (95%) LOS: 14.2%
Total: 40000 W: 8337 L: 8476 D: 23187
40000 @ 15+0.05 th 1 BishopPin (-10, 0)
13-07-14 mco tuning diff
ELO: 2.15 +-2.2 (95%) LOS: 97.2%
Total: 40000 W: 8513 L: 8266 D: 23221
40000 @ 15+0.05 th 1 RookOn7th (+5, +10)
13-07-14 mco tuning diff
ELO: -0.28 +-2.2 (95%) LOS: 40.3%
Total: 40000 W: 8506 L: 8538 D: 22956
40000 @ 15+0.05 th 1 RookOn7th (-5, -10)
13-07-14 mco tuning diff
ELO: -5.84 +-6.4 (95%) LOS: 3.6%
Total: 5000 W: 1057 L: 1141 D: 2802
20000 @ 15+0.05 th 1 QueenOn7th (+3, +8)
13-07-14 mco tuning diff
ELO: -2.87 +-3.1 (95%) LOS: 3.5%
Total: 20000 W: 4101 L: 4266 D: 11633
20000 @ 15+0.05 th 1 ueenOn7th disabled
13-07-14 mco tuning diff
ELO: 0.07 +-3.1 (95%) LOS: 51.7%
Total: 20000 W: 4238 L: 4234 D: 11528
20000 @ 15+0.05 th 1 RookOnPawn (+6 + 10)
13-07-14 mco tuning diff
ELO: -1.06 +-3.2 (95%) LOS: 25.6%
Total: 20000 W: 4303 L: 4364 D: 11333
20000 @ 15+0.05 th 1 RookOnPawn (-5, -10)
13-07-14 rap Allows_A diff
LLR: -2.95 (-2.94,2.94)
Total: 31414 W: 6603 L: 6621 D: 18190
sprt @ 15+0.05 th 1 Try to avoid some false posiove in allows()