Stockfish Testing Queue

Finished - 43823 tests

17-09-28 Voy statsUpdate diff
LLR: -2.96 (-2.94,2.94) [0.00,5.00]
Total: 10403 W: 1847 L: 1928 D: 6628
sprt @ 10+0.1 th 1 Don't penalize prior move if current move is exact value.
17-09-28 sni RP-vs-BN diff
LLR: -2.95 (-2.94,2.94) [0.00,4.00]
Total: 5753 W: 999 L: 1136 D: 3618
sprt @ 10+0.1 th 1 Make R+P a little bit closer to B+N
17-09-28 pb0 correctionFactor diff
LLR: -2.96 (-2.94,2.94) [0.00,5.00]
Total: 13374 W: 2307 L: 2376 D: 8691
sprt @ 10+0.1 th 1 Take 2
17-09-26 vdv furtherExcept diff
LLR: 3.38 (-2.94,2.94) [-3.00,1.00]
Total: 213377 W: 37463 L: 37641 D: 138273
sprt @ 10+0.1 th 1 stc, take 2
17-09-28 IIv imbalance_tune diff
2690/50000 iterations
5829/100000 games played
100000 @ 10+0.1 th 1 Tune imbalance tables, starting from an alternative starting point which was neutral in SPRT test.
17-09-28 Voy ttPVcutoff3 diff
LLR: -2.96 (-2.94,2.94) [0.00,5.00]
Total: 3493 W: 593 L: 704 D: 2196
sprt @ 10+0.1 th 1 Take 3...not sure why these tests are failing so bad...
17-09-28 vdv furtherExcept diff
LLR: 2.96 (-2.94,2.94) [-3.00,1.00]
Total: 52063 W: 9192 L: 9123 D: 33748
sprt @ 10+0.1 th 1 stc, take 3
17-09-28 Voy ttPVcutoff2 diff
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 3163 W: 500 L: 611 D: 2052
sprt @ 10+0.1 th 1 Take 2...
17-09-26 fau tuneKingShSt diff
42182/45000 iterations
88135/90000 games played
90000 @ 10+0.1 th 1 Tuning King Shelter and King Storm after recent changes
17-09-28 sg unsupportable_pawns2 diff
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 12491 W: 2174 L: 2246 D: 8071
sprt @ 10+0.1 th 1 Last try seems really bad so try opposite (rank <= 3). Optimized by removing !supportable condition and its calculation because its always true for rank <= 3
17-09-28 Voy ttPVcutoff diff
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 5353 W: 897 L: 999 D: 3457
sprt @ 10+0.1 th 1 Allow TT-cutoff for PV moves. This may now be stronger since we added new stats updating for TT-cutoffs..
17-09-28 pb0 correctionFactor diff
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 8059 W: 932 L: 1024 D: 6103
sprt @ 60+0.6 th 1 LTC: Dynamic correction factor according to game phase (non-pawn material at root position)
17-09-28 Elb bttm diff
LLR: -1.34 (-2.94,2.94) [0.00,5.00]
Total: 9000 W: 1466 L: 1486 D: 6048
sprt @ 5+0.05 th 5 Take into account better worker threads for time management
17-09-28 pb0 correctionFactor diff
LLR: 2.96 (-2.94,2.94) [0.00,5.00]
Total: 22921 W: 4134 L: 3914 D: 14873
sprt @ 10+0.1 th 1 Dynamic correction factor according to game phase (bench should be correct now)
17-09-28 sni risk7 diff
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 11340 W: 1937 L: 2014 D: 7389
sprt @ 10+0.1 th 1 Simpler version
17-09-28 pb0 captureKiller2 diff
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 10476 W: 1762 L: 1843 D: 6871
sprt @ 10+0.1 th 1 A much less invasive approach.
17-09-28 nit king_danger_test diff
LLR: -2.96 (-2.94,2.94) [0.00,5.00]
Total: 18009 W: 3226 L: 3275 D: 11508
sprt @ 10+0.1 th 1 experimenting with the king danger reduction
17-09-27 sni opb diff
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 65155 W: 11493 L: 11346 D: 42316
sprt @ 10+0.1 th 1 take 3
17-09-27 IIv imbalance diff
LLR: -2.96 (-2.94,2.94) [0.00,4.00]
Total: 46311 W: 8282 L: 8282 D: 29747
sprt @ 10+0.1 th 1 As imbalance tables are certainly hard to tune, I'll take few tries by guess and luck - take 2 - priority -1 so far.
17-09-27 nit king_danger_test diff
LLR: -1.87 (-2.94,2.94) [0.00,5.00]
Total: 15633 W: 2784 L: 2797 D: 10052
sprt @ 10+0.1 th 1
17-09-27 IIv imbalance diff
LLR: -2.95 (-2.94,2.94) [0.00,4.00]
Total: 26853 W: 4826 L: 4891 D: 17136
sprt @ 10+0.1 th 1 As imbalance tables are certainly hard to tune, I'll take few tries by guess and luck - take 1.
17-09-27 And statsExpectation diff
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 26957 W: 4664 L: 4677 D: 17616
sprt @ 10+0.1 th 1 Reduce LMR for good moves following bad ones, if there is an expectation of beating beta based on the stat eval
17-09-27 Voy seTweak diff
LLR: -2.97 (-2.94,2.94) [0.00,5.00]
Total: 26121 W: 4596 L: 4612 D: 16913
sprt @ 10+0.1 th 1 Reduce depth by one for captureOrPromotion when checking for failing low
17-09-27 Voy seTweak2 diff
LLR: -2.96 (-2.94,2.94) [0.00,5.00]
Total: 3200 W: 500 L: 611 D: 2089
sprt @ 10+0.1 th 1 stc
17-09-27 Elb opb2 diff
LLR: -2.96 (-2.94,2.94) [0.00,5.00]
Total: 13054 W: 2280 L: 2350 D: 8424
sprt @ 10+0.1 th 1 Variant of opposite colored bishops and queens patch
17-09-27 sni opb diff
LLR: -2.96 (-2.94,2.94) [0.00,5.00]
Total: 29314 W: 5132 L: 5135 D: 19047
sprt @ 10+0.1 th 1 take 2
17-09-27 Voy ttCapturePiece diff
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 16549 W: 2812 L: 2868 D: 10869
sprt @ 10+0.1 th 1 stc
17-09-27 sni bothflanks5 diff
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 5845 W: 974 L: 1074 D: 3797
sprt @ 10+0.1 th 1 Take 2
17-09-27 Fis tune33 diff
9741/10000 iterations
20000/20000 games played
20000 @ 10+0.1 th 1 tune a new pawn idea
17-09-27 sni opb diff
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 21403 W: 2631 L: 2681 D: 16091
sprt @ 60+0.6 th 1 LTC: opposite bishops and queens
17-09-27 sni bothflanks5 diff
LLR: -2.95 (-2.94,2.94) [0.00,4.00]
Total: 20061 W: 3547 L: 3635 D: 12879
sprt @ 10+0.1 th 1 Increase the bothFlank coefficient in initiative()
17-09-27 pb0 captureKiller2 diff
LLR: -2.96 (-2.94,2.94) [0.00,5.00]
Total: 15792 W: 1938 L: 2006 D: 11848
sprt @ 60+0.6 th 1 LTC: Another shot at Stefan's Killer Captures idea...
17-09-27 sni opb diff
LLR: 2.95 (-2.94,2.94) [0.00,5.00]
Total: 4931 W: 927 L: 783 D: 3221
sprt @ 10+0.1 th 1 Opposite bishops and queens
17-09-26 sni protection diff
LLR: -2.80 (-2.94,2.94) [0.00,5.00]
Total: 11552 W: 1408 L: 1483 D: 8661
sprt @ 60+0.6 th 1 Take 4. Speculative LTC.
17-09-26 pb0 captureKiller2 diff
LLR: 2.95 (-2.94,2.94) [0.00,5.00]
Total: 89138 W: 15869 L: 15374 D: 57895
sprt @ 10+0.1 th 1 Another shot at Stefan's Killer Captures idea...
17-09-26 Voy lmrCapture2 diff
LLR: -2.96 (-2.94,2.94) [0.00,4.00]
Total: 54501 W: 9616 L: 9590 D: 35295
sprt @ 10+0.1 th 1 take 2
17-09-26 pb0 captureKiller3 diff
LLR: -2.96 (-2.94,2.94) [0.00,5.00]
Total: 46298 W: 8213 L: 8144 D: 29941
sprt @ 10+0.1 th 1 Take 2: one killer move only, bench: 4418863
17-09-26 sni opb diff
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 17006 W: 3004 L: 3057 D: 10945
sprt @ 10+0.1 th 1 Only apply the "opposite colored bishop" scaling factor in positions without material imbalance.
17-09-26 And lmrcapture1 diff
LLR: -2.97 (-2.94,2.94) [0.00,5.00]
Total: 28503 W: 5048 L: 5054 D: 18401
sprt @ 10+0.1 th 1 Reduce LMR on captures for PvNodes that follow a good capture or a first move.
17-09-26 Voy lmrCapture diff
LLR: -2.96 (-2.94,2.94) [0.00,5.00]
Total: 33323 W: 5853 L: 5839 D: 21631
sprt @ 10+0.1 th 1 Allow lmr capture only when depth <8
17-09-26 sg lmr_capture diff
LLR: -2.96 (-2.94,2.94) [0.00,5.00]
Total: 30993 W: 5549 L: 5543 D: 19901
sprt @ 10+0.1 th 1 Less reduction for capture only if last move has negative statScore
17-09-25 sni protection diff
LLR: -1.42 (-2.94,2.94) [0.00,5.00]
Total: 111412 W: 19756 L: 19352 D: 72304
sprt @ 10+0.1 th 1 Take 4
17-09-26 sni protection diff
LLR: -2.97 (-2.94,2.94) [0.00,5.00]
Total: 15455 W: 2676 L: 2737 D: 10042
sprt @ 10+0.1 th 1 Take 5
17-09-26 sg lmr_capture diff
LLR: -2.14 (-2.94,2.94) [0.00,5.00]
Total: 25031 W: 4387 L: 4373 D: 16271
sprt @ 10+0.1 th 1 Less reduction for captures only if last move was a capture too.
17-09-26 Elb maxStats diff
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 6601 W: 1111 L: 1208 D: 4282
sprt @ 10+0.1 th 1 Take 2: also use minStats
17-09-26 pb0 lmrEscape diff
LLR: -2.96 (-2.94,2.94) [0.00,5.00]
Total: 20140 W: 3491 L: 3532 D: 13117
sprt @ 10+0.1 th 1 detect escapes at cut nodes too
17-09-26 Elb maxStats diff
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 12632 W: 2152 L: 2224 D: 8256
sprt @ 10+0.1 th 1 Decrease reduction further if stats bigger than opponents maxStats.
17-09-25 sni protection diff
LLR: -2.24 (-2.94,2.94) [0.00,5.00]
Total: 42187 W: 7545 L: 7462 D: 27180
sprt @ 10+0.1 th 1 Malus if opponent has strong protections
17-09-26 Voy beta2' diff
LLR: -2.96 (-2.94,2.94) [0.00,5.00]
Total: 21038 W: 3188 L: 3231 D: 14619
sprt @ 20+0.2 th 1 This seems to be TC sensitive...so try it with 20'-0.2
17-09-20 vdv timeExcept2 diff
LLR: -0.96 (-2.94,2.94) [-4.00,0.00]
Total: 52405 W: 8574 L: 8791 D: 35040
sprt @ 5+0.05 th 5 stc, take 2, threaded version