Stockfish Testing Queue

Finished - 45441 tests

16-08-29 Voy signAgreement diff
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 13935 W: 2557 L: 2622 D: 8756
sprt @ 10+0.1 th 1 stc
16-08-29 sni pawn_weaknesses3 diff
LLR: -2.96 (-2.94,2.94) [0.00,5.00]
Total: 17138 W: 3133 L: 3185 D: 10820
sprt @ 10+0.1 th 1 Principle of multiple weaknesses, bonus = 8 * x * x * x
16-08-29 Elb rook_ending diff
LLR: -2.94 (-2.94,2.94) [0.00,5.00]
Total: 8297 W: 1173 L: 1263 D: 5861
sprt @ 10+0.1 th 1 Try out rook ending idea by Dave Kittinger: https://groups.google.com/forum/?fromgroups=#!topic/fishcooking/2zrTvxSdWwA
16-08-29 sg probcut2 diff
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 23309 W: 4270 L: 4295 D: 14744
sprt @ 10+0.1 th 1 Try now the maximum
16-08-29 Voy hpt diff
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 6667 W: 1223 L: 1320 D: 4124
sprt @ 10+0.1 th 1 History pruning tweak...based off and against recent simplification.
16-08-27 sni pawn_chains3 diff
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 69994 W: 13100 L: 12921 D: 43973
sprt @ 10+0.1 th 1 Try weight = 1 for the bonus
16-08-27 Fis lazyDelta2 diff
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 11865 W: 1901 L: 1977 D: 7987
sprt @ 5+0.05 th 7 Since lazyDelta is positive but struggling set it to pri -1 and try this more aggressive version.
16-08-29 sg probcut2 diff
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 7733 W: 1381 L: 1473 D: 4879
sprt @ 10+0.1 th 1 Try minimum of old and new threshold. Test against my passed patch.
16-08-28 pec center diff
LLR: -2.96 (-2.94,2.94) [0.00,5.00]
Total: 26390 W: 4323 L: 4342 D: 17725
sprt @ 5+0.05 th 7 Center window towards best move score as currently picked up upon stop. Did this with maybe too much safety but let see how it performs.
16-08-26 SC simpleOppbis diff
LLR: -2.95 (-2.94,2.94) [-3.00,1.00]
Total: 99854 W: 13407 L: 13690 D: 72757
sprt @ 60+0.6 th 1 Only check for strong side having only bishop in opposite bishops endgames. LTC.
16-08-29 Elb lmr_hcor diff
LLR: -2.95 (-2.94,2.94) [0.00,4.00]
Total: 39748 W: 7388 L: 7407 D: 24953
sprt @ 10+0.1 th 1 LMR history correction value tweak (take 2)
16-08-28 Fis helperDepth diff
LLR: -2.96 (-2.94,2.94) [0.00,5.00]
Total: 16455 W: 2670 L: 2728 D: 11057
sprt @ 5+0.05 th 7 Dynamically select helper search depth based on number of threads, max depth completed, and other depths being currently searched.
16-08-28 sg probcut diff
LLR: 2.95 (-2.94,2.94) [0.00,4.00]
Total: 36618 W: 5083 L: 4836 D: 26699
sprt @ 60+0.6 th 1 LTC: Tweak probcut threshold. Test with SPRT[0,4] as proposed by Marco
16-08-29 sni pawn_weaknesses3 diff
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 28618 W: 5178 L: 5181 D: 18259
sprt @ 10+0.1 th 1 Principle of multiple weaknesses, bonus = 16 * x * x
16-08-28 Elb lmr_hcor diff
LLR: -2.96 (-2.94,2.94) [0.00,4.00]
Total: 41786 W: 7767 L: 7779 D: 26240
sprt @ 10+0.1 th 1 LMR history correction value tweak.
16-08-27 Voy hps diff
LLR: 2.96 (-2.94,2.94) [-3.00,1.00]
Total: 56428 W: 7610 L: 7537 D: 41281
sprt @ 60+0.6 th 1 LTC: A simplification...but am hoping for elo gain.
16-08-29 sni knight_fork diff
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 5720 W: 1003 L: 1104 D: 3613
sprt @ 10+0.1 th 1 Bonus for knight checks which are also forking a piece
16-08-29 aji threat_evasion diff
LLR: -2.96 (-2.94,2.94) [0.00,5.00]
Total: 6228 W: 1109 L: 1208 D: 3911
sprt @ 10+0.1 th 1 don't movecount prune evasions from pawn threats: STC
16-08-29 sni pawn_chains3' diff
LLR: -2.96 (-2.94,2.94) [0.00,5.00]
Total: 13903 W: 2542 L: 2608 D: 8753
sprt @ 10+0.1 th 1 Take 4: like take 2, but with psqt compensation for higher ranks
16-08-29 aji threat_extension diff
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 2001 W: 344 L: 463 D: 1194
sprt @ 10+0.1 th 1 Extend knight and pawn threats that don't have a negative see: STC
16-08-29 Ser threefold_repetition diff
LLR: -2.94 (-2.94,2.94) [0.00,5.00]
Total: 22730 W: 4272 L: 4298 D: 14160
sprt @ 10+0.1 th 1 Threefold repetition detection. Fixes the bug with erroneous 2-fold rep.
16-08-28 SC capture_score diff
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 15441 W: 2950 L: 3008 D: 9483
sprt @ 10+0.1 th 1 A byproduct of these stats experiments
16-08-28 sni windmill diff
LLR: -2.96 (-2.94,2.94) [0.00,5.00]
Total: 19986 W: 3711 L: 3750 D: 12525
sprt @ 10+0.1 th 1 Windmill: take 2, half bonus. Priority -3.
16-08-27 fau FauziPawnsTuning diff
LLR: -2.95 (-2.94,2.94) [0.00,4.00]
Total: 61235 W: 11398 L: 11342 D: 38495
sprt @ 10+0.1 th 1 First test on Pawns Tuning
16-08-28 sg probcut diff
LLR: 2.96 (-2.94,2.94) [0.00,5.00]
Total: 23265 W: 4415 L: 4188 D: 14662
sprt @ 10+0.1 th 1 Tweak probcut threshold
16-08-28 sni windmill diff
LLR: -2.97 (-2.94,2.94) [0.00,5.00]
Total: 13013 W: 2380 L: 2450 D: 8183
sprt @ 10+0.1 th 1 Bonus for taking material by discovered check, like in the windmill combinaison
16-08-28 sg initiative diff
LLR: -2.97 (-2.94,2.94) [0.00,5.00]
Total: 21539 W: 3969 L: 4002 D: 13568
sprt @ 10+0.1 th 1 Try half weight. Take 2
16-08-27 sni pawn_chains3 diff
LLR: -2.94 (-2.94,2.94) [0.00,5.00]
Total: 15274 W: 2827 L: 2886 D: 9561
sprt @ 10+0.1 th 1 Try weight = 2 for the bonus (like in the yellow patch), but with a penalty for the supporting pawn
16-08-27 Voy hps diff
LLR: 2.94 (-2.94,2.94) [-3.00,1.00]
Total: 15852 W: 2910 L: 2781 D: 10161
sprt @ 10+0.1 th 1 A simplification...but am hoping for elo gain.
16-08-27 sni pawn_chains3 diff
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 11610 W: 2180 L: 2255 D: 7175
sprt @ 10+0.1 th 1 Weight = 3 for the bonus
16-08-27 sg initiative diff
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 7566 W: 1361 L: 1454 D: 4751
sprt @ 10+0.1 th 1 initiative penalty if pawns of weaker side defended
16-08-27 IIv deep_search diff
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 6465 W: 1165 L: 1263 D: 4037
sprt @ 10+0.1 th 1 Take 3.
16-08-27 sni queen_threats2 diff
LLR: -0.13 (-2.94,2.94) [0.00,5.00]
Total: 83 W: 13 L: 18 D: 52
sprt @ 10+0.1 th 1 Reintroduce queen threats, when they are part of a double attack
16-08-26 jos probcut_test^ diff
ELO: -5.12 +-2.4 (95%) LOS: 0.0%
Total: 30000 W: 5332 L: 5774 D: 18894
30000 @ 10+0.1 th 1 Probcut at odd plies vs. probcut at even plies. Let's see if there's something to it.
16-08-26 sni pawn_chains2 diff
LLR: -2.96 (-2.94,2.94) [0.00,5.00]
Total: 68288 W: 12847 L: 12675 D: 42766
sprt @ 10+0.1 th 1 Prefer pawn chains towards the center (take2). Submitted with priority -3.
16-08-27 jos probcut_test diff
LLR: -2.96 (-2.94,2.94) [0.00,5.00]
Total: 12399 W: 2290 L: 2362 D: 7747
sprt @ 10+0.1 th 1 Test against master.
16-08-26 sni pawn_chains2 diff
LLR: -2.96 (-2.94,2.94) [0.00,5.00]
Total: 7231 W: 1299 L: 1394 D: 4538
sprt @ 10+0.1 th 1 Prefer pawn chains towards the center (take 3). Submitted with priority -3.
16-08-27 SC updateStatsTweak2 diff
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 14229 W: 2615 L: 2679 D: 8935
sprt @ 10+0.1 th 1 Take 3 and call it a day.
16-08-27 sni pawn_weaknesses2 diff
LLR: -2.96 (-2.94,2.94) [0.00,5.00]
Total: 5877 W: 1059 L: 1160 D: 3658
sprt @ 10+0.1 th 1 Weak pawns in endgame: take 3
16-08-26 luc lower_thread_skips diff
LLR: -0.93 (-2.94,2.94) [0.00,5.00]
Total: 8000 W: 1303 L: 1310 D: 5387
sprt @ 5+0.05 th 7 Helpers skip depths matching those of higher-numbered threads
16-08-26 IIv deep_search diff
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 10531 W: 1877 L: 1957 D: 6697
sprt @ 10+0.1 th 1 As the tuning has its limitations, and the test is pretty neutral at STC, I'll take few more tries here.
16-08-26 luc more_flanks diff
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 12370 W: 2232 L: 2304 D: 7834
sprt @ 10+0.1 th 1 More differentiation in KingFlank for king tropism...
16-08-26 SC ttEval diff
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 10988 W: 1996 L: 2074 D: 6918
sprt @ 10+0.1 th 1 Testing with SPRT since I noticed that bench changes at very high depth probably due to different replacements...
16-08-26 SC updateStatsTweak2 diff
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 27656 W: 5092 L: 5098 D: 17466
sprt @ 10+0.1 th 1 Take 2.
16-08-26 IIv deep_search diff
LLR: -2.96 (-2.94,2.94) [0.00,5.00]
Total: 28147 W: 5200 L: 5204 D: 17743
sprt @ 10+0.1 th 1 Tuned values.
16-08-26 cru evaluate_space_w15 diff
LLR: -2.95 (-2.94,2.94) [0.00,4.00]
Total: 26131 W: 4844 L: 4910 D: 16377
sprt @ 10+0.1 th 1 Continuing to decrease parameter, score for evaluate_space goes up.
16-08-26 sni pawn_chains2 diff
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 17521 W: 3298 L: 3347 D: 10876
sprt @ 10+0.1 th 1 Prefer pawn chains towards the center (take 1)
16-08-26 SC lazyEval diff
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 9368 W: 1721 L: 1806 D: 5841
sprt @ 10+0.1 th 1 "Thou shall not evaluate too much if thou search deep enough." Surprisingly enough was going well on average TC locally.
16-08-26 Elb simple_eg3 diff
LLR: -2.96 (-2.94,2.94) [-3.00,1.00]
Total: 45853 W: 6661 L: 6878 D: 32314
sprt @ 10+0.1 th 1 Take 3 Endgame simplification: retire KBPPKB endgame, let opposite bishop endgames be handled in KBPsK endgame. And in the process, improve evaluations for them. Lowered priority for take 2.
16-08-26 fau FauziTuning1 diff
LLR: -2.95 (-2.94,2.94) [0.00,4.00]
Total: 18093 W: 3361 L: 3455 D: 11277
sprt @ 10+0.1 th 1 First tuning test on a single value change