Stockfish Testing Queue

Finished - 27341 tests

12-04-16 Vo CA diff
LLR: -2.97 (-2.94,2.94) [0.00,5.00]
Total: 8172 W: 1436 L: 1527 D: 5209
sprt @ 10+0.1 th 1 Take 2. Add a weight for from_sq.
10-04-16 Vo fmh_leaf diff
LLR: -2.96 (-2.94,2.94) [0.00,5.00]
Total: 109204 W: 19978 L: 19637 D: 69589
sprt @ 10+0.1 th 1 Don't update fmh at the leaves.
12-04-16 pb dccandidates diff
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 32771 W: 5885 L: 5871 D: 21015
sprt @ 10+0.1 th 1 move-order: score discovering check capture as good capture with large bonus
12-04-16 My ho diff
LLR: -2.96 (-2.94,2.94) [0.00,5.00]
Total: 3089 W: 523 L: 636 D: 1930
sprt @ 10+0.1 th 1 Experiment with history in capture ordering
10-04-16 El fmh_rerun diff
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 37296 W: 4870 L: 4866 D: 27560
sprt @ 60+0.6 th 1 This is a rerun of http://tests.stockfishchess.org/tests/view/56f062b80ebc59301a353f2e which ended yellow, but was run with wrong hash value. Just to be sure, now try with correct hash.
12-04-16 sg cmh2 diff
LLR: -2.96 (-2.94,2.94) [0.00,5.00]
Total: 6959 W: 1189 L: 1285 D: 4485
sprt @ 10+0.1 th 1 Try to extend cmh to second last opponent move.
12-04-16 sg fmh2 diff
LLR: -2.96 (-2.94,2.94) [0.00,5.00]
Total: 9915 W: 1815 L: 1898 D: 6202
sprt @ 10+0.1 th 1 Try to extend fmh to second last own move.
12-04-16 Vo CA2 diff
LLR: -2.96 (-2.94,2.94) [0.00,5.00]
Total: 9001 W: 1615 L: 1702 D: 5684
sprt @ 10+0.1 th 1 -Operation Green Patch- Sort by: MVV, LVA, Victim's Rank, MovePiece Rank
11-04-16 Vo CA diff
LLR: -2.97 (-2.94,2.94) [0.00,5.00]
Total: 44781 W: 8189 L: 8123 D: 28469
sprt @ 10+0.1 th 1 Use rank of from_sq to help sort captures.
11-04-16 th loose_values diff
LLR: -2.96 (-2.94,2.94) [0.00,4.00]
Total: 17195 W: 3093 L: 3191 D: 10911
sprt @ 10+0.1 th 1 Take 1 with SPSA output.
11-04-16 sn pawn_control3 diff
LLR: -2.96 (-2.94,2.94) [0.00,5.00]
Total: 22090 W: 4003 L: 4034 D: 14053
sprt @ 10+0.1 th 1 Pawn control: take 2
08-04-16 jk master diff
ELO: 15.27 +-1.5 (95%) LOS: 100.0%
Total: 40000 W: 5008 L: 3251 D: 31741
40000 @ 60+0.6 th 1 Regression test (1cbba8d6fadaae2c)
11-04-16 pb capture_killer diff
LLR: -1.33 (-2.94,2.94) [0.00,5.00]
Total: 6128 W: 1132 L: 1162 D: 3834
sprt @ 10+0.1 th 1 Like killer-move heuristic for quite movesintroduce a capture/recapture heuristic
10-04-16 th loose_tune2 diff
9933/10000 iterations
20000/20000 games played
20000 @ 20+0.2 th 1 Take 2 with higher ck. Low throughput and I will stop if values are within +-0.5 after 5000 games.
10-04-16 My EO diff
LLR: -2.97 (-2.94,2.94) [0.00,5.00]
Total: 9352 W: 1635 L: 1721 D: 5996
sprt @ 10+0.1 th 1 Take 2 VoyagerOne's sign suggestion
10-04-16 Ro PassedPawnMb diff
LLR: -2.96 (-2.94,2.94) [0.00,4.00]
Total: 8342 W: 1495 L: 1624 D: 5223
sprt @ 10+0.1 th 1 Test values from a SPSA tuning session by mbootsector.
08-04-16 Vo QuickSEE diff
LLR: -2.97 (-2.94,2.94) [0.00,5.00]
Total: 27294 W: 4890 L: 4900 D: 17504
sprt @ 10+0.1 th 1 If we are looking for the sign of see and if the queen captures a piece < QueenValue and then gets captured by pawn or knight. Then the see_sign should be negative. The idea is to exit the see algorithm earlier. Interestingly there is a 11 node bench difference...
10-04-16 sn pawn_control3 diff
LLR: -2.96 (-2.94,2.94) [0.00,5.00]
Total: 25113 W: 4579 L: 4597 D: 15937
sprt @ 10+0.1 th 1 Pawn control
10-04-16 sg king_safety diff
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 15370 W: 2808 L: 2867 D: 9695
sprt @ 10+0.1 th 1 Tweak undefended penalty
10-04-16 My sft diff
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 9165 W: 1610 L: 1696 D: 5859
sprt @ 10+0.1 th 1 Revisit double opposite bishop scaling
10-04-16 jo QueenOn7th diff
LLR: -2.95 (-2.94,2.94) [0.00,4.00]
Total: 11916 W: 2128 L: 2244 D: 7544
sprt @ 10+0.1 th 1 Take 2.
10-04-16 My EO diff
LLR: -2.97 (-2.94,2.94) [0.00,5.00]
Total: 5175 W: 902 L: 1006 D: 3267
sprt @ 10+0.1 th 1 A different tweak on capture ordering.
10-04-16 jo passed3 diff
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 4946 W: 852 L: 956 D: 3138
sprt @ 10+0.1 th 1 Retest an old patch.
09-04-16 Vo Attack diff
LLR: -2.96 (-2.94,2.94) [0.00,5.00]
Total: 27956 W: 3762 L: 3787 D: 20407
sprt @ 60+0.6 th 1 See how this yellow patch will fare at LTC. Low Throughput.
10-04-16 sn queen_approach diff
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 1968 W: 322 L: 440 D: 1206
sprt @ 10+0.1 th 1 Count not only the queen contact checks, but also the queen approaching moves in the larger king ring.
08-04-16 jo QueenOn7th diff
LLR: -2.95 (-2.94,2.94) [0.00,4.00]
Total: 110078 W: 20237 L: 20015 D: 69826
sprt @ 10+0.1 th 1 Apply a small midgame bonus for a queen on the 7th rank.
08-04-16 El attack_unit diff
LLR: -2.96 (-2.94,2.94) [0.00,4.00]
Total: 95366 W: 17458 L: 17288 D: 60620
sprt @ 10+0.1 th 1 Try test for attack_units with info from tuning session.
10-04-16 Ro UndefKingRing2b diff
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 13818 W: 2510 L: 2576 D: 8732
sprt @ 10+0.1 th 1 Can we gain strength by looking at the remaining squares at distance 2 (test with normal [0,5])
10-04-16 fb AlignedPawns02 diff
LLR: -2.95 (-2.94,2.94) [0.00,4.00]
Total: 11330 W: 2022 L: 2140 D: 7168
sprt @ 10+0.1 th 1 Changing from Rank 5 to Rank 3
09-04-16 Ro UndefKingRing2 diff
LLR: 2.96 (-2.94,2.94) [-3.00,1.00]
Total: 61817 W: 11352 L: 11301 D: 39164
sprt @ 10+0.1 th 1 Simplification experiment.
09-04-16 Vo ToSqStats diff
LLR: -2.94 (-2.94,2.94) [0.00,5.00]
Total: 15701 W: 2804 L: 2862 D: 10035
sprt @ 10+0.1 th 1 New stat table idea: [Color][tosq][prevSq] First attempt (bonus might be a bit too high)
09-04-16 Vo Attack2 diff
LLR: -2.96 (-2.94,2.94) [0.00,5.00]
Total: 9761 W: 1697 L: 1781 D: 6283
sprt @ 10+0.1 th 1 It seems like this idea may have some merit...so I don't want to throw in the towel just yet.
09-04-16 Vo Attack diff
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 56608 W: 10414 L: 10296 D: 35898
sprt @ 10+0.1 th 1 It might help if i didn't comment out my code...
09-04-16 th loose_tune diff
9937/10000 iterations
20000/20000 games played
20000 @ 20+0.2 th 1 Tuning the recent LooseEnemies parameter.
09-04-16 Ha pinned diff
LLR: -2.95 (-2.94,2.94) [0.00,4.00]
Total: 23167 W: 4175 L: 4252 D: 14740
sprt @ 10+0.1 th 1 Give attack units bonus for each pinned piece.
09-04-16 El hpb_tweak diff
LLR: -2.96 (-2.94,2.94) [0.00,5.00]
Total: 24962 W: 4521 L: 4540 D: 15901
sprt @ 10+0.1 th 1 Take 3: History based pruning tweak
09-04-16 sn king_ring7 diff
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 14147 W: 2521 L: 2586 D: 9040
sprt @ 10+0.1 th 1 Use larger king ring for counting undefended squares (take 2, weight=10)
08-04-16 My pp diff
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 34663 W: 6444 L: 6419 D: 21800
sprt @ 10+0.1 th 1 Following recent simplification, try a variant of a patch by jhellis on pawn promotion with mating material.
08-04-16 la pp_tune diff
LLR: -2.95 (-2.94,2.94) [0.00,4.00]
Total: 39658 W: 7357 L: 7376 D: 24925
sprt @ 10+0.1 th 1 Tuned locally with a bespoke tuner. Stop criterion: loss function < 0.08 (2442 iter / 9768 games). Tested for 1000 games (10+0.1 s): +230 -183 =587
08-04-16 sn king_ring7 diff
LLR: -2.96 (-2.94,2.94) [0.00,5.00]
Total: 23840 W: 4361 L: 4384 D: 15095
sprt @ 10+0.1 th 1 Use larger king ring for counting undefended squares
08-04-16 Vo Attack diff
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 5468 W: 972 L: 1074 D: 3422
sprt @ 10+0.1 th 1 Final try...
08-04-16 Vo SeeTable diff
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 22722 W: 4112 L: 4140 D: 14470
sprt @ 10+0.1 th 1 One more shot at this.
08-04-16 My RA diff
LLR: -2.96 (-2.94,2.94) [0.00,5.00]
Total: 3320 W: 569 L: 681 D: 2070
sprt @ 10+0.1 th 1 Penalise rooks when other pieces of our own are in the way.
08-04-16 El hpb_tweak diff
LLR: -2.96 (-2.94,2.94) [0.00,5.00]
Total: 23163 W: 4157 L: 4184 D: 14822
sprt @ 10+0.1 th 1 Take 2: History based pruning tweak
08-04-16 Ro ImprovingMinor diff
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 9358 W: 1681 L: 1766 D: 5911
sprt @ 10+0.1 th 1 Take 1
08-04-16 My MO3 diff
LLR: -2.96 (-2.94,2.94) [0.00,5.00]
Total: 38400 W: 7021 L: 6982 D: 24397
sprt @ 10+0.1 th 1 Take 2 before I maybe move onto something slightly different
08-04-16 El hpb_tweak diff
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 18551 W: 3329 L: 3375 D: 11847
sprt @ 10+0.1 th 1 History based pruning tweak
08-04-16 My MO4 diff
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 13182 W: 2363 L: 2432 D: 8387
sprt @ 10+0.1 th 1 Last try - similar to VoyagerOne targeting capture order, but with smaller bonus and inverted sign which seems most promising from runs so far.
06-04-16 Ha passed diff
LLR: 2.96 (-2.94,2.94) [-3.00,1.00]
Total: 67213 W: 9135 L: 9077 D: 49001
sprt @ 60+0.6 th 1 LTC: Small passed pawn simplification.
08-04-16 pb see_pin_aware_both_side diff
LLR: -2.96 (-2.94,2.94) [0.00,5.00]
Total: 10889 W: 1944 L: 2023 D: 6922
sprt @ 10+0.1 th 1 SEE aware of pins, take 2 with bugfix and also including pins of the opposite side if the king hasn't moved.