Stockfish Testing Queue

Pending - 1 tests 1.0 hrs

05-12-17 pb queenBeforeStanding diff
LLR: -0.95 (-2.94,2.94) [0.00,5.00]
Total: 9000 W: 1618 L: 1620 D: 5762
sprt @ 10+0.1 th 1 Kingdanger: no Rook/Bishop-Check when square is defended by our Queen (see commit notes). Test against passed patch.

Active - 1 tests

15-12-17 pb flatten_out2 diff
ELO: 0.04 +-3.4 (95%) LOS: 50.8%
Total: 9776 W: 1195 L: 1194 D: 7387
20000 @ 60+0.6 th 1 Avoid fast saturation (clamping) by reducing bonus for higher depths. LTC Experiment at low throughput.

Finished - 621 tests

13-12-17 pb pinned_passeds diff
LLR: -2.96 (-2.94,2.94) [0.00,5.00]
Total: 30690 W: 5544 L: 5539 D: 19607
sprt @ 10+0.1 th 1 no bonus to pinned passed pawns
12-12-17 pb enhanced_verify diff
LLR: 2.96 (-2.94,2.94) [-3.00,1.00]
Total: 41899 W: 5359 L: 5265 D: 31275
sprt @ 60+0.6 th 1 LTC: Enhance verification search for better detecting of zugzwang. [-3;1] bounds, see commit notes.
12-12-17 pb enhanced_verify diff
LLR: 2.96 (-2.94,2.94) [-3.00,1.00]
Total: 18220 W: 3379 L: 3253 D: 11588
sprt @ 10+0.1 th 1 STC: Enhance verification search for better detecting of zugzwang. [-3;1] bounds, see commit notes.
11-12-17 pb rook_pin diff
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 18107 W: 3356 L: 3403 D: 11348
sprt @ 10+0.1 th 1 Allow bishop check attacks through opponent rooks. Similar to http://tests.stockfishchess.org/tests/view/5a26c78d0ebc590ccbb8b866 but little more accurate. Test against passed Simplify other checks
07-12-17 pb kingQueenLine2 diff
LLR: -2.96 (-2.94,2.94) [0.00,5.00]
Total: 3668 W: 615 L: 725 D: 2328
sprt @ 10+0.1 th 1 Take 2 for this idea
07-12-17 pb pureBRattacks diff
LLR: -2.96 (-2.94,2.94) [0.00,5.00]
Total: 16384 W: 2891 L: 2947 D: 10546
sprt @ 10+0.1 th 1 Take 2
07-12-17 pb proximity diff
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 9837 W: 1739 L: 1822 D: 6276
sprt @ 10+0.1 th 1 More Kingdanger when Queen threatens with check in proximity of the king
07-12-17 pb kingdangerOther diff
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 10379 W: 1850 L: 1931 D: 6598
sprt @ 10+0.1 th 1 Take 2
06-12-17 pb kingdangerOther diff
LLR: -2.96 (-2.94,2.94) [0.00,5.00]
Total: 38286 W: 7023 L: 6984 D: 24279
sprt @ 10+0.1 th 1 Kingdanger tweak. Test against passed patch. (fixed base bench)
06-12-17 pb pureBRattacks diff
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 16801 W: 3060 L: 3113 D: 10628
sprt @ 10+0.1 th 1 Pure attacks (without x-rays) for Bishop and Rook, does this better? Test against passed patch.
05-12-17 pb queenCheckOnce diff
LLR: -3.56 (-2.94,2.94) [0.00,5.00]
Total: 53438 W: 9654 L: 9578 D: 34206
sprt @ 10+0.1 th 1 Assign QueenCheck once per checkSquare. Test against passed patch.
04-12-17 pb queenBetween diff
LLR: 2.96 (-2.94,2.94) [0.00,5.00]
Total: 63448 W: 8287 L: 7965 D: 47196
sprt @ 60+0.6 th 1 LTC: Don't consider queen as check-blocker
04-12-17 pb kingQueenLine2 diff
LLR: -2.96 (-2.94,2.94) [0.00,5.00]
Total: 14266 W: 2505 L: 2570 D: 9191
sprt @ 10+0.1 th 1 Additional penalty if Queen is in check-line with no pieces between her and King
04-12-17 pb queenBetween diff
LLR: 2.95 (-2.94,2.94) [0.00,5.00]
Total: 22636 W: 4212 L: 3990 D: 14434
sprt @ 10+0.1 th 1 Don't consider queen as check-blocker
30-11-17 pb safe_attacks_king diff
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 11111 W: 1961 L: 2039 D: 7111
sprt @ 10+0.1 th 1 Take 3. Test against passed safe_attacks2 patch.
30-11-17 pb safe_attacks_king diff
LLR: -2.97 (-2.94,2.94) [0.00,5.00]
Total: 20129 W: 3619 L: 3659 D: 12851
sprt @ 10+0.1 th 1 Take 2. Test against passed safe_attacks2 patch.
30-11-17 pb safe_attacks_king diff
LLR: -2.96 (-2.94,2.94) [0.00,5.00]
Total: 9761 W: 1671 L: 1755 D: 6335
sprt @ 10+0.1 th 1 Additional bonus if opponent king stays on same line. Test against passed safe_attacks2 patch.
30-11-17 pb safe_attacks_noxray diff
LLR: -2.97 (-2.94,2.94) [0.00,5.00]
Total: 14342 W: 2526 L: 2591 D: 9225
sprt @ 10+0.1 th 1 Take 2: avoid some unnecessary computations
29-11-17 pb safe_attacks_noxray diff
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 16795 W: 2960 L: 3014 D: 10821
sprt @ 10+0.1 th 1 Don't give bonus if queen must enter the crucial square first (with rook or bishop behind) where opponent queen can force the exchange. Test against passed safe_attacks2 patch.
24-11-17 pb probCutDepth diff
LLR: -2.96 (-2.94,2.94) [0.00,5.00]
Total: 29402 W: 3670 L: 3695 D: 22037
sprt @ 60+0.6 th 1 LTC Take 3: Reduce depth when capture returns an overwhelming good see-value.
24-11-17 pb qsearch_capt_stat diff
LLR: -2.96 (-2.94,2.94) [0.00,5.00]
Total: 44676 W: 8106 L: 8041 D: 28529
sprt @ 10+0.1 th 1 Take 5: assert cut's with see_ge and let QRECAPTURES stage unchanged (to few kick-in's)
24-11-17 pb lmr_prune_tips diff
LLR: -4.53 (-2.94,2.94) [0.00,5.00]
Total: 25852 W: 4626 L: 4708 D: 16518
sprt @ 10+0.1 th 1 Lmr-prune more towards the leaves
24-11-17 pb probCutDepth2 diff
LLR: -2.97 (-2.94,2.94) [0.00,5.00]
Total: 17182 W: 3042 L: 3095 D: 11045
sprt @ 10+0.1 th 1 While LTC is running, test if applying the same idea also on lower depths is stronger
24-11-17 pb probCutDepth diff
LLR: 2.97 (-2.94,2.94) [0.00,5.00]
Total: 10906 W: 2051 L: 1879 D: 6976
sprt @ 10+0.1 th 1 Take 3: Reduce depth when capture returns an overwhelming good see-value. Local test Score:983 - 931 - 3502 [0.505] 5416, Elo +3.34
17-11-17 pb helper_reduct_on_tips diff
LLR: -2.96 (-2.94,2.94) [0.00,5.00]
Total: 81053 W: 14702 L: 14483 D: 51868
sprt @ 5+0.05 th 3 Inspired by the idea of SC to assist helpers to finish faster their ID-cycle by pruning more aggressively: apply it near the leaves where probably searching new territory, so avoid the risk to override more valuable tt-entries.
21-11-17 pb qsearch_capt_stat diff
LLR: -2.94 (-2.94,2.94) [0.00,5.00]
Total: 100447 W: 18353 L: 18049 D: 64045
sprt @ 10+0.1 th 1 Take 2, fixed logic: return at least the first capture
23-11-17 pb qsearch_capt_stat diff
LLR: -2.96 (-2.94,2.94) [0.00,5.00]
Total: 13062 W: 2287 L: 2357 D: 8418
sprt @ 10+0.1 th 1 Take 4
23-11-17 pb probCutDepth diff
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 22270 W: 3949 L: 3980 D: 14341
sprt @ 10+0.1 th 1 Take 2
22-11-17 pb probCutDepth diff
LLR: -2.96 (-2.94,2.94) [0.00,5.00]
Total: 6911 W: 1222 L: 1318 D: 4371
sprt @ 10+0.1 th 1 Probcut depth tweak
21-11-17 pb qsearch_capt_stat diff
LLR: -2.96 (-2.94,2.94) [0.00,5.00]
Total: 22292 W: 3988 L: 4019 D: 14285
sprt @ 10+0.1 th 1 Take 3
21-11-17 pb qsearch_capt_stat diff
LLR: -0.06 (-2.94,2.94) [0.00,5.00]
Total: 373 W: 63 L: 64 D: 246
sprt @ 10+0.1 th 1 qsearch: Don't return captures with very bad history scores
21-11-17 pb probcut_threshold diff
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 13617 W: 2468 L: 2535 D: 8614
sprt @ 10+0.1 th 1 Apply good-bad capture logic on probcut too
21-11-17 pb probcut_ttsave diff
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 11783 W: 2107 L: 2182 D: 7494
sprt @ 10+0.1 th 1 Save pos in TT on probcut
20-11-17 pb helper_reduct_on_tips diff
LLR: -2.97 (-2.94,2.94) [0.00,5.00]
Total: 34267 W: 6116 L: 6097 D: 22054
sprt @ 5+0.05 th 3 Take 2: with Vondele's recommendations and a little bit more aggressive in pruning
20-11-17 pb bestCapture diff
LLR: -2.96 (-2.94,2.94) [0.00,5.00]
Total: 21131 W: 3765 L: 3801 D: 13565
sprt @ 10+0.1 th 1 Take 2 (fixed logic)
20-11-17 pb bestCapture diff
LLR: -2.14 (-2.94,2.94) [0.00,5.00]
Total: 7168 W: 1282 L: 1342 D: 4544
sprt @ 10+0.1 th 1 Inspired by Vondele's timeCapture series: use logic at low depth's for applying it at lmr
14-11-17 pb counterCapture diff
LLR: -2.96 (-2.94,2.94) [0.00,5.00]
Total: 30025 W: 5455 L: 5452 D: 19118
sprt @ 10+0.1 th 1 Take 6: use it for quiet continuations too
14-11-17 pb counterCapture diff
LLR: -2.96 (-2.94,2.94) [0.00,5.00]
Total: 25712 W: 4629 L: 4645 D: 16438
sprt @ 10+0.1 th 1 Take 7
13-11-17 pb counterCapture diff
LLR: -2.96 (-2.94,2.94) [0.00,5.00]
Total: 4496 W: 739 L: 845 D: 2912
sprt @ 10+0.1 th 1 Take 5
13-11-17 pb counterCapture diff
LLR: -2.97 (-2.94,2.94) [0.00,5.00]
Total: 25321 W: 4556 L: 4574 D: 16191
sprt @ 10+0.1 th 1 Take 4
13-11-17 pb counterCapture diff
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 12994 W: 2281 L: 2351 D: 8362
sprt @ 10+0.1 th 1 Take 3
10-11-17 pb captKillerStat diff
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 39865 W: 7229 L: 7184 D: 25452
sprt @ 10+0.1 th 1 Capture killer idea with capture stats to assist
10-11-17 pb captKillerStat2 diff
LLR: -2.97 (-2.94,2.94) [0.00,5.00]
Total: 13600 W: 2403 L: 2471 D: 8726
sprt @ 10+0.1 th 1 Take 2: 2 killers
10-11-17 pb counterCapture diff
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 11546 W: 2037 L: 2113 D: 7396
sprt @ 10+0.1 th 1 Take 2: Bugfix (missed ttMove check, so ttMove was returned twice)
10-11-17 pb counterCapture diff
LLR: -2.96 (-2.94,2.94) [0.00,5.00]
Total: 11430 W: 2015 L: 2092 D: 7323
sprt @ 10+0.1 th 1 Introduce capture counter moves. Saves some move generation. Take 1: Fill counter-capture only when different 'to' square
10-11-17 pb probCutStat diff
LLR: -2.97 (-2.94,2.94) [0.00,5.00]
Total: 12192 W: 2138 L: 2212 D: 7842
sprt @ 10+0.1 th 1 Take 2
10-11-17 pb probCutStat diff
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 5432 W: 956 L: 1058 D: 3418
sprt @ 10+0.1 th 1 Use capture-stat to assist probcut
09-11-17 pb stats_experiment diff
LLR: -3.31 (-2.94,2.94) [0.00,5.00]
Total: 15525 W: 2793 L: 2867 D: 9865
sprt @ 10+0.1 th 1 A further 'empty-framework'-experiment.
09-11-17 pb capture_history_tweak2 diff
LLR: -2.94 (-2.94,2.94) [0.00,5.00]
Total: 18471 W: 3332 L: 3378 D: 11761
sprt @ 10+0.1 th 1 Retry a slight modified version for this strong yellow http://tests.stockfishchess.org/tests/view/59fb99810ebc590ccbb8a07a from Stefan (empty framework)
09-11-17 pb tweak_killers diff
LLR: -2.96 (-2.94,2.94) [0.00,5.00]
Total: 10493 W: 1802 L: 1883 D: 6808
sprt @ 10+0.1 th 1 Skip second killer only if we have a tt move (bugfixed)