Stockfish Testing Queue

Pending - 0 tests 0.0 hrs

None

Active - 0 tests

Finished - 1296 tests

19-01-15 pb0 inactiveBishop diff
LLR: -2.95 (-2.94,2.94) [0.50,4.50]
Total: 10987 W: 2363 L: 2473 D: 6151
sprt @ 10+0.1 th 1 Penalty for bishop being inactive due pawn-structure
19-01-03 pb0 swapKillers diff
LLR: -2.94 (-2.94,2.94) [0.50,4.50]
Total: 9057 W: 1901 L: 2020 D: 5136
sprt @ 10+0.1 th 1 Take 2: reset first killer when it gets refuted
19-01-02 pb0 failLowCnt diff
LLR: -2.96 (-2.94,2.94) [0.50,4.50]
Total: 9664 W: 2043 L: 2160 D: 5461
sprt @ 10+0.1 th 1 Yet another shot at this idea
19-01-02 pb0 cm_and_killer diff
LLR: -2.95 (-2.94,2.94) [0.50,4.50]
Total: 15056 W: 3241 L: 3331 D: 8484
sprt @ 10+0.1 th 1 Consider first killer which is identical to countermove
19-01-02 pb0 swapKillers diff
LLR: -2.96 (-2.94,2.94) [0.50,4.50]
Total: 17606 W: 3767 L: 3845 D: 9994
sprt @ 10+0.1 th 1 Swap killers when first killer gets refuted
18-12-21 pb0 killerExchange3 diff
LLR: -2.96 (-2.94,2.94) [0.00,5.00]
Total: 7028 W: 1483 L: 1580 D: 3965
sprt @ 10+0.1 th 1 Killers: take 3
18-12-21 pb0 killerExchange3 diff
LLR: -2.96 (-2.94,2.94) [0.00,5.00]
Total: 23465 W: 5148 L: 5164 D: 13153
sprt @ 10+0.1 th 1 Killers: take 2 (fixed bench)
18-12-20 pb0 killerExchange3 diff
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 34318 W: 7569 L: 7531 D: 19218
sprt @ 10+0.1 th 1 Killers: the good, the bad and better one (hopely now no uninit. access)
18-12-18 pb0 voting_nonresolved diff
LLR: -2.96 (-2.94,2.94) [0.00,5.00]
Total: 49895 W: 9799 L: 9701 D: 30395
sprt @ 5+0.05 th 8 Voting: give less weigth to threads not having resolved fail/low/high
18-12-18 pb0 inactiveQueen diff
LLR: -2.96 (-2.94,2.94) [0.00,5.00]
Total: 7068 W: 1496 L: 1593 D: 3979
sprt @ 10+0.1 th 1 Tired to see Alphazero force SF's Queen to inactive squares.
18-12-17 pb0 penalty diff
LLR: 2.95 (-2.94,2.94) [0.00,5.00]
Total: 35073 W: 5886 L: 5625 D: 23562
sprt @ 60+0.6 th 1 Spec. LTC for this yellow http://tests.stockfishchess.org/tests/view/5c121e730ebc5902ba11aad8 (rebased to current master)
18-12-17 pb0 bonus_one diff
LLR: -2.95 (-2.94,2.94) [0.00,4.00]
Total: 7269 W: 1554 L: 1690 D: 4025
sprt @ 10+0.1 th 1 Same bonus for quiet/capture and continuations histories.
18-12-17 pb0 timePlanning diff
LLR: -2.96 (-2.94,2.94) [0.00,5.00]
Total: 10542 W: 2261 L: 2341 D: 5940
sprt @ 10+0.1 th 1 Assign more thinking time when current PV does simplify a lot. Take 3 (fixed logic)
18-12-17 pb0 timePlanning diff
LLR: -0.33 (-2.94,2.94) [0.00,5.00]
Total: 321 W: 64 L: 77 D: 180
sprt @ 10+0.1 th 1 Assign more thinking time when current PV does simplify a lot. Take 3
18-12-17 pb0 timePlanning diff
LLR: -0.27 (-2.94,2.94) [0.00,5.00]
Total: 385 W: 82 L: 92 D: 211
sprt @ 10+0.1 th 1 Assign more thinking time when current PV does simplify a lot. Take 3
18-12-17 pb0 timePlanning diff
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 29304 W: 6409 L: 6396 D: 16499
sprt @ 10+0.1 th 1 Assign more thinking time when current PV does simplify a lot. Idea: an early on simplified game is likely drawish & terminating with plenty time on the clock (or remaining time in endgame does not help at all to convert into anything).
18-12-17 pb0 timePlanning diff
LLR: -2.96 (-2.94,2.94) [0.00,5.00]
Total: 2263 W: 462 L: 584 D: 1217
sprt @ 10+0.1 th 1 Assign more thinking time when current PV does simplify a lot. Take 2: much larger impact.
18-12-13 pb0 penalty diff
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 72193 W: 15848 L: 15625 D: 40720
sprt @ 10+0.1 th 1 Not sure if this has been tried already: apply same logic to ttHit
18-12-13 pb0 conservative diff
LLR: -2.97 (-2.94,2.94) [0.00,5.00]
Total: 38920 W: 8586 L: 8526 D: 21808
sprt @ 10+0.1 th 1 Prevent a singular event from influencing stats to much (no change sign)
18-12-10 pb0 consecutiveFL diff
LLR: -2.96 (-2.94,2.94) [0.00,5.00]
Total: 14346 W: 3114 L: 3175 D: 8057
sprt @ 10+0.1 th 1 Another shot at this idea
18-12-10 pb0 fastFL diff
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 13884 W: 2996 L: 3059 D: 7829
sprt @ 10+0.1 th 1 Take 3
18-12-10 pb0 contRoot diff
LLR: -2.96 (-2.94,2.94) [0.00,5.00]
Total: 11528 W: 2448 L: 2523 D: 6557
sprt @ 10+0.1 th 1 Does it help to set proper (ss-[1,2])->continuationHistory at root?
18-12-04 pb0 smp_spread0skips diff
ELO: 0.24 +-3.7 (95%) LOS: 55.1%
Total: 10000 W: 1475 L: 1468 D: 7057
10000 @ 20+0.2 th 15 Does it help having few helpers no skipping plies at all? Test with 15 th thereof 3 not skipping. Big skip-sizes always seemed to perform well on very short tc's but scaled rather badly, so testing at smp LTC with fixed num of games.
18-12-06 pb0 fastFL diff
LLR: -2.96 (-2.94,2.94) [0.00,5.00]
Total: 10720 W: 2290 L: 2369 D: 6061
sprt @ 10+0.1 th 1 Fast Fail Low: take 2
18-12-06 pb0 fastFL diff
LLR: -2.96 (-2.94,2.94) [0.00,5.00]
Total: 9476 W: 2018 L: 2103 D: 5355
sprt @ 10+0.1 th 1 Fast Fail Low: forfait from searching moves bejond moveCount 10
18-12-06 pb0 fastFH diff
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 6965 W: 1475 L: 1572 D: 3918
sprt @ 10+0.1 th 1 Take 5
18-12-06 pb0 fastFH diff
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 3143 W: 636 L: 752 D: 1755
sprt @ 10+0.1 th 1 Take 4
18-12-04 pb0 smp_spread0skips diff
ELO: 0.90 +-3.7 (95%) LOS: 68.2%
Total: 10000 W: 1522 L: 1496 D: 6982
10000 @ 20+0.2 th 8 Does it help having few helpers no skipping plies at all? Test also with 8 threads (thereof 2 not skipping) since in the 15 th test there's lot of crashes/noise.
18-12-03 pb0 fastFH diff
LLR: -2.96 (-2.94,2.94) [0.00,5.00]
Total: 1084 W: 192 L: 320 D: 572
sprt @ 10+0.1 th 1 take 3
18-12-03 pb0 fastFH diff
LLR: -2.96 (-2.94,2.94) [0.00,5.00]
Total: 3948 W: 807 L: 919 D: 2222
sprt @ 10+0.1 th 1 Abort search directly on ply 1 when asp. search looks like to fail high
18-12-03 pb0 fastFH diff
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 1908 W: 376 L: 499 D: 1033
sprt @ 10+0.1 th 1 Abort search on ply 1 when asp. search looks like to fh, take 2
18-11-30 pb0 killer6 diff
LLR: -2.96 (-2.94,2.94) [-3.00,1.00]
Total: 7168 W: 1469 L: 1647 D: 4052
sprt @ 10+0.1 th 1 Simplification of recently passed patch. Testing on top of it (fixed bench)
18-11-30 pb0 noDepthRegress diff
LLR: -2.96 (-2.94,2.94) [0.00,5.00]
Total: 8593 W: 1733 L: 1822 D: 5038
sprt @ 5+0.05 th 4 Take 4: regress on TT depth only if bounds or value changes
18-11-29 pb0 noDepthRegress diff
LLR: -2.96 (-2.94,2.94) [0.00,5.00]
Total: 7687 W: 1461 L: 1554 D: 4672
sprt @ 5+0.05 th 8 Take 4
18-11-28 pb0 noDepthRegress diff
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 79463 W: 15567 L: 15336 D: 48560
sprt @ 5+0.05 th 8 Take 3
18-11-28 pb0 noDepthRegress diff
LLR: 2.95 (-2.94,2.94) [-3.00,1.00]
Total: 106846 W: 23310 L: 23356 D: 60180
sprt @ 10+0.1 th 1 Take 3, test for non-regression at single thread
18-11-28 pb0 noDepthRegress diff
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 1077 W: 159 L: 282 D: 636
sprt @ 5+0.05 th 8 Avoid regressing in TT-depth, fixed logic (smp-test)
18-11-28 pb0 noDepthRegress diff
LLR: -2.94 (-2.94,2.94) [0.00,5.00]
Total: 307 W: 15 L: 137 D: 155
sprt @ 5+0.05 th 8 Avoid regressing in TT-depth (smp)
18-11-27 pb0 cacheAttacks2 diff
LLR: -2.96 (-2.94,2.94) [0.00,5.00]
Total: 16754 W: 3701 L: 3750 D: 9303
sprt @ 10+0.1 th 1 Cache attacks in see for consecutive calls with the same move
18-11-26 pb0 asp_win_after_skip diff
LLR: -2.96 (-2.94,2.94) [0.00,5.00]
Total: 23098 W: 4529 L: 4552 D: 14017
sprt @ 5+0.05 th 8 SMP helpers: use larger aspiration windows after skipping, take 3
18-11-23 pb0 asp_win_after_skip diff
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 10669 W: 2010 L: 2089 D: 6570
sprt @ 5+0.05 th 8 SMP helpers: use larger aspiration windows when skipping (fixed logic)
18-11-23 pb0 asp_win_after_skip diff
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 11310 W: 2230 L: 2306 D: 6774
sprt @ 5+0.05 th 8 SMP helpers: use larger aspiration windows when skipping, take 2 (fix)
18-11-23 pb0 ideep_mainthread diff
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 33909 W: 7543 L: 7506 D: 18860
sprt @ 10+0.1 th 1 Skip odd rootDepths on mainthread if we have enough time at disposal
18-11-23 pb0 asp_win_after_skip diff
LLR: -0.17 (-2.94,2.94) [0.00,5.00]
Total: 40 W: 4 L: 11 D: 25
sprt @ 5+0.05 th 8 SMP helper threads: use larger aspiration windows when skipping plies
18-11-22 pb0 threatByKing diff
LLR: -0.41 (-2.94,2.94) [0.00,5.00]
Total: 1088 W: 228 L: 241 D: 619
sprt @ 10+0.1 th 1 As suggested by Brian: don't assign ThreatByKing bonus if king would go into check.
18-11-21 pb0 discoInQS diff
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 54118 W: 9052 L: 8959 D: 36107
sprt @ 60+0.6 th 1 LTC: Speculative LTC for http://tests.stockfishchess.org/tests/view/5bf2b4570ebc5902bbbde0b8
18-11-21 pb0 discoInQS diff
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 87241 W: 19307 L: 19008 D: 48926
sprt @ 10+0.1 th 1 Don't prune moves discovering an attack to enemy king or queen
18-11-21 pb0 discoInQS diff
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 10990 W: 2400 L: 2477 D: 6113
sprt @ 10+0.1 th 1 take 3: centralize logic and use it for shallow pruning too
18-11-21 pb0 discoInQS diff
LLR: -2.94 (-2.94,2.94) [0.00,5.00]
Total: 13380 W: 2933 L: 2998 D: 7449
sprt @ 10+0.1 th 1 Don't prune moves discovering an attack to enemy king or queen, take 2
18-11-19 pb0 discoInQS diff
LLR: -2.96 (-2.94,2.94) [0.00,5.00]
Total: 118270 W: 26226 L: 25774 D: 66270
sprt @ 10+0.1 th 1 don't prune discovering check moves with negavive see-value in qsearch