Stockfish Testing Queue

Pending - 0 tests 0.0 hrs

None

Active - 0 tests

Finished - 1332 tests

24-02-18 Ro UnsafeBlocker diff
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 18322 W: 3967 L: 4008 D: 10347
sprt @ 10+0.1 th 1 Adding a few more unsafechecks For example Rg1 Ng4 against h5 Kh8 Ng4 is not in the mobilityArea[WHITE] but should be considered as an unsafecheck. Was already tested here http://tests.stockfishchess.org/tests/view/5a3940ca0ebc590ccbb8bf36 The difference is that I don't allow a bonus for pawn blockers such as White Rg1 g2 against Kg8
21-02-18 Ro BishopPairMobility diff
LLR: -2.96 (-2.94,2.94) [0.00,5.00]
Total: 11474 W: 2528 L: 2603 D: 6343
sprt @ 10+0.1 th 1 Take 2: smaller penalty and bonus
20-02-18 Ro BishopPairMobility diff
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 8801 W: 1896 L: 1984 D: 4921
sprt @ 10+0.1 th 1 Bishop pair value correction based on duo mobiity (fied bench)
17-02-18 Ro EffectiveMobility diff
LLR: -0.71 (-2.94,2.94) [0.00,5.00]
Total: 831 W: 159 L: 186 D: 486
sprt @ 10+0.1 th 1 A new mobility experiment
16-02-18 Ro RemoveDefended diff
LLR: -2.95 (-2.94,2.94) [-3.00,1.00]
Total: 17411 W: 3844 L: 4045 D: 9522
sprt @ 10+0.1 th 1 Simplify defended by always including majors
16-02-18 Ro TallPawn diff
LLR: -2.96 (-2.94,2.94) [0.00,5.00]
Total: 21985 W: 4849 L: 4872 D: 12264
sprt @ 10+0.1 th 1 Take 2, lower penalty
16-02-18 Ro TallPawn diff
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 17084 W: 3748 L: 3795 D: 9541
sprt @ 10+0.1 th 1 (Another) bishop experiment
13-02-18 Ro OppBThreats diff
LLR: -2.96 (-2.94,2.94) [0.00,5.00]
Total: 25889 W: 5570 L: 5575 D: 14744
sprt @ 10+0.1 th 1 Bonus for bishop attacks not defended by bishop
13-02-18 Ro BlockedPawns diff
LLR: -2.94 (-2.94,2.94) [0.00,5.00]
Total: 9545 W: 2047 L: 2131 D: 5367
sprt @ 10+0.1 th 1 Consider a storming pawn "blocked" if faced by a square occupied by our pawn (even if not backmost) or defended twice by pawns.
03-02-18 Ro WeakPawnsMajor diff
LLR: -2.95 (-2.94,2.94) [-3.00,1.00]
Total: 79276 W: 17384 L: 17706 D: 44186
sprt @ 10+0.1 th 1 Move the weak unopposed penalty back to pawns.cpp
10-02-18 Ro TopPassers2 diff
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 10168 W: 2213 L: 2294 D: 5661
sprt @ 10+0.1 th 1
10-02-18 Ro PPKD1 diff
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 4301 W: 926 L: 1037 D: 2338
sprt @ 10+0.1 th 1 How bad is this ?
09-02-18 Ro WeakPP diff
LLR: -2.95 (-2.94,2.94) [0.00,4.00]
Total: 18046 W: 3994 L: 4087 D: 9965
sprt @ 10+0.1 th 1 pp safety is currently only looking at the squares to queen. Include also the pawn itself when testing for the pp safety.
04-02-18 Ro BPReworkTuned1 diff
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 45181 W: 10153 L: 10059 D: 24969
sprt @ 10+0.1 th 1 Testing the tuned values (with some simplifications)
03-02-18 Ro BPRework diff
44099/50000 iterations
94031/100000 games played
100000 @ 20+0.2 th 1 Fixed compile error. Bishop experiment. Let the imbalance bishop/pawns be 0, (since it is color blind) and tune instead the pawn count by color in evaluate.cpp. Low thoughput
30-01-18 Ro ttuned diff
LLR: -2.95 (-2.94,2.94) [0.00,4.00]
Total: 94900 W: 16460 L: 16303 D: 62137
sprt @ 60+0.6 th 1 LTC
30-01-18 Ro HalfLeverCount1 diff
LLR: -2.96 (-2.94,2.94) [0.00,5.00]
Total: 36402 W: 8109 L: 8060 D: 20233
sprt @ 10+0.1 th 1 Try the opposite
30-01-18 Ro HalfLeverCount1 diff
LLR: -2.96 (-2.94,2.94) [0.00,5.00]
Total: 8054 W: 1734 L: 1826 D: 4494
sprt @ 10+0.1 th 1 Assume levers have a 50% chance to be exchanged. so adjust the pawn count estimate in evaluate space or in evaluate initiative accordingly
30-01-18 Ro ttuned diff
LLR: 2.96 (-2.94,2.94) [0.00,4.00]
Total: 39229 W: 8802 L: 8482 D: 21945
sprt @ 10+0.1 th 1
28-01-18 Ro QTune diff
28483/30000 iterations
60000/60000 games played
60000 @ 20+0.2 th 1 Tuning threats, priority -1 if fishtest is empty
30-01-18 Ro InitiativeLever diff
LLR: -2.96 (-2.94,2.94) [0.00,5.00]
Total: 6342 W: 1380 L: 1481 D: 3481
sprt @ 10+0.1 th 1 Take 3: with factor 4
29-01-18 Ro InitiativeLever diff
LLR: -2.96 (-2.94,2.94) [0.00,5.00]
Total: 13525 W: 2982 L: 3047 D: 7496
sprt @ 10+0.1 th 1 Take 2: normalise for rank_3 and double the weight bench: 4960171 master: 5059457
29-01-18 Ro InitiativeLever diff
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 6336 W: 1329 L: 1429 D: 3578
sprt @ 10+0.1 th 1 Take 1: levers in initiative
29-01-18 Ro PPTuned diff
LLR: -2.95 (-2.94,2.94) [0.00,4.00]
Total: 61110 W: 13664 L: 13585 D: 33861
sprt @ 10+0.1 th 1 From spsa tuning "as is"
28-01-18 Ro DistanceFactor2a diff
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 31903 W: 7108 L: 7081 D: 17714
sprt @ 10+0.1 th 1 Adjust distance by 2 if defender king is chasing from behind bench 5103871
28-01-18 Ro QTune diff
27608/30000 iterations
59468/60000 games played
60000 @ 20+0.2 th 1 Tuning passed pawns, low thoughput
28-01-18 Ro DistanceFactor2a diff
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 50618 W: 11368 L: 11247 D: 28003
sprt @ 10+0.1 th 1 increase distance of defender king if chasing from behind
28-01-18 Ro DistanceFactor diff
LLR: -2.38 (-2.94,2.94) [0.00,5.00]
Total: 2370 W: 488 L: 583 D: 1299
sprt @ 10+0.1 th 1 Try some more distance heuristics
28-01-18 Ro RealKingDistance diff
LLR: -2.96 (-2.94,2.94) [0.00,5.00]
Total: 38817 W: 8565 L: 8505 D: 21747
sprt @ 10+0.1 th 1 take 1b fixed
27-01-18 Ro RealKingDistance diff
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 7548 W: 1617 L: 1711 D: 4220
sprt @ 10+0.1 th 1 Increase "distance" by one if king cannot come any closer
27-01-18 Ro PPDTuned2 diff
LLR: -2.96 (-2.94,2.94) [0.00,5.00]
Total: 22820 W: 5043 L: 5062 D: 12715
sprt @ 10+0.1 th 1 Making sure we can't lower the new bound
27-01-18 Ro PPDTuned diff
LLR: 3.31 (-2.94,2.94) [0.00,5.00]
Total: 23987 W: 5550 L: 5281 D: 13156
sprt @ 10+0.1 th 1 Simpler version where we limit the king distance to 5 and we introduce an array for the rank factor. (such that the rank factor for a pawn on rank 7 is smaller than master)
26-01-18 Ro PPDTuned diff
LLR: -2.97 (-2.94,2.94) [0.00,5.00]
Total: 57411 W: 12984 L: 12829 D: 31598
sprt @ 10+0.1 th 1 Tuned values almost "as is". Factor for a pawn on the 7th rank is much lower than master. Penalty for king at distance > 4 is much lower too. Use [0,5] for a quick result before trying something simpler.
26-01-18 Ro LateralExposure diff
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 3300 W: 692 L: 808 D: 1800
sprt @ 10+0.1 th 1 Bonus for safe entry points for majors from which unsupported pawn can be attacked
25-01-18 Ro PPDTune diff
28941/30000 iterations
60000/60000 games played
60000 @ 20+0.2 th 1 Try SPSA to hint a possibly better formula
25-01-18 Ro KingInBetween diff
LLR: -2.96 (-2.94,2.94) [0.00,5.00]
Total: 24626 W: 5425 L: 5435 D: 13766
sprt @ 10+0.1 th 1 bonus when our king is in the path between their king and our passed pawn
25-01-18 Ro PPMinDistance diff
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 6311 W: 1331 L: 1431 D: 3549
sprt @ 10+0.1 th 1 Take 2
25-01-18 Ro KingDistanceDefender diff
LLR: -2.95 (-2.94,2.94) [0.00,4.00]
Total: 31945 W: 7058 L: 7096 D: 17791
sprt @ 10+0.1 th 1 SImply lower the coefficient
25-01-18 Ro PPMinDistance diff
LLR: -2.96 (-2.94,2.94) [0.00,5.00]
Total: 6862 W: 1483 L: 1581 D: 3798
sprt @ 10+0.1 th 1 The opponent king distance penalty is quite huge. Try to lower this when applicable.
25-01-18 Ro PPDistance diff
LLR: -2.95 (-2.94,2.94) [0.00,4.00]
Total: 6879 W: 1453 L: 1590 D: 3836
sprt @ 10+0.1 th 1 Take 3
24-01-18 Ro PPDistance2 diff
LLR: -2.95 (-2.94,2.94) [0.00,4.00]
Total: 10727 W: 2355 L: 2477 D: 5895
sprt @ 10+0.1 th 1 Take 2
24-01-18 Ro PPDistance diff
LLR: -2.96 (-2.94,2.94) [0.00,4.00]
Total: 10125 W: 2220 L: 2345 D: 5560
sprt @ 10+0.1 th 1 king distance bonus adjusted
24-01-18 Ro WeakPawnsAndMajors diff
LLR: -2.95 (-2.94,2.94) [0.00,4.00]
Total: 10942 W: 2350 L: 2471 D: 6121
sprt @ 10+0.1 th 1 Apply the weak unopposed count more often
23-01-18 Ro TopPassers diff
LLR: -2.96 (-2.94,2.94) [0.00,5.00]
Total: 13350 W: 2352 L: 2421 D: 8577
sprt @ 10+0.1 th 1 If many good passers, score only the bests
23-01-18 Ro CandidateTweak diff
LLR: -2.94 (-2.94,2.94) [0.00,4.00]
Total: 35539 W: 6385 L: 6420 D: 22734
sprt @ 10+0.1 th 1
22-01-18 Ro CandidateBlocker diff
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 13478 W: 2362 L: 2430 D: 8686
sprt @ 10+0.1 th 1 Take 2
22-01-18 Ro CandidateBlocker diff
LLR: -2.96 (-2.94,2.94) [0.00,5.00]
Total: 9486 W: 1656 L: 1741 D: 6089
sprt @ 10+0.1 th 1 A candidate passed pawn improvement
20-01-18 Ro MinorControl3 diff
LLR: -2.96 (-2.94,2.94) [0.00,5.00]
Total: 7658 W: 1385 L: 1478 D: 4795
sprt @ 10+0.1 th 1 Try a mg bonus only
20-01-18 Ro MinorControl3 diff
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 5003 W: 871 L: 975 D: 3157
sprt @ 10+0.1 th 1 Minor against major
20-01-18 Ro MinorControl3 diff
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 8886 W: 1593 L: 1680 D: 5613
sprt @ 10+0.1 th 1 Bonus for each attacked square challenged by a minor