Stockfish Testing Queue

Pending - 3 tests 0.0 hrs

19-08-18 Roc PinnedQueen2 diff
LLR: -2.02 (-2.94,2.94) [0.00,4.00]
Total: 25651 W: 5686 L: 5697 D: 14268
sprt @ 10+0.1 th 1 See git notes.
19-08-18 Roc 64738f80c5b7e53eebf8320 diff
LLR: -0.85 (-2.94,2.94) [0.50,4.50]
Total: 59891 W: 13348 L: 13099 D: 33444
sprt @ 10+0.1 th 1 Take 2b; leave RookOnFile as master, and reduce ThreatOnPawn.
19-07-31 Roc ROFControl diff
LLR: -0.88 (-2.94,2.94) [0.50,4.50]
Total: 9825 W: 2195 L: 2195 D: 5435
sprt @ 10+0.1 th 1 R_1 Bonus for restricting rook access to fully open file. Since the rook may be already on open file, increase RookOnFile[1] bonus accordingly to compensate.

Active - 0 tests

Finished - 2148 tests

19-08-18 Roc AttacksOnPawn3 diff
LLR: -2.96 (-2.94,2.94) [0.50,4.50]
Total: 29180 W: 6388 L: 6409 D: 16383
sprt @ 10+0.1 th 1 AOP3 Take 4
19-08-18 Roc AttacksOnpawn diff
LLR: -0.70 (-2.94,2.94) [0.50,4.50]
Total: 2840 W: 609 L: 634 D: 1597
sprt @ 10+0.1 th 1 take 3
19-08-18 Roc AttacksOnpawn diff
LLR: -2.96 (-2.94,2.94) [0.50,4.50]
Total: 21344 W: 4699 L: 4758 D: 11887
sprt @ 10+0.1 th 1 AOp1b Fixed bench problem in first attempt.
19-08-08 Roc UnsafeChecks4 diff
LLR: -2.95 (-2.94,2.94) [0.50,4.50]
Total: 17034 W: 3794 L: 3874 D: 9366
sprt @ 10+0.1 th 1 UC4_125-175 Take 2
19-08-08 Roc UnsafeChecks4 diff
LLR: -2.95 (-2.94,2.94) [0.50,4.50]
Total: 15456 W: 3384 L: 3472 D: 8600
sprt @ 10+0.1 th 1 UC4-100-200 Less penalty if attacked by one More penalty if attacked by 2
19-08-01 Roc CandidatePasserScore3 diff
LLR: -2.95 (-2.94,2.94) [0.50,4.50]
Total: 56402 W: 12624 L: 12508 D: 31270
sprt @ 10+0.1 th 1 CPS3 Same as master. However remove 1/3 instead of 1/2 if the candidate does not depend on another "passed" pawn.
19-08-01 Roc CandidatePasserScore3 diff
LLR: -2.95 (-2.94,2.94) [0.50,4.50]
Total: 14824 W: 3234 L: 3325 D: 8265
sprt @ 10+0.1 th 1 Lat try looks neutral, try reducing by 1/4 instead of 1/3
19-08-01 Roc CandiatePasserScore2 diff
LLR: -2.96 (-2.94,2.94) [0.50,4.50]
Total: 7473 W: 1601 L: 1729 D: 4143
sprt @ 10+0.1 th 1 CPS2 Allow full passed pawn bonus in more cases.
19-07-31 Roc ROFControl1b diff
LLR: -2.95 (-2.94,2.94) [0.50,4.50]
Total: 20587 W: 4609 L: 4671 D: 11307
sprt @ 10+0.1 th 1 R1b exclude squares occupied by own pieces
19-07-31 Roc ROFControl2 diff
LLR: -2.95 (-2.94,2.94) [0.50,4.50]
Total: 4528 W: 929 L: 1071 D: 2528
sprt @ 10+0.1 th 1 Bonus for restricting horizontal access of a rook to a fully open file.
19-07-29 Roc KProx3 diff
LLR: -2.95 (-2.94,2.94) [0.50,4.50]
Total: 10559 W: 2296 L: 2408 D: 5855
sprt @ 10+0.1 th 1 KP3 Take3: Try factor 5
19-07-27 Roc KProx2 diff
LLR: -2.95 (-2.94,2.94) [0.50,4.50]
Total: 32200 W: 7099 L: 7104 D: 17997
sprt @ 10+0.1 th 1 KP2 Take 2: larger bonus
19-07-27 Roc KProx1 diff
LLR: -2.96 (-2.94,2.94) [0.50,4.50]
Total: 71643 W: 16008 L: 15817 D: 39818
sprt @ 10+0.1 th 1 KP1 Add small bonus when enemy king is on the wrong rank side of the passed pawn. For example White e4, Black Ke3
19-07-25 Roc TunedPassedFile diff
LLR: -2.95 (-2.94,2.94) [0.00,4.00]
Total: 70044 W: 15638 L: 15524 D: 38882
sprt @ 10+0.1 th 1 TPF-1b Using correct sign when applying the change will help ;)
19-07-25 Roc TunedPassedFile diff
LLR: -0.19 (-2.94,2.94) [0.00,4.00]
Total: 328 W: 66 L: 75 D: 187
sprt @ 10+0.1 th 1 TPF
19-07-24 Roc PPFileTune diff
27819/30000 iterations
60000/60000 games played
60000 @ 20+0.2 th 1 Applied 3 passed pawn PRs and tune PassedRank and PassedFile
19-07-16 Roc SafePawnThreats diff
LLR: -2.95 (-2.94,2.94) [0.50,4.50]
Total: 74369 W: 16700 L: 16494 D: 41175
sprt @ 10+0.1 th 1 SPT_1 Replace the safe bitboard (safe squares to make a pawn attack) with a different one, and some code shuffling.
19-07-20 Roc PawnCleanUp diff
LLR: 2.95 (-2.94,2.94) [-3.00,1.00]
Total: 124666 W: 27586 L: 27669 D: 69411
sprt @ 10+0.1 th 1 Non functional simplification of the passed pawn calculation, and some code clean up
19-07-21 Roc ShelterX diff
LLR: -2.94 (-2.94,2.94) [0.50,4.50]
Total: 14564 W: 3168 L: 3260 D: 8136
sprt @ 10+0.1 th 1 A follow up on impede shelter idea from Fisherman, exact computation according to current position
19-07-21 Roc ClosedCenterStorm diff
LLR: -2.95 (-2.94,2.94) [0.50,4.50]
Total: 43945 W: 9892 L: 9837 D: 24216
sprt @ 10+0.1 th 1 CCS_30_0 Take 2, mg bonus only
19-07-21 Roc ClosedCenterStorm diff
LLR: -2.95 (-2.94,2.94) [0.50,4.50]
Total: 53659 W: 12020 L: 11917 D: 29722
sprt @ 10+0.1 th 1 CCS_30_30 Penalize king safety when closed enemy center pointing to the king with the advanced center pawn on relative rank 5. (for example, the french structure, white d4, e5 against Black d5 e6 and Black king on the king side)
19-07-20 Roc PassedFileV2 diff
LLR: 2.95 (-2.94,2.94) [-3.00,1.00]
Total: 18525 W: 3176 L: 3052 D: 12297
sprt @ 60+0.6 th 1 LTC for green STC, removing additive term, tested against PR 2250
19-07-20 Roc SafePawnThreats diff
LLR: -2.95 (-2.94,2.94) [0.50,4.50]
Total: 49726 W: 11130 L: 11047 D: 27549
sprt @ 10+0.1 th 1 Take 3 and last, trying to use the stronglyProtected bitboard in the safe pawn threats expression
19-07-20 Roc PassedFileV2 diff
LLR: 2.96 (-2.94,2.94) [-3.00,1.00]
Total: 81459 W: 18174 L: 18171 D: 45114
sprt @ 10+0.1 th 1 Fixed a mistake in previous test, which dramatically changed the std::min(f, ~f) result.
19-07-20 Roc SafePawnThreats diff
LLR: -2.95 (-2.94,2.94) [0.50,4.50]
Total: 12424 W: 2774 L: 2877 D: 6773
sprt @ 10+0.1 th 1 SPT-2 take 2 on this idea, always include undefended squares as master
19-07-20 Roc PassedFileV2 diff
LLR: -2.96 (-2.94,2.94) [-3.00,1.00]
Total: 1943 W: 400 L: 575 D: 968
sprt @ 10+0.1 th 1 Tested as a simplification against PR #2250
19-07-20 Roc SafePawnThreat2 diff
LLR: -2.95 (-2.94,2.94) [-3.00,1.00]
Total: 43103 W: 9606 L: 9858 D: 23639
sprt @ 10+0.1 th 1 SPT2 Simplify the pawn threat calculation
19-07-15 Roc pawnsimpl2 diff
LLR: -2.95 (-2.94,2.94) [-3.00,1.00]
Total: 32741 W: 7027 L: 7255 D: 18459
sprt @ 10+0.1 th 1 for the new Attacked2Unsupported, passers were recently excluded along with a larger bonus. Verify if the larger bonus alone is OK.
19-07-15 Roc BishopPawnWeight diff
LLR: -2.95 (-2.94,2.94) [-3.00,1.00]
Total: 2387 W: 465 L: 636 D: 1286
sprt @ 10+0.1 th 1 See if using the central pawn mass gives a similar benefit as counting our central blocked pawns. Divide BishopPawn by 2 to compensate,
19-07-15 Roc EdgePawns diff
LLR: -2.95 (-2.94,2.94) [0.00,4.00]
Total: 10837 W: 2370 L: 2492 D: 5975
sprt @ 10+0.1 th 1 Take 2, try the contrary, always include unblocked pawns on edge files. So compared to master, this would include unblocked a/h/2/3, b/g/2/3 in the mobility area. Run as parameter tweak.
19-07-15 Roc EdgePawns diff
LLR: -2.96 (-2.94,2.94) [0.50,4.50]
Total: 9237 W: 1974 L: 2093 D: 5170
sprt @ 10+0.1 th 1 EP_1 What is the impact of excluding pawns on edge files from mobility area
19-07-15 Roc EdgedBishop diff
LLR: -2.95 (-2.94,2.94) [0.50,4.50]
Total: 26281 W: 5848 L: 5882 D: 14551
sprt @ 10+0.1 th 1 EB_0_40
19-07-15 Roc BishopBlockers3 diff
LLR: -2.96 (-2.94,2.94) [0.50,4.50]
Total: 13525 W: 2898 L: 2996 D: 7631
sprt @ 10+0.1 th 1 Include a few more same_color_pawns in the bishop calculation.
19-07-15 Roc BishopBlockers2 diff
LLR: -2.96 (-2.94,2.94) [0.50,4.50]
Total: 56543 W: 12624 L: 12508 D: 31411
sprt @ 10+0.1 th 1 fixed version 1b
19-07-14 Roc SpaceInvaders2b diff
LLR: 2.95 (-2.94,2.94) [-3.00,1.00]
Total: 63879 W: 14076 L: 14038 D: 35765
sprt @ 10+0.1 th 1 Take 2, with higher weight to compensate
19-07-15 Roc BishopBlockers2 diff
LLR: -0.52 (-2.94,2.94) [0.50,4.50]
Total: 41 W: 1 L: 39 D: 1
sprt @ 10+0.1 th 1 BB2 include, in the bishop pawn formula, some safe enemy pawns
19-07-14 Roc SpaceInvaders2 diff
LLR: 2.95 (-2.94,2.94) [-3.00,1.00]
Total: 49761 W: 8409 L: 8335 D: 33017
sprt @ 60+0.6 th 1 LTC for green STC simplification
19-07-14 Roc SpaceInvaders2 diff
LLR: 2.95 (-2.94,2.94) [-3.00,1.00]
Total: 51299 W: 11320 L: 11257 D: 28722
sprt @ 10+0.1 th 1 SI-2 Try a more ambitius simplification
19-07-14 Roc SpaceInvaders diff
LLR: -0.45 (-2.94,2.94) [-3.00,1.00]
Total: 25094 W: 4209 L: 4273 D: 16612
sprt @ 60+0.6 th 1 LTC, running as simplification
19-07-14 Roc SpaceInvaders diff
LLR: -2.95 (-2.94,2.94) [0.50,4.50]
Total: 31065 W: 7004 L: 7013 D: 17048
sprt @ 10+0.1 th 1 SI-1 Let me know if this was already tried.
19-07-14 Roc KSafetyQ_Tuned diff
LLR: -2.96 (-2.94,2.94) [0.50,4.50]
Total: 10711 W: 2327 L: 2439 D: 5945
sprt @ 10+0.1 th 1 KQ_T1 Testing some raw spsa tuned values
19-07-14 Roc pawnonly_eval diff
LLR: -2.95 (-2.94,2.94) [0.50,4.50]
Total: 14065 W: 3051 L: 3146 D: 7868
sprt @ 10+0.1 th 1 PE_1b skip some initialisations and some evaluations when only king and pawn
19-07-14 Roc pawnonly_eval diff
LLR: -2.95 (-2.94,2.94) [0.50,4.50]
Total: 12444 W: 2674 L: 2777 D: 6993
sprt @ 10+0.1 th 1 PE_2 same as previous test but always use the king shelter and storm evaluation
19-07-14 Roc KSafetyQ diff
45078/50000 iterations
99148/100000 games played
100000 @ 20+0.2 th 1 Different parameter values depending that opponent has queen or not, hoping to see some trend.
19-07-13 Roc HomeProtection diff
LLR: -2.95 (-2.94,2.94) [0.50,4.50]
Total: 19048 W: 4150 L: 4220 D: 10678
sprt @ 10+0.1 th 1 HP_3 Take 3
19-07-13 Roc HomeProtection diff
LLR: -2.96 (-2.94,2.94) [0.50,4.50]
Total: 17828 W: 3858 L: 3935 D: 10035
sprt @ 10+0.1 th 1 HP_2 take 2
19-07-13 Roc HomeProtection diff
LLR: -2.95 (-2.94,2.94) [0.50,4.50]
Total: 9595 W: 2066 L: 2183 D: 5346
sprt @ 10+0.1 th 1 HP_1 Higher bonus for restricting on our half.
19-07-13 Roc PawnControlMob diff
LLR: -2.95 (-2.94,2.94) [0.50,4.50]
Total: 9000 W: 1928 L: 2048 D: 5024
sprt @ 10+0.1 th 1 PCM
19-07-13 Roc ProOutpost2 diff
LLR: -2.96 (-2.94,2.94) [0.50,4.50]
Total: 21390 W: 4700 L: 4759 D: 11931
sprt @ 10+0.1 th 1 PO2d Try smaller bonus in the shielded case.
19-07-13 Roc ProOutpost2 diff
LLR: -2.96 (-2.94,2.94) [0.50,4.50]
Total: 44507 W: 9887 L: 9831 D: 24789
sprt @ 10+0.1 th 1 Require outpost to be pawn protected or screened by a pawn, (re) fixed version