Stockfish Testing Queue

Finished - 30415 tests

16-03-12 Elb pp_tune diff
38592/40000 iterations
78214/80000 games played
80000 @ 20+0.2 th 1 Tune passed pawn bonus values
16-03-11 Mys red diff
LLR: -2.97 (-2.94,2.94) [0.00,5.00]
Total: 33917 W: 6156 L: 6137 D: 21624
sprt @ 10+0.1 th 1 Let's see where else we can use fmh! reductions?
16-03-12 sg fmh_reductions diff
LLR: 2.96 (-2.94,2.94) [0.00,5.00]
Total: 94788 W: 17527 L: 16995 D: 60266
sprt @ 10+0.1 th 1 Use followup move history for history reduction. Replace cmh with average of cmh and fmh. Take 3
16-03-11 Voy fm-hp diff
LLR: -2.96 (-2.94,2.94) [0.00,5.00]
Total: 99070 W: 18322 L: 18021 D: 62727
sprt @ 10+0.1 th 1 Try my luck with using fmh with history pruning.
16-03-12 jos values_spsa2 diff
LLR: -2.95 (-2.94,2.94) [0.00,4.00]
Total: 80204 W: 14712 L: 14593 D: 50899
sprt @ 10+0.1 th 1 Test some tuned values. Tuned with 2-measurement SPSA (SPSA2).
16-03-12 sg fmh_reductions diff
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 87334 W: 11675 L: 11504 D: 64155
sprt @ 60+0.6 th 1 LTC: Use followup move history for history reduction. Compensate to get the same average rHist. Take 2
16-03-12 Voy fmh-bonus diff
LLR: -2.96 (-2.94,2.94) [0.00,5.00]
Total: 44516 W: 8138 L: 8073 D: 28305
sprt @ 10+0.1 th 1 Speculate that the weight of cmh > fmh. Finding the optimal weight for fmh will help to incorporate fmh with LMR/History Pruning.
16-03-12 sni attacked diff
LLR: -2.96 (-2.94,2.94) [0.00,5.00]
Total: 13696 W: 2461 L: 2528 D: 8707
sprt @ 10+0.1 th 1 Take 2: try small endgame bonus S(0,12)
16-03-12 sni attacked diff
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 11037 W: 2004 L: 2082 D: 6951
sprt @ 10+0.1 th 1 Take 3: try huge endgame bonus S(0,58)
16-03-12 tzi oppBishops_passedPawns1 diff
LLR: -2.96 (-2.94,2.94) [0.00,5.00]
Total: 6476 W: 1144 L: 1242 D: 4090
sprt @ 10+0.1 th 1 reduce ScaleFactor if there are no passed Pawns
16-03-12 sni attacked diff
LLR: -2.96 (-2.94,2.94) [0.00,5.00]
Total: 3134 W: 512 L: 624 D: 1998
sprt @ 10+0.1 th 1 Bonus for over-attacking weak pawns
16-03-12 Mys mo diff
LLR: -2.95 (-2.94,2.94) [0.00,4.00]
Total: 13605 W: 1684 L: 1796 D: 10125
sprt @ 60+0.6 th 1 LTC: Evasions tweak [0, 4]
16-03-11 Elb tw_init diff
LLR: -2.96 (-2.94,2.94) [0.00,4.00]
Total: 78558 W: 14411 L: 14298 D: 49849
sprt @ 10+0.1 th 1 Initiative parameter tweak - take 2
16-03-11 pb0 quite_raises diff
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 39417 W: 7226 L: 7182 D: 25009
sprt @ 10+0.1 th 1 Usually a quite move which produces an alpha raise is a good one, even if subsequently overtrumped by another one. So trying to give some bonus too in history.
16-03-11 sg fmh_reductions diff
LLR: 2.96 (-2.94,2.94) [0.00,5.00]
Total: 35274 W: 6524 L: 6248 D: 22502
sprt @ 10+0.1 th 1 Use followup move history for history reduction. Compensate to get the same average rHist. Take 2
16-03-11 Mys mo diff
LLR: 2.95 (-2.94,2.94) [0.00,4.00]
Total: 39896 W: 7431 L: 7137 D: 25328
sprt @ 10+0.1 th 1 Evasions tweak (pls if this has been tried let me know!)
16-03-10 Voy rf-1.95 diff
ELO: 0.28 +-3.5 (95%) LOS: 56.2%
Total: 10000 W: 1322 L: 1314 D: 7364
10000 @ 60+0.6 th 1 Try coefficient 1.95. Get a quick idea how this scales. At higher depth the reduction table changes quite a bit.
16-03-11 Elb tw_init diff
LLR: -2.95 (-2.94,2.94) [0.00,4.00]
Total: 18450 W: 3198 L: 3292 D: 11960
sprt @ 10+0.1 th 1 Tweak of initiative parameters with tuned values
16-03-11 sg fmh_reductions diff
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 8275 W: 1487 L: 1577 D: 5211
sprt @ 10+0.1 th 1 Use followup move history for history reduction. First try it without compensation.
16-03-11 Mys fmhp diff
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 3418 W: 576 L: 687 D: 2155
sprt @ 10+0.1 th 1 fmh based history pruning
16-03-11 Fis sc_simple diff
LLR: 2.96 (-2.94,2.94) [-3.00,1.00]
Total: 14324 W: 1935 L: 1806 D: 10583
sprt @ 60+0.6 th 1 LTC Simplify safe checks based on idea of jcalovski from http://tests.stockfishchess.org/tests/view/56e12c3d0ebc59301a353b01
16-03-10 Voy fmh-lmr diff
LLR: -2.96 (-2.94,2.94) [0.00,5.00]
Total: 22288 W: 4048 L: 4078 D: 14162
sprt @ 10+0.1 th 1 Use fmh in lmr.
16-03-10 Fis sc_simple diff
LLR: 2.96 (-2.94,2.94) [-3.00,1.00]
Total: 11796 W: 2211 L: 2074 D: 7511
sprt @ 10+0.1 th 1 Simplify safe checks based on idea of jcalovski from http://tests.stockfishchess.org/tests/view/56e12c3d0ebc59301a353b01
16-03-10 fbv AlignedPawns293 diff
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 80077 W: 14809 L: 14589 D: 50679
sprt @ 10+0.1 th 1 Fourth formula
16-03-10 Mys ch1 diff
LLR: -2.95 (-2.94,2.94) [0.00,4.00]
Total: 61189 W: 11347 L: 11292 D: 38550
sprt @ 10+0.1 th 1 A weird idea, but diminish check bonus for subsequent checks instead...
16-03-10 sg followup_moves_history diff
LLR: 2.95 (-2.94,2.94) [0.00,5.00]
Total: 5323 W: 778 L: 642 D: 3903
sprt @ 60+0.6 th 1 LTC: Collect analog to counter move history stats for follow up moves and use them at quiet move ordering. We can use the current CounterMoveHistoryStats array because only half of the entries are used (counter moves indexed by two moves from opposite colors). Instead followup moves are indexed by two moves of same color.
16-03-10 sg followup_moves_history diff
LLR: 2.96 (-2.94,2.94) [0.00,5.00]
Total: 7955 W: 1538 L: 1378 D: 5039
sprt @ 10+0.1 th 1 Collect analog to counter move history stats for follow up moves and use them at quiet move ordering. We can use the current CounterMoveHistoryStats array because only half of the entries are used (counter moves indexed by two moves from opposite colors). Instead followup moves are indexed by two moves of same color.
16-03-10 Voy rf-1.90 diff
LLR: -2.96 (-2.94,2.94) [0.00,4.00]
Total: 51284 W: 9363 L: 9344 D: 32577
sprt @ 10+0.1 th 1 stc
16-03-10 Mys ccm diff
LLR: -0.75 (-2.94,2.94) [0.00,5.00]
Total: 2396 W: 356 L: 378 D: 1662
sprt @ 5+0.05 th 7 Clear counter moves for ALL non-main threads. Similar idea to Fisherman
16-03-10 fbv AlignedPawns291-293 diff
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 9172 W: 1672 L: 1758 D: 5742
sprt @ 10+0.1 th 1 Fourth formula Take 2 { S(14, 54), S(21, 81) }
16-03-10 Voy lmrTweakD diff
LLR: -1.47 (-2.94,2.94) [0.00,5.00]
Total: 2159 W: 363 L: 416 D: 1380
sprt @ 10+0.1 th 1 Based on prior tests...it seems that the strength comes from increasing reduction.
16-03-09 Roc RookInFront diff
LLR: -2.97 (-2.94,2.94) [0.00,5.00]
Total: 27323 W: 5040 L: 5048 D: 17235
sprt @ 10+0.1 th 1 Fixed mixed up signatures.
16-03-08 Voy sbRetest diff
LLR: -2.96 (-2.94,2.94) [0.00,4.00]
Total: 144214 W: 19438 L: 19201 D: 105575
sprt @ 60+0.6 th 1 LTC: This patch back in October passed at STC, but I never tested it at LTC. http://tests.stockfishchess.org/tests/view/5632c8ad0ebc59227df16b8a Thanks for the reminder pb.
16-03-09 sni attack_units diff
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 8002 W: 1489 L: 1580 D: 4933
sprt @ 10+0.1 th 1 Use material instead of queens in attackUnits, take 2
16-03-09 sni attack_units diff
LLR: -2.96 (-2.94,2.94) [0.00,5.00]
Total: 3563 W: 660 L: 772 D: 2131
sprt @ 10+0.1 th 1 Use material instead of queens in attackUnits
16-03-09 Voy lmrTweakC diff
LLR: -2.94 (-2.94,2.94) [0.00,5.00]
Total: 8648 W: 1557 L: 1645 D: 5446
sprt @ 10+0.1 th 1 Version C.
16-03-07 Voy rf-1.95 diff
LLR: -2.96 (-2.94,2.94) [0.00,4.00]
Total: 122549 W: 22658 L: 22392 D: 77499
sprt @ 10+0.1 th 1 Try coefficient 1.95
16-03-09 fbv AlignedPawns141 diff
LLR: -2.96 (-2.94,2.94) [0.00,5.00]
Total: 57480 W: 10671 L: 10548 D: 36261
sprt @ 10+0.1 th 1 Third formula.
16-03-09 Mys ch diff
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 19413 W: 3534 L: 3576 D: 12303
sprt @ 10+0.1 th 1 Any strength in scoring an extra check?
16-03-09 pb0 parabola_shift diff
ELO: -1.80 +-4.1 (95%) LOS: 19.3%
Total: 10218 W: 1837 L: 1890 D: 6491
20000 @ 10+0.1 th 1 Although this bonus formula seems faulty (breaks the granularity) local results appear good. Quick check
16-03-09 fbv AlignedPawns257 diff
LLR: -2.96 (-2.94,2.94) [0.00,5.00]
Total: 22128 W: 4007 L: 4038 D: 14083
sprt @ 10+0.1 th 1 Second formula
16-03-09 sni pinned diff
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 21841 W: 3942 L: 3974 D: 13925
sprt @ 10+0.1 th 1 Try (11,13) for (protected pins, unprotected pins)
16-03-09 sni pinned diff
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 22932 W: 4174 L: 4201 D: 14557
sprt @ 10+0.1 th 1 Try (10,12) for (protected pins, unprotected pins)
16-03-09 fbv AlignedPawns139 diff
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 15293 W: 2843 L: 2902 D: 9548
sprt @ 10+0.1 th 1 New formula
16-03-08 sni pinned diff
LLR: -2.96 (-2.94,2.94) [0.00,5.00]
Total: 7603 W: 1326 L: 1419 D: 4858
sprt @ 10+0.1 th 1 Unprotected pinned pieces are more dangerous
16-03-08 Roc RookInFront diff
LLR: -1.91 (-2.94,2.94) [0.00,5.00]
Total: 12330 W: 2263 L: 2291 D: 7776
sprt @ 10+0.1 th 1 Take 2 (just change the mobility)
16-03-08 Roc RookInFront diff
LLR: -1.42 (-2.94,2.94) [0.00,5.00]
Total: 12079 W: 2239 L: 2247 D: 7593
sprt @ 10+0.1 th 1 Mobility calculation tweak for rook and queen
16-03-08 Elb block_lines2 diff
LLR: -0.93 (-2.94,2.94) [0.00,5.00]
Total: 4342 W: 783 L: 804 D: 2755
sprt @ 10+0.1 th 1 Take2: if all the rook mobility squares are controlled by the enemy, give a penalty.
16-03-08 Voy lmrTweak2b diff
LLR: -3.26 (-2.94,2.94) [0.00,5.00]
Total: 40442 W: 7521 L: 7484 D: 25437
sprt @ 10+0.1 th 1 Fixed faulty logic.
16-03-08 pb0 lowplies_history diff
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 24388 W: 4411 L: 4432 D: 15545
sprt @ 10+0.1 th 1 Introducing a 2nd layer in history for better move ordering for plies near the root, take 3.