Stockfish Testing Queue

Finished - 37194 tests

15-05-07 SC MVV_evasions diff
LLR: -2.96 (-2.94,2.94) [-3.00,1.00]
Total: 14748 W: 2809 L: 2996 D: 8943
sprt @ 15+0.05 th 1 1) In scoring evasions value of captured piece should not play any role, so MVV does not apply (you can always just capture one piece) 2) LVA is (according to Lucas) already taken into account by move generator and stable sorting. Maybe is possible to brutally simplify evasions scoring.
15-05-07 sni capture_score diff
LLR: -2.95 (-2.94,2.94) [-1.50,4.50]
Total: 8469 W: 1574 L: 1657 D: 5238
sprt @ 15+0.05 th 1 Respin of DU-jdto test (take 2 : with 100 * distanceFromKing)
15-05-07 Roc NoPinPawnThreatV2 diff
LLR: -2.95 (-2.94,2.94) [0.00,6.00]
Total: 11195 W: 1755 L: 1805 D: 7635
sprt @ 60+0.05 th 1 Take 2. @LTC
15-05-07 lbr TTSmartSave diff
LLR: 2.96 (-2.94,2.94) [-5.00,0.00]
Total: 18944 W: 3607 L: 3564 D: 11773
sprt @ 15+0.05 th 1 Take2. Don't overwrite more valuable TT data w/ less valuable. => Check that it's not a regression under normal testing conditions (16mb STC)
15-05-06 Hai capture_score diff
LLR: -3.28 (-2.94,2.94) [-1.50,4.50]
Total: 32240 W: 6187 L: 6218 D: 19835
sprt @ 15+0.05 th 1 Score captures based on king position
15-05-07 sni capture_to_center diff
LLR: -2.95 (-2.94,2.94) [-1.50,4.50]
Total: 14053 W: 2675 L: 2743 D: 8635
sprt @ 15+0.05 th 1 Prefer captures in the central files as a tertiary sorting criteria.Credits to Ralph Stößer, Alain Savard and Stefano Cardanobile.
15-05-06 SC bestMoveChanges diff
LLR: -5.46 (-2.94,2.94) [-1.50,4.50]
Total: 34689 W: 6526 L: 6629 D: 21534
sprt @ 15+0.05 th 1 Use log2 instead of H of move count to increment besteMoveChanges. Results from tuning time management parameters.
15-05-06 nas linear_quadratic_parame diff
9894/10000 iterations
20000/20000 games played
20000 @ 15+0.05 th 1 linear and quadratic parameter tuning
15-05-06 SC MVV_rank_tuned diff
LLR: -2.96 (-2.94,2.94) [0.00,4.00]
Total: 29421 W: 5606 L: 5660 D: 18155
sprt @ 15+0.05 th 1 Higher values as indicated by SPSA (using a much higher values, since SPSA did not converge and I have other evidence indicating that something around 250 should be better).
15-05-03 Voy Pufferfish3 diff
LLR: -2.96 (-2.94,2.94) [-1.50,4.50]
Total: 77401 W: 14984 L: 14882 D: 47535
sprt @ 15+0.05 th 1 Pufferfish2 failed like a midget at basketball tryout...So lets do the opposite. On repetitive moves don't update counter and killers.
15-05-05 Roc NoPinPawnThreatV2 diff
LLR: 3.11 (-2.94,2.94) [-1.50,4.50]
Total: 38772 W: 7538 L: 7323 D: 23911
sprt @ 15+0.05 th 1 Take 2.
15-05-06 jos RookOnAdjacentOpenFile diff
LLR: -2.97 (-2.94,2.94) [-1.50,4.50]
Total: 9113 W: 1729 L: 1811 D: 5573
sprt @ 15+0.05 th 1 Bonus for rook on a open file adjacent to the enemy king.
15-05-06 sg lmr diff
LLR: -2.95 (-2.94,2.94) [-1.50,4.50]
Total: 17016 W: 3281 L: 3341 D: 10394
sprt @ 15+0.05 th 1 LMR: less reduction if cmh > 0 (replaced countermove condition)
15-05-06 Fis TTSmartSave diff
LLR: 2.97 (-2.94,2.94) [0.00,6.00]
Total: 13381 W: 2149 L: 1987 D: 9245
sprt @ 60+0.05 th 1 Take2. Don't overwrite more valuable TT data w/ less valuable. 8MB LTC
15-05-05 lbr noclear diff
LLR: 2.95 (-2.94,2.94) [0.00,5.00]
Total: 17893 W: 2984 L: 2792 D: 12117
sprt @ 60+0.05 th 1 never clear stats
15-05-04 SC MVV_rank_tuning diff
10310/10000 iterations
19648/20000 games played
20000 @ 30+0.05 th 1 Tune the rank penalty in MVV/rank for scoring captures.
15-04-27 SC bestMoveChangesTuning diff
39617/20000 iterations
71716/80000 games played
80000 @ 60+0.2 th 1 Given that the manually tuned bestMoveChanges performed much better than the trivial one and the BMCtime is not looking really promising, I'll give a try at tuning per SPSA. Using nodestime as specified in the guidelines, and using a longer tc in the hope to be more sensitive on time management. (Both as in the previous two patches).
15-05-04 Fis TTSmartSave diff
LLR: 2.96 (-2.94,2.94) [-1.50,4.50]
Total: 21132 W: 4108 L: 3946 D: 13078
sprt @ 15+0.05 th 1 Take2. Don't overwrite more valuable TT data w/ less valuable. 2MB
15-05-05 SC see_depth diff
LLR: -2.95 (-2.94,2.94) [-3.00,1.00]
Total: 3059 W: 506 L: 670 D: 1883
sprt @ 15+0.05 th 1 Retire see_sign, take 2. This time non functionally, with incosistent speed changes.
15-05-05 gli Anglerfish diff
LLR: 2.95 (-2.94,2.94) [0.00,6.00]
Total: 11196 W: 1852 L: 1699 D: 7645
sprt @ 60+0.05 th 1 [LTC for VoyagerOne because I'm super excited to see the results]: Only clear history stats if played move is a capture. (This time I should have the correct bench)
15-05-04 sg pruning diff
LLR: -2.94 (-2.94,2.94) [0.00,6.00]
Total: 15770 W: 2528 L: 2556 D: 10686
sprt @ 60+0.05 th 1 if at pv node and no preliminary best move is found don't prune move (Take 2)
15-05-05 lbr noclear diff
LLR: 2.95 (-2.94,2.94) [-1.00,4.00]
Total: 5375 W: 1119 L: 977 D: 3279
sprt @ 15+0.05 th 1 never clear stats
15-05-04 lan half_rook-psqt diff
LLR: -2.95 (-2.94,2.94) [0.00,4.00]
Total: 11602 W: 2180 L: 2297 D: 7125
sprt @ 15+0.05 th 1 Updated hand-tuned rook psqt (half-table)
15-05-04 Roc NoPinPawnThreatV2 diff
LLR: -2.95 (-2.94,2.94) [-1.50,4.50]
Total: 7311 W: 1355 L: 1441 D: 4515
sprt @ 15+0.05 th 1 Small tweak to mbootsector test
15-04-29 Roc BishopMobV4 diff
LLR: 2.96 (-2.94,2.94) [-1.50,4.50]
Total: 113087 W: 21907 L: 21498 D: 69682
sprt @ 15+0.05 th 1 Opponent Bishop long-term mobility is penalized by same color pawns. This test considers our wedges (example c3-d4-e3) as an additional long term structural hurdle (for example, for a Bg7). Based on an idea by Fauzi.
15-05-04 SC statistical_see_sign diff
LLR: -2.95 (-2.94,2.94) [-1.50,4.50]
Total: 9596 W: 1789 L: 1869 D: 5938
sprt @ 15+0.05 th 1 statistical_see was something like -250 ELO. Is there a chance for statistical_see_sign?
15-05-04 Voy Anglerfish diff
LLR: 2.95 (-2.94,2.94) [-1.50,4.50]
Total: 2778 W: 606 L: 492 D: 1680
sprt @ 15+0.05 th 1 Only clear history stats if played move is a capture. (This time I should have the correct bench)
15-05-04 SC see_depth diff
LLR: -2.95 (-2.94,2.94) [-3.00,1.00]
Total: 639 W: 72 L: 236 D: 331
sprt @ 15+0.05 th 1 Try to retire see_sign by adding a depth argument to the signature of see, at which the swap algorithm is stopped. If see_sign(m) >= 0 then see(m, 2) would be also >= 0 but the inverse is not true. Let's see how far this go. One could then tune the swap depth for the specific tuning purposes.
15-05-03 sg lmr diff
LLR: -2.97 (-2.94,2.94) [0.00,6.00]
Total: 25745 W: 4068 L: 4052 D: 17625
sprt @ 60+0.05 th 1 if at pv node and no preliminary best move is found exclude counter move from LMR
15-05-03 sg pruning diff
LLR: 2.96 (-2.94,2.94) [-1.50,4.50]
Total: 18039 W: 3471 L: 3318 D: 11250
sprt @ 15+0.05 th 1 if at pv node and no preliminary best move is found don't prune move (Take 2)
15-05-03 lbr passed diff
LLR: -2.95 (-2.94,2.94) [-1.00,4.00]
Total: 15995 W: 3062 L: 3146 D: 9787
sprt @ 15+0.05 th 1 tuned passed pawns
15-04-28 Roc OutpostV3 diff
LLR: -2.97 (-2.94,2.94) [-1.50,4.50]
Total: 55987 W: 10812 L: 10768 D: 34407
sprt @ 15+0.05 th 1 New definition for outpost. Pre calculated in the Pawn Entry structure.
15-05-03 SC statistical_see diff
LLR: -2.96 (-2.94,2.94) [-3.00,1.00]
Total: 210 W: 3 L: 159 D: 48
sprt @ 15+0.05 th 1 Replace Position::see() and Position::see_sign() by a statistical formula in Position::see(). This is quite far fetched, but one never knows. I was in particular surprised than nps went considerably down.
15-05-01 lbr psq diff
LLR: 3.11 (-2.94,2.94) [-0.50,4.50]
Total: 58764 W: 11530 L: 11185 D: 36049
sprt @ 15+0.05 th 1 last try
15-05-02 mbo no_pinned_pawn_threats diff
LLR: -2.96 (-2.94,2.94) [0.00,6.00]
Total: 16592 W: 2646 L: 2671 D: 11275
sprt @ 60+0.05 th 1 Do not add threat bonus from pinned pawns. Inspired by a loss to Komodo 9. Game is on talkchess viewtopic.php?start=40&t=56109
15-05-02 lbr mobility diff
LLR: -2.95 (-2.94,2.94) [-0.50,4.50]
Total: 34423 W: 6598 L: 6602 D: 21223
sprt @ 15+0.05 th 1 tuned mobility
15-05-02 jos piece_values diff
LLR: -2.96 (-2.94,2.94) [0.00,4.00]
Total: 16255 W: 3140 L: 3241 D: 9874
sprt @ 15+0.05 th 1 Proposed piece values.
15-04-30 Voy Monkfish diff
LLR: -3.03 (-2.94,2.94) [-1.50,4.50]
Total: 64092 W: 12216 L: 12154 D: 39722
sprt @ 15+0.05 th 1 If cmh and history signs disagree, then use cmh value.
15-05-02 Voy Pufferfish2 diff
LLR: -2.94 (-2.94,2.94) [-1.50,4.50]
Total: 8424 W: 1586 L: 1669 D: 5169
sprt @ 15+0.05 th 1 Don't update history stats for repetitive move, except for killer and counter moves.
15-05-01 sg lmr diff
LLR: 3.32 (-2.94,2.94) [-1.50,4.50]
Total: 31351 W: 6121 L: 5918 D: 19312
sprt @ 15+0.05 th 1 if at pv node and no preliminary best move is found exclude counter move from LMR
15-05-01 jos tuned_space-weight diff
ELO: 0.30 +-3.0 (95%) LOS: 57.7%
Total: 20000 W: 3881 L: 3864 D: 12255
20000 @ 15+0.05 th 1 Measure some interim values. Space advantage seems much more valuable towards the endgame.
15-05-01 mbo no_pinned_pawn_threats diff
LLR: 2.95 (-2.94,2.94) [-1.50,4.50]
Total: 24894 W: 4839 L: 4667 D: 15388
sprt @ 15+0.05 th 1 Do not add threat bonus from pinned pawns. Inspired by a loss to Komodo 9. Game is on talkchess viewtopic.php?start=40&t=56109
15-04-30 Roc BishopMobV4 diff
LLR: -2.95 (-2.94,2.94) [-1.50,4.50]
Total: 12627 W: 2352 L: 2424 D: 7851
sprt @ 15+0.05 th 1 Take 2: add also some mobility when working with a lever.
15-04-27 sg pruning diff
LLR: -2.95 (-2.94,2.94) [-1.50,4.50]
Total: 93862 W: 18236 L: 18089 D: 57537
sprt @ 15+0.05 th 1 if at pv node and no preliminary best move is found don't prune killers and counter move
15-05-01 Voy Monkfish2 diff
LLR: -2.96 (-2.94,2.94) [-1.50,4.50]
Total: 3203 W: 546 L: 643 D: 2014
sprt @ 15+0.05 th 1 Move Ordering: If cmh value is available use that, else use history.
15-04-30 SC MVV_MAV diff
LLR: 2.95 (-2.94,2.94) [0.00,6.00]
Total: 25770 W: 4184 L: 3964 D: 17622
sprt @ 60+0.05 th 1 On bench PieceValue - 200* relativeRank is a better approximation of pos.see than PieceValue alone (Pearson correlation from 0.55 to 0.59). See whether it is enough to beat MVV.
15-05-01 sg bishops diff
LLR: -2.95 (-2.94,2.94) [-1.50,4.50]
Total: 9229 W: 1708 L: 1789 D: 5732
sprt @ 15+0.05 th 1 Inspired by fauzi's post i try it the other way around. Penalty for bishop if hindered by own blocked pawns
15-04-18 sg king_safety diff
LLR: -2.97 (-2.94,2.94) [-1.50,4.50]
Total: 73859 W: 14291 L: 14199 D: 45369
sprt @ 15+0.05 th 1 Retest take 2 with weight=1: Very small penalty if pawn in king ring is attacked by opponent
15-04-30 Voy Pufferfish diff
LLR: -2.96 (-2.94,2.94) [-1.50,4.50]
Total: 18292 W: 3451 L: 3508 D: 11333
sprt @ 15+0.05 th 1 Don't update history stats for repetitive move.
15-04-28 jki see diff
ELO: -10.03 +-3.0 (95%) LOS: 0.0%
Total: 20000 W: 3534 L: 4111 D: 12355
20000 @ 15+0.05 th 1 SEE vs. MVV