Stockfish Testing Queue

Finished - 37193 tests

13-02-24 alhpa_pruning diff
ELO: 1.63 +-4.5 (95%) LOS: 88.1%
Total: 23000 W: 4232 L: 4124 D: 14644
30001 @ 20+0.05 th 1 From lucas_see_pv: test only the pruning condition on alpha
13-02-24 lucaas_see_pv_1 diff
ELO: 5.18 +-6.8 (95%) LOS: 99.2%
Total: 10000 W: 1952 L: 1803 D: 6245
10000 @ 20+0.05 th 1 From lucas_see_pv: test only the pruning condition on PV nodes
13-02-25 space diff
ELO: -2.17 +-5.4 (95%) LOS: 10.5%
Total: 16000 W: 3128 L: 3228 D: 9644
16000 @ 15+0.05 th 1 Reduce space weight
13-02-25 scale_with_gameplay diff
ELO: 1.46 +-6.8 (95%) LOS: 75.5%
Total: 10000 W: 1865 L: 1823 D: 6312
10000 @ 20+0.05 th 1 Last try with scaling midgame instead of end game, always 1%
13-02-25 singular_tweak4 diff
ELO: 1.94 +-4.4 (95%) LOS: 92.0%
Total: 24000 W: 4620 L: 4486 D: 14894
24000 @ 15+0.05 th 1 Re-test against current master
13-02-25 rook_evaluation diff
ELO: -6.62 +-5.4 (95%) LOS: 0.0%
Total: 16000 W: 2921 L: 3226 D: 9853
16000 @ 15+0.05 th 1 From Eelco
13-02-25 alhpa_pruning diff
ELO: 3.20 +-4.8 (95%) LOS: 98.5%
Total: 19999 W: 3675 L: 3491 D: 12833
20000 @ 20+0.05 th 1 Rationale behind condition on alpha is not so clear to me. So remove it.
13-02-25 space diff
ELO: -1.95 +-5.4 (95%) LOS: 12.4%
Total: 15999 W: 2979 L: 3069 D: 9951
16000 @ 15+0.05 th 1 Reduce weight even more
13-02-25 easy_move diff
ELO: -155.54 +-23.1 (95%) LOS: 0.0%
Total: 1000 W: 68 L: 488 D: 444
16000 @ 45+0.05 th 1
13-02-26 xray diff
ELO: -26.46 +-21.5 (95%) LOS: 0.0%
Total: 1000 W: 177 L: 253 D: 570
16000 @ 15+0.05 th 1 X-ray through more types of pieces
13-02-26 scale_with_gameplay diff
ELO: -3.13 +-7.2 (95%) LOS: 7.5%
Total: 9000 W: 1541 L: 1622 D: 5837
10000 @ 20+0.05 th 1 Becaue 1% midgame is not bad, increase to 2%
13-02-26 6c7e7a2 diff
ELO: -0.87 +-4.4 (95%) LOS: 26.6%
Total: 24000 W: 4596 L: 4656 D: 14748
24000 @ 15+0.05 th 1 Re-test successful scale version
13-02-26 trapped_bishop diff
ELO: -2.50 +-5.4 (95%) LOS: 6.9%
Total: 15999 W: 2958 L: 3073 D: 9968
16000 @ 15+0.05 th 1 Generalize trapped bishop logic
13-02-27 master diff
ELO: 2.08 +-5.6 (95%) LOS: 90.0%
Total: 15000 W: 2504 L: 2414 D: 10082
16000 @ 60+0.05 th 1 Regression test after Lucas SEE prune merge
13-02-27 easy_move diff
ELO: 8.12 +-5.8 (95%) LOS: 100.0%
Total: 14000 W: 2475 L: 2148 D: 9377
16000 @ 45+0.05 th 1 Tweak easy move take #2
13-02-27 space diff
ELO: -2.48 +-5.4 (95%) LOS: 7.4%
Total: 16000 W: 3033 L: 3147 D: 9820
16000 @ 15+0.05 th 1 More subtle space tweak
13-02-27 xray diff
ELO: -33.92 +-12.5 (95%) LOS: 0.0%
Total: 3000 W: 491 L: 783 D: 1726
16000 @ 15+0.05 th 1 More x-ray take 2
13-02-27 king_safety_tune diff
ELO: -0.87 +-5.4 (95%) LOS: 30.2%
Total: 16000 W: 2964 L: 3004 D: 10032
16000 @ 15+0.05 th 1 Slightly reduce king safety scores
13-02-28 see_prune_depth diff
ELO: 9.56 +-6.8 (95%) LOS: 100.0%
Total: 10000 W: 1949 L: 1674 D: 6377
10000 @ 20+0.05 th 1 After Lucas patch on SEE pruning in PV nodes, it seems this is a sensible parameter
13-02-28 pinned_null diff
ELO: 4.24 +-4.8 (95%) LOS: 99.9%
Total: 20000 W: 3380 L: 3136 D: 13484
20000 @ 60+0.05 th 1 Retest at longer TC
13-02-28 trapped_bishop diff
ELO: -1.72 +-5.4 (95%) LOS: 15.9%
Total: 16000 W: 3080 L: 3159 D: 9761
16000 @ 15+0.05 th 1 Tweak mobility as well
13-02-28 rook_pawn diff
ELO: -3.02 +-5.4 (95%) LOS: 3.9%
Total: 16000 W: 3059 L: 3198 D: 9743
16000 @ 15+0.05 th 1 Tweak rook/queen attacking pawn bonus
13-02-28 non_pv_rule50 diff
ELO: -10.22 +-9.6 (95%) LOS: 0.0%
Total: 5000 W: 922 L: 1069 D: 3009
16000 @ 15+0.05 th 1 Lower rule50 draw in non-pv
13-03-02 yet_easy_move diff
ELO: -0.97 +-6.8 (95%) LOS: 30.9%
Total: 10000 W: 1559 L: 1587 D: 6854
10000 @ 60+0.05 th 1 Yet another attempt at easy move improving
13-03-02 yet_see_prune diff
ELO: 14.14 +-5.4 (95%) LOS: 100.0%
Total: 16000 W: 3012 L: 2361 D: 10627
16000 @ 60+0.05 th 1 Very speculative (next week I'm out): I bet on Lucas patch to fail and mine to win :-)
13-03-02 easy_move diff
ELO: 3.13 +-12.4 (95%) LOS: 81.6%
Total: 3000 W: 465 L: 438 D: 2097
10000 @ 90+0.1 th 1 Only try easy move once
13-03-02 simplify_dangerous diff
ELO: -1.15 +-5.4 (95%) LOS: 24.7%
Total: 16000 W: 2986 L: 3039 D: 9975
16000 @ 15+0.05 th 1 Simplify dangerous moves a bit
13-03-04 master diff
ELO: 15.00 +-4.8 (95%) LOS: 100.0%
Total: 20000 W: 3769 L: 2906 D: 13325
20000 @ 60+0.05 th 1 Regression test vs sf_2.3.1
13-03-04 max_time diff
ELO: 2.17 +-5.9 (95%) LOS: 89.6%
Total: 13472 W: 2266 L: 2182 D: 9024
16000 @ 60+0.05 th 1 More aggressive trying to finish iterations
13-03-04 see_prune2 diff
ELO: 3.13 +-4.4 (95%) LOS: 98.8%
Total: 24000 W: 4714 L: 4498 D: 14788
24000 @ 15+0.05 th 1 3 worked, why not 4?
13-03-06 razor5 diff
ELO: -1.98 +-5.4 (95%) LOS: 12.4%
Total: 16000 W: 3050 L: 3141 D: 9809
16000 @ 15+0.05 th 1 Bump razor depth up to 5 ply
13-03-06 razor3 diff
ELO: -0.61 +-5.4 (95%) LOS: 36.0%
Total: 16000 W: 3021 L: 3049 D: 9930
16000 @ 15+0.05 th 1 Bump razor depth down to 3 ply
13-03-06 staticnull5 diff
ELO: 3.58 +-5.4 (95%) LOS: 98.2%
Total: 16000 W: 3169 L: 3004 D: 9827
16000 @ 15+0.05 th 1 Bump static null depth up to 5
13-03-06 staticnull3 diff
ELO: -2.87 +-5.6 (95%) LOS: 5.3%
Total: 15000 W: 2870 L: 2994 D: 9136
16000 @ 15+0.05 th 1 Bump static null depth down to 3
13-03-06 non_pv_rule50 diff
ELO: -11.01 +-12.4 (95%) LOS: 0.2%
Total: 3000 W: 521 L: 616 D: 1863
16000 @ 15+0.05 th 1 Less aggressive rule50 reduction
13-03-06 pinned_null3 diff
ELO: 2.67 +-5.4 (95%) LOS: 93.9%
Total: 16000 W: 3212 L: 3089 D: 9699
16000 @ 15+0.05 th 1 Only worry about discovered check threats
13-03-07 max_time diff
ELO: 0.93 +-5.4 (95%) LOS: 72.5%
Total: 16000 W: 2598 L: 2555 D: 10847
16000 @ 60+0.05 th 1 Little less aggressive trying to finish iteration
13-03-07 easy_move diff
ELO: 3.04 +-5.4 (95%) LOS: 97.2%
Total: 16000 W: 2756 L: 2616 D: 10628
16000 @ 60+0.05 th 1 Re-test easy move vs current master
13-03-08 pinned_null3 diff
ELO: -0.02 +-5.4 (95%) LOS: 49.5%
Total: 16000 W: 3083 L: 3084 D: 9833
16000 @ 15+0.05 th 1 Try excluding discovered checks in static null pruning
13-03-08 pinned_null3 diff
ELO: -2.61 +-5.4 (95%) LOS: 6.8%
Total: 16000 W: 3166 L: 3286 D: 9548
16000 @ 15+0.05 th 1 Reduce depth we care about discovered checks
13-03-08 pinned_null diff
ELO: -3.17 +-5.4 (95%) LOS: 3.1%
Total: 16000 W: 2969 L: 3115 D: 9916
16000 @ 15+0.05 th 1 Try lower depth on original version too
13-03-09 max_time diff
ELO: 4.89 +-5.4 (95%) LOS: 99.9%
Total: 16000 W: 2700 L: 2475 D: 10825
16000 @ 60+0.05 th 1 Even more aggressive tweak: previous tests show we have to increase the parameter
13-03-10 easy_move_once diff
ELO: 3.21 +-5.4 (95%) LOS: 97.9%
Total: 16000 W: 2745 L: 2597 D: 10658
16000 @ 60+0.05 th 1 Let's see if the good in previous easy move patch is due to this
13-03-10 staticnull5 diff
ELO: 0.96 +-5.4 (95%) LOS: 72.6%
Total: 16000 W: 2691 L: 2647 D: 10662
16000 @ 60+0.05 th 1 Retest static null depth at 5 with longer TC
13-03-11 combined diff
ELO: -1.74 +-5.4 (95%) LOS: 15.5%
Total: 16000 W: 3075 L: 3155 D: 9770
16000 @ 15+0.05 th 1 pinned_null3, see_prune2 and staticnull5
13-03-11 combined2 diff
ELO: -2.19 +-5.4 (95%) LOS: 10.0%
Total: 16000 W: 3065 L: 3166 D: 9769
16000 @ 15+0.05 th 1 pinned_null, see_prune2 and staticnull5
13-03-11 nullpv diff
ELO: 0.67 +-12.2 (95%) LOS: 56.8%
Total: 3117 W: 621 L: 615 D: 1881
16000 @ 15+0.05 th 1 Try null moves in PV
13-03-11 probcut diff
ELO: -4.86 +-5.4 (95%) LOS: 0.2%
Total: 16000 W: 2946 L: 3170 D: 9884
16000 @ 15+0.05 th 1 Bump up probcut depth
13-03-11 probcut diff
ELO: -1.85 +-5.4 (95%) LOS: 13.9%
Total: 16000 W: 3026 L: 3111 D: 9863
16000 @ 15+0.05 th 1 Bump down probcut depth
13-03-11 probcut diff
ELO: -3.39 +-5.4 (95%) LOS: 2.3%
Total: 16000 W: 2964 L: 3120 D: 9916
16000 @ 15+0.05 th 1 Increase rBeta for probcut