Stockfish Testing Queue

Finished - 30366 tests

13-09-11 mco unstoppable diff
LLR: 2.96 (-2.94,2.94) [-1.50,4.50]
Total: 16927 W: 3462 L: 3308 D: 10157
sprt @ 15+0.05 th 1 Unstoppable: Like the not bad second attempt but with reduced bonus
13-09-11 jos kingRing diff
LLR: -2.95 (-2.94,2.94) [-1.50,4.50]
Total: 17551 W: 3598 L: 3656 D: 10297
sprt @ 15+0.05 th 1 Expand kingRing to 2 squares around the King.
13-09-11 rap rook_pawns diff
LLR: -2.95 (-2.94,2.94) [0.00,6.00]
Total: 22295 W: 3803 L: 3797 D: 14695
sprt @ 60+0.05 th 1 Bonus to extreme pawn when we have more material. Same bench due to few hits.
13-09-11 gli master diff
ELO: 24.87 +-2.9 (95%) LOS: 100.0%
Total: 20000 W: 4253 L: 2824 D: 12923
20000 @ 60+0.05 th 1 Regression test after Uri and Marco check extension changes
13-09-11 gli uri_less_lmr diff
LLR: -2.95 (-2.94,2.94) [-1.50,4.50]
Total: 1516 W: 247 L: 349 D: 920
sprt @ 15+0.05 th 1 Simplified version of Uri's LMR condition
13-09-11 uri uri_less_reduce diff
LLR: -2.95 (-2.94,2.94) [-1.50,4.50]
Total: 3107 W: 541 L: 638 D: 1928
sprt @ 15+0.05 th 1 try one of not reducing extended move by a full ply
13-09-11 uri uri_less_reduce diff
LLR: -2.96 (-2.94,2.94) [-1.50,4.50]
Total: 3206 W: 613 L: 712 D: 1881
sprt @ 15+0.05 th 1 this time not reduce checking move that are extended by a full ply(the difference is that I reduce singular extended moves)
13-09-11 uri uri_less_reduce diff
LLR: -2.95 (-2.94,2.94) [-1.50,4.50]
Total: 8257 W: 1599 L: 1683 D: 4975
sprt @ 15+0.05 th 1 this time very similiar condition to the condition that I extend one ply. The only difference is that I may extend some singular move that is also checking with negative SEE so this move can be reduced.
13-09-11 mco unstoppable diff
Pending...
sprt @ 60+0.05 th 1 LTC Unstoppable: Like the not bad second attempt but with reduced bonus
13-09-12 lbr reduction diff
LLR: -2.96 (-2.94,2.94) [-1.50,4.50]
Total: 7020 W: 1379 L: 1467 D: 4174
sprt @ 15+0.05 th 1 extend checks by one ply less. rebase on new master.
13-09-12 gli uri_less_lmr diff
LLR: -2.95 (-2.94,2.94) [0.00,6.00]
Total: 12489 W: 2102 L: 2144 D: 8243
sprt @ 60+0.05 th 1 Long TC: Uri LMR condition take 2
13-09-12 mco 490f67a3f89449e243c3 diff
ELO: -1.53 +-3.0 (95%) LOS: 16.0%
Total: 20000 W: 3885 L: 3973 D: 12142
20000 @ 15+0.05 th 1 Regression test "Move draw by material check"
13-09-12 mco 49e110c52b19f5c diff
ELO: -2.97 +-3.0 (95%) LOS: 2.8%
Total: 20000 W: 3898 L: 4069 D: 12033
20000 @ 15+0.05 th 1 Regression test "Fix random moves when time < 10ms"
13-09-12 mco unstoppable diff
LLR: -2.95 (-2.94,2.94) [0.00,6.00]
Total: 8451 W: 1386 L: 1448 D: 5617
sprt @ 60+0.05 th 1 LTC Unstoppable: Like the not bad second attempt but with reduced bonus
13-09-12 mco master diff
ELO: 24.34 +-2.8 (95%) LOS: 100.0%
Total: 20000 W: 4224 L: 2825 D: 12951
20000 @ 60+0.05 th 1 Regression test with dubious patches reverted
13-09-12 pec uri_ext diff
LLR: 0.13 (-2.94,2.94) [0.00,6.00]
Total: 16275 W: 2858 L: 2773 D: 10644
sprt @ 60+0.05 th 1 Check extension: Compare with alpha instead of VALUE_ZERO. Repeat run: check to see if original run was an outlier. If it was outlier this run is expected to fail in approximately the same number of games as original run passed ( < 10000 )
13-09-12 rap passed_B diff
LLR: -1.76 (-2.94,2.94) [-1.50,4.50]
Total: 85479 W: 17114 L: 16942 D: 51423
sprt @ 15+0.05 th 1 Bonus for passed pawns when up in material.
13-09-12 rap passed_c diff
LLR: 2.95 (-2.94,2.94) [-1.50,4.50]
Total: 36463 W: 7575 L: 7365 D: 21523
sprt @ 15+0.05 th 1 Increase bonus for passed pawns in case we have more pieces.
13-09-12 rap passed_D diff
LLR: -2.95 (-2.94,2.94) [-1.50,4.50]
Total: 47766 W: 9524 L: 9499 D: 28743
sprt @ 15+0.05 th 1 MG bonus for supported passed pawn
13-09-12 jos kingRing2 diff
LLR: 2.96 (-2.94,2.94) [-1.50,4.50]
Total: 32092 W: 6593 L: 6396 D: 19103
sprt @ 15+0.05 th 1 Expand kingRing only sideways. Last try.
13-09-12 gli master diff
ELO: 24.87 +-2.9 (95%) LOS: 100.0%
Total: 20000 W: 4273 L: 2844 D: 12883
20000 @ 60+0.05 th 1 (Rescheduled +28 elo regression test to see if it was outlier) Regression test after singular extension 8 plies
13-09-13 jos kingRing2 diff
LLR: -2.96 (-2.94,2.94) [0.00,6.00]
Total: 14792 W: 2508 L: 2539 D: 9745
sprt @ 60+0.05 th 1 LTC: Expand kingRing only sideways. Last try.
13-09-13 mco passed_c diff
LLR: 2.97 (-2.94,2.94) [0.00,6.00]
Total: 2953 W: 564 L: 446 D: 1943
sprt @ 60+0.05 th 1 LTC: Increase bonus for passed pawns in case we have more pieces.
13-09-13 uri less_reduce_uri diff
LLR: -2.96 (-2.94,2.94) [-1.50,4.50]
Total: 11956 W: 2374 L: 2448 D: 7134
sprt @ 15+0.05 th 1 simply reduce uri_extension moves by one ply. Maybe I could write it in less lines but I did not like to xor a lot of conditions (move give check, check is safe and evaluation<alpha)
13-09-13 uri reduce_singular_extensi diff
LLR: -2.95 (-2.94,2.94) [-1.50,4.50]
Total: 4711 W: 910 L: 1004 D: 2797
sprt @ 15+0.05 th 1 reduce singular extensions to half ply
13-09-13 uri more_late diff
LLR: -2.97 (-2.94,2.94) [-1.50,4.50]
Total: 9924 W: 1936 L: 2016 D: 5972
sprt @ 15+0.05 th 1 late move reduction does not drop relative to qsearch when the remaining depth is 1 and I try changing it
13-09-13 uri limit_ext diff
LLR: -2.96 (-2.94,2.94) [-1.50,4.50]
Total: 6346 W: 1272 L: 1362 D: 3712
sprt @ 15+0.05 th 1 do not extend in plies that are after the nominal depth(if stockfish search to depth 20 it means do not extend lines from ply 21)
13-09-13 uri limit_ext diff
LLR: -2.96 (-2.94,2.94) [-1.50,4.50]
Total: 9526 W: 1872 L: 1953 D: 5701
sprt @ 15+0.05 th 1 less risky try to limit extensions and this time I try limiting it to plies that are not after initial_depth*3/2 meaning that if you search to depth 20 you do not extend after ply 30
13-09-13 mco master diff
ELO: 26.21 +-2.9 (95%) LOS: 100.0%
Total: 20000 W: 4377 L: 2871 D: 12752
20000 @ 60+0.05 th 1 Regression test with latest passed bonus patch
13-09-13 mco 4803d5772c120121da diff
LLR: 2.96 (-2.94,2.94) [-1.50,4.50]
Total: 12501 W: 2583 L: 2441 D: 7477
sprt @ 15+0.05 th 1 Retest new check extensions as a single patch
13-09-13 uri passed_pawn_change diff
LLR: -2.96 (-2.94,2.94) [-1.50,4.50]
Total: 2849 W: 540 L: 640 D: 1669
sprt @ 15+0.05 th 1 changed the formula for passed pawn bonus and now there is also less bonus for having less pieces and the exact number of pieces is also important.
13-09-13 jki cext diff
LLR: 2.95 (-2.94,2.94) [-1.50,4.50]
Total: 9849 W: 2045 L: 1911 D: 5893
sprt @ 15+0.05 th 1 Always extend checks full ply when previous move was not extended a full ply
13-09-13 jki cext diff
LLR: 2.95 (-2.94,2.94) [-1.50,4.50]
Total: 11477 W: 2337 L: 2199 D: 6941
sprt @ 15+0.05 th 1 One more try
13-09-13 uri limit_ext1 diff
LLR: -2.96 (-2.94,2.94) [-1.50,4.50]
Total: 6075 W: 1152 L: 1242 D: 3681
sprt @ 15+0.05 th 1 another try to limit extensions(to the first plies) and this time there is an exception for pv nodes.
13-09-13 mco 4803d5772c120121da diff
LLR: 2.95 (-2.94,2.94) [0.00,6.00]
Total: 19782 W: 3449 L: 3250 D: 13083
sprt @ 60+0.05 th 1 LTC: Retest new check extensions as a single patch
13-09-13 mco cext diff
LLR: -2.95 (-2.94,2.94) [0.00,6.00]
Total: 12393 W: 2123 L: 2165 D: 8105
sprt @ 60+0.05 th 1 LTC: Always extend checks full ply when previous move was not extended a full ply
13-09-13 gli queen_only diff
LLR: -2.97 (-2.94,2.94) [-1.50,4.50]
Total: 12477 W: 2432 L: 2505 D: 7540
sprt @ 15+0.05 th 1 Only treat queen promotions specially
13-09-13 gli improving diff
LLR: -2.95 (-2.94,2.94) [-1.50,4.50]
Total: 7590 W: 1484 L: 1570 D: 4536
sprt @ 15+0.05 th 1 More aggressive LMR when in check
13-09-13 jki cext diff
LLR: -2.96 (-2.94,2.94) [0.00,6.00]
Total: 26567 W: 4511 L: 4485 D: 17571
sprt @ 60+0.05 th 1 LTC: One more try
13-09-14 lbr fut_pv diff
LLR: -2.97 (-2.94,2.94) [0.00,6.00]
Total: 1667 W: 298 L: 397 D: 972
sprt @ 15+0.05 th 1 futility pruning at PV nodes.
13-09-14 gli blocked diff
LLR: -2.97 (-2.94,2.94) [-1.50,4.50]
Total: 7835 W: 1570 L: 1656 D: 4609
sprt @ 15+0.05 th 1 Another try on blocked pawns
13-09-14 uri null_increase diff
LLR: -2.96 (-2.94,2.94) [-1.50,4.50]
Total: 7673 W: 1504 L: 1590 D: 4579
sprt @ 15+0.05 th 1 increase null move pruning also when the difference is more than half pawn from beta
13-09-14 uri reduce_promotions diff
LLR: -2.96 (-2.94,2.94) [-1.50,4.50]
Total: 11937 W: 2336 L: 2410 D: 7191
sprt @ 15+0.05 th 1 reduced the depth after promotions. idea is from my engine movei and I am going to try also a different formula of reducing the depth by 1
13-09-14 uri reduce_promotions diff
LLR: -2.95 (-2.94,2.94) [-1.50,4.50]
Total: 10340 W: 1976 L: 2054 D: 6310
sprt @ 15+0.05 th 1 this time I reduce the depth for promotion moves by one ply
13-09-14 mco search diff
LLR: 2.96 (-2.94,2.94) [-1.50,4.50]
Total: 26590 W: 5426 L: 5245 D: 15919
sprt @ 15+0.05 th 1 Don't blunder under extreme time pressure
13-09-14 eel outsidePasser diff
LLR: -2.95 (-2.94,2.94) [-1.50,4.50]
Total: 60457 W: 11858 L: 11800 D: 36799
sprt @ 15+0.05 th 1 Not locally tested this exact variation using r, something similar a little bit.
13-09-14 jki cext diff
LLR: -2.90 (-2.94,2.94) [-1.50,4.50]
Total: 8289 W: 1624 L: 1706 D: 4959
sprt @ 15+0.05 th 1 Almost always extend a full ply
13-09-14 jki cext diff
LLR: -2.97 (-2.94,2.94) [-1.50,4.50]
Total: 17861 W: 3567 L: 3625 D: 10669
sprt @ 15+0.05 th 1 Almost always extend a full ply II
13-09-14 mco search diff
LLR: 2.96 (-2.94,2.94) [-1.50,4.50]
Total: 5767 W: 1397 L: 1268 D: 3102
sprt @ 5+0.05 th 1 Super short TC: Don't blunder under extreme time pressure (verify the patch at 5 seconds per game)
13-09-14 mco search diff
LLR: -2.96 (-2.94,2.94) [0.00,6.00]
Total: 79862 W: 13623 L: 13339 D: 52900
sprt @ 60+0.05 th 1 LTC: Don't blunder under extreme time pressure