Finished
PeregrRecklesscorrect-singulardiff8.0+0.08
LLR: -2.34 (-2.25, 2.89) [0.00, 3.00]
Games: 44482 W: 11055 L: 11095 D: 22332
Ptnml(0-2): 83, 5338, 11436, 5304, 80
Don't correct if the position is insanely singular (assume singularity is the enemy of generalization)
PeregrRecklesscorrection-pawns-onlydiff8.0+0.08
LLR: -1.36 (-2.25, 2.89) [0.00, 3.00]
Games: 4480 W: 1061 L: 1127 D: 2292
Ptnml(0-2): 8, 512, 1264, 450, 6
Basically assume that any correction to pawns only endgames is wrong and not needed
PeregrRecklessrfp-quietdiff8.0+0.08
LLR: -2.27 (-2.25, 2.89) [0.00, 3.00]
Games: 70010 W: 17391 L: 17371 D: 35248
Ptnml(0-2): 158, 8314, 18018, 8380, 135
PeregrReckless[Stockfish] openbench-pgodiff40.0+0.40
Elo: -49.19 +- 3.04 (95%) [N=40000]
Games: 13090 W: 2430 L: 4271 D: 6389
Ptnml(0-2): 35, 2658, 3012, 793, 47
LTC vs SF dev
PeregrRecklessfauzi-cntdiff8.0+0.08
LLR: 2.93 (-2.25, 2.89) [0.00, 3.00]
Games: 21060 W: 5369 L: 5151 D: 10540
Ptnml(0-2): 38, 2418, 5422, 2592, 60
one-shot my idea in SF to use allnode in ctfcnt + fauzi new separative tweak
PeregrRecklessreturn-mate-pcdiff8.0+0.08
LLR: -0.11 (-2.25, 2.89) [0.00, 3.00]
Games: 59750 W: 15017 L: 14893 D: 29840
Ptnml(0-2): 140, 7012, 15450, 7130, 143
Since we no longer do probcut at pvnodes, return mates from probcut as long as it doesn't hurt finding best mates or gains Elo
PeregrRecklessnegative-guarddiff40.0+0.40
LLR: -0.25 (-2.25, 2.89) [0.00, 3.00]
Games: 41294 W: 10035 L: 9965 D: 21294
Ptnml(0-2): 18, 4429, 11682, 4501, 17
spec LTC, try reverse
PeregrRecklessnegative-no-guarddiff40.0+0.40
LLR: -2.25 (-2.25, 2.89) [0.00, 3.00]
Games: 31328 W: 7567 L: 7627 D: 16134
Ptnml(0-2): 6, 3574, 8565, 3512, 7
[Rebased LTC] Do not guard TT moves with negative extensions against further reduction…
PeregrRecklessnegative-no-guarddiff8.0+0.08
LLR: 2.96 (-2.25, 2.89) [0.00, 3.00]
Games: 101084 W: 25581 L: 25187 D: 50316
Ptnml(0-2): 242, 11790, 26063, 12226, 221
Do not guard TT moves with negative extensions against further reduction.
PeregrRecklessrecap-failed-pcdiff8.0+0.08
LLR: -2.31 (-2.25, 2.89) [0.00, 3.00]
Games: 24994 W: 6174 L: 6255 D: 12565
Ptnml(0-2): 58, 2999, 6469, 2908, 63
PeregrRecklesslmr-failed-pc2diff8.0+0.08
LLR: -2.30 (-2.25, 2.89) [0.00, 3.00]
Games: 29906 W: 7441 L: 7511 D: 14954
Ptnml(0-2): 71, 3634, 7604, 3582, 62
PeregrRecklesslmr-failed-pcdiff8.0+0.08
LLR: -2.26 (-2.25, 2.89) [0.00, 3.00]
Games: 120860 W: 30421 L: 30289 D: 60150
Ptnml(0-2): 290, 14304, 31170, 14316, 350
PeregrRecklesslate-capture-gooddiff8.0+0.08
LLR: -2.26 (-2.25, 2.89) [0.00, 3.00]
Games: 50022 W: 12491 L: 12514 D: 25017
Ptnml(0-2): 102, 5935, 12969, 5894, 111
PeregrRecklessnoisy-pcm2diff8.0+0.08
LLR: 2.91 (-2.25, 2.89) [0.00, 3.00]
Games: 110686 W: 27804 L: 27394 D: 55488
Ptnml(0-2): 238, 12869, 28698, 13321, 217
PeregrRecklessnoisy-pcm3diff8.0+0.08
LLR: -2.26 (-2.25, 2.89) [0.00, 3.00]
Games: 22728 W: 5649 L: 5731 D: 11348
Ptnml(0-2): 51, 2684, 5968, 2618, 43
PeregrRecklessnoisy-pcmdiff8.0+0.08
LLR: -1.21 (-2.25, 2.89) [0.00, 3.00]
Games: 18690 W: 4593 L: 4623 D: 9474
Ptnml(0-2): 40, 2218, 4855, 2196, 36
PeregrRecklessreverse-iirdiff8.0+0.08
LLR: -2.29 (-2.25, 2.89) [0.00, 3.00]
Games: 100618 W: 25056 L: 24971 D: 50591
Ptnml(0-2): 204, 11868, 26036, 12041, 160
PeregrRecklesssplit-iirdiff8.0+0.08
LLR: -2.29 (-2.25, 2.89) [0.00, 3.00]
Games: 85184 W: 21323 L: 21271 D: 42590
Ptnml(0-2): 179, 10112, 21968, 10144, 189
PeregrRecklessiir-first-movediff8.0+0.08
LLR: -2.28 (-2.25, 2.89) [0.00, 3.00]
Games: 11660 W: 2867 L: 2977 D: 5816
Ptnml(0-2): 31, 1469, 2938, 1363, 29
no IIR for the first move
PeregrRecklessfaillow-consistentdiff8.0+0.08
LLR: -2.26 (-2.25, 2.89) [0.00, 3.00]
Games: 15474 W: 3841 L: 3938 D: 7695
Ptnml(0-2): 27, 1868, 4040, 1779, 23
Only if fail-low is consistent across all of the previous occasions 88%
PeregrRecklessiir-consistent-upperdiff8.0+0.08
LLR: -2.35 (-2.25, 2.89) [0.00, 3.00]
Games: 8580 W: 2073 L: 2192 D: 4315
Ptnml(0-2): 31, 1051, 2231, 960, 17
Make all iir'able logic about the fact that this position consistently keeps getting an upper bound
PeregrRecklesscutnode-no-entry2diff8.0+0.08
LLR: -0.93 (-2.25, 2.89) [0.00, 3.00]
Games: 2500 W: 590 L: 639 D: 1271
Ptnml(0-2): 5, 323, 641, 278, 3
Assume duplicate effect here if cutnode has no ttmove
PeregrRecklesscutnode-no-entrydiff8.0+0.08
LLR: -2.31 (-2.25, 2.89) [0.00, 3.00]
Games: 7052 W: 1686 L: 1805 D: 3561
Ptnml(0-2): 27, 858, 1858, 773, 10
Insist on finding a move that cuts in (expected) cutnodes
PeregrRecklesslmr-pv2diff40.0+0.40
LLR: -2.05 (-2.25, 2.89) [0.00, 3.00]
Games: 36642 W: 8901 L: 8939 D: 18802
Ptnml(0-2): 11, 4203, 9928, 4171, 8
[LTC] Try to avoid interfering with the fact that depth was already decremented.
PeregrRecklessmaybe-thisdiff8.0+0.08
LLR: -2.26 (-2.25, 2.89) [0.00, 3.00]
Games: 10020 W: 2433 L: 2541 D: 5046
Ptnml(0-2): 15, 1227, 2631, 1125, 12
Restore the original depth for TT and history updates after it was decremented.