Finished
PeregrReckless[Stockfish] sf17diff40.0+0.40
Elo: -20.41 +- 3.38 (95%) [N=10000]
Games: 10020 W: 2247 L: 2835 D: 4938
Ptnml(0-2): 8, 1516, 2560, 908, 18
[Threads=1 Hash=64] vs. SF 17 with PGO
PeregrRecklessreplies-to-capturesdiff40.0+0.40
LLR: -2.31 (-2.25, 2.89) [0.00, 3.00]
Games: 24690 W: 6005 L: 6082 D: 12603
Ptnml(0-2): 4, 2859, 6698, 2778, 6
[LTC] Assume most of what matters against captures is the history of the quiet move against it
PeregrRecklessreplies-to-capturesdiff8.0+0.08
LLR: 2.93 (-2.25, 2.89) [0.00, 3.00]
Games: 170066 W: 42584 L: 42038 D: 85444
Ptnml(0-2): 411, 20030, 43641, 20504, 447
Assume most of what matters against captures is the history of the quiet move against it
PeregrRecklessreplies-to-captures2diff8.0+0.08
LLR: -1.37 (-2.25, 2.89) [0.00, 3.00]
Games: 8936 W: 2230 L: 2290 D: 4416
Ptnml(0-2): 22, 1063, 2362, 995, 26
in movepick-scoring, care less about the quiet history of the move if the position is very tactical and we need a good reply or a good follow-up of moves
PeregrRecklessfollowup-plan3diff8.0+0.08
LLR: -2.27 (-2.25, 2.89) [0.00, 3.00]
Games: 12058 W: 2911 L: 3017 D: 6130
Ptnml(0-2): 28, 1476, 3120, 1384, 21
Take 3
PeregrRecklessfollowup-plan2diff40.0+0.40
LLR: -1.19 (-2.25, 2.89) [0.00, 3.00]
Games: 159682 W: 38702 L: 38438 D: 82542
Ptnml(0-2): 43, 17939, 43626, 18177, 56
Appreciate deeper follow-ups only if they help negative reductions
PeregrRecklessfollowup-plan2diff8.0+0.08
LLR: 2.93 (-2.25, 2.89) [0.00, 3.00]
Games: 100810 W: 25195 L: 24804 D: 50811
Ptnml(0-2): 210, 11817, 25956, 12216, 206
Appreciate deeper follow-ups only if they help negative reductions
PeregrRecklessfollowup-plandiff8.0+0.08
LLR: -0.78 (-2.25, 2.89) [0.00, 3.00]
Games: 8376 W: 2054 L: 2082 D: 4240
Ptnml(0-2): 19, 1042, 2091, 1020, 16
PeregrRecklessnoisy-tiebreaksdiff40.0+0.40
LLR: -2.25 (-2.25, 2.89) [0.00, 3.00]
Games: 30426 W: 7382 L: 7445 D: 15599
Ptnml(0-2): 12, 3591, 8064, 3540, 6
PeregrRecklessnoisy-tiebreaksdiff8.0+0.08
LLR: 2.93 (-2.25, 2.89) [0.00, 3.00]
Games: 90814 W: 22783 L: 22410 D: 45621
Ptnml(0-2): 212, 10730, 23191, 11021, 253
break MVV tiebreaks with undefended vs defended
PeregrRecklessquiet-after-noisydiff8.0+0.08
LLR: -2.27 (-2.25, 2.89) [0.00, 3.00]
Games: 29050 W: 7230 L: 7299 D: 14521
Ptnml(0-2): 55, 3491, 7507, 3412, 60
Appreciate the quiet moves as counters to noisy moves more.
PeregrRecklesscutnode-ttpvdiff8.0+0.08
LLR: 2.91 (-2.25, 2.89) [-2.75, 0.25]
Games: 40594 W: 10260 L: 10162 D: 20172
Ptnml(0-2): 101, 4856, 10301, 4922, 117
PeregrRecklesscutnode-ttpvdiff40.0+0.40
LLR: 2.89 (-2.25, 2.89) [-3.00, 0.00]
Games: 50578 W: 12183 L: 12127 D: 26268
Ptnml(0-2): 17, 5724, 13759, 5764, 25
PeregrRecklessreverse-bonus-wrong2diff40.0+0.40
LLR: -2.30 (-2.25, 2.89) [0.00, 3.00]
Games: 11526 W: 2769 L: 2873 D: 5884
Ptnml(0-2): 6, 1364, 3126, 1262, 5
PeregrRecklessreverse-bonus-wrong2diff8.0+0.08
LLR: 2.96 (-2.25, 2.89) [0.00, 3.00]
Games: 172718 W: 43180 L: 42624 D: 86914
Ptnml(0-2): 424, 20515, 43985, 20951, 484
reverse our bonus about the potential singular move that turned out bad even when failing low, but based on its tt_depth
PeregrRecklessreverse-bonus-wrongdiff8.0+0.08
LLR: -2.33 (-2.25, 2.89) [0.00, 3.00]
Games: 7786 W: 1884 L: 2005 D: 3897
Ptnml(0-2): 20, 1005, 1962, 888, 18
reverse the bonus of the tt_move with non-upperbound even when failing low, based on its tt_depth
PeregrRecklessrelative-to-absolutediff8.0+0.08
LLR: -2.27 (-2.25, 2.89) [0.00, 3.00]
Games: 43990 W: 10964 L: 11001 D: 22025
Ptnml(0-2): 94, 5274, 11320, 5189, 118
Maintain the same hit-rate when penalizing previous move in early ttcut by using the absolute history value instead of the relative move_count.
PeregrRecklessunbias-ttmdiff8.0+0.08
LLR: -2.28 (-2.25, 2.89) [0.00, 3.00]
Games: 56026 W: 14018 L: 14029 D: 27979
Ptnml(0-2): 108, 6606, 14589, 6609, 101
tt_moves may already have poor history, so exclude them from the penalty
PeregrRecklesslate-bestdiff40.0+0.40
LLR: -2.30 (-2.25, 2.89) [0.00, 3.00]
Games: 33788 W: 8248 L: 8305 D: 17235
Ptnml(0-2): 10, 3802, 9327, 3745, 10
[LTC] account for lateness in bonus
PeregrReckless[Stockfish] sf17diff40.0+0.40
Elo: -30.60 +- 5.28 (95%) [N=10000]
Games: 4064 W: 837 L: 1194 D: 2033
Ptnml(0-2): 4, 672, 1045, 299, 12
latest main, [Threads=1 Hash=64] vs. SF 17 with PGO
PeregrRecklesslate-bestdiff8.0+0.08
LLR: 2.96 (-2.25, 2.89) [0.00, 3.00]
Games: 69056 W: 17307 L: 16983 D: 34766
Ptnml(0-2): 139, 8057, 17862, 8281, 189
account for lateness in bonus
PeregrRecklessstandpat-no-confidentdiff40.0+0.40
LLR: 2.90 (-2.25, 2.89) [0.00, 3.00]
Games: 56254 W: 13659 L: 13382 D: 29213
Ptnml(0-2): 18, 6220, 15375, 6495, 19
PeregrRecklessstandpat-no-confidentdiff8.0+0.08
LLR: 2.95 (-2.25, 2.89) [0.00, 3.00]
Games: 147786 W: 36870 L: 36374 D: 74542
Ptnml(0-2): 326, 17387, 37963, 17899, 318
PeregrRecklessreverse-qs-pvdiff40.0+0.40
LLR: -2.29 (-2.25, 2.89) [0.00, 3.00]
Games: 21658 W: 5152 L: 5234 D: 11272
Ptnml(0-2): 6, 2499, 5901, 2417, 6
PeregrRecklessfix-reverse-qsdiff40.0+0.40
LLR: -0.55 (-2.25, 2.89) [-3.00, 0.00]
Games: 54254 W: 13204 L: 13334 D: 27716
Ptnml(0-2): 12, 5167, 16899, 5037, 12
[25% TP] bench is different at depth 22, fixes the assert triggered because of the reverse qsearch logic passing always cutnode as true