OpenBench Testing Framework

Finished
don-t-allow-more-than-3-reductions-in-aspiration
8.0+0.08
Elo: -0.48 +- 2.40 [0.00, 5.00]
Games: 15316 W: 3566 L: 3587 D: 8163
Ptnml(0-2): 33, 1333, 4950, 1306, 36
don't allow more than 3 reductions in aspiration
thread-voting
8.0+0.08
Elo: 1.16 +- 2.71 [0.00, 5.00]
Games: 14984 W: 3381 L: 3331 D: 8272
Ptnml(0-2): 12, 1698, 4023, 1746, 13
Add thread voting After finishing search, have each thread vote for its own best move. Then select the best result. Based on the Beserk impl which also seems to be used in Stockfish, Stormphrax, etc. vote
thread-voting
N=40000
Elo: 33.58 +- 10.85 [0.00, 5.00]
Games: 1318 W: 403 L: 276 D: 639
Ptnml(0-2): 12, 102, 317, 203, 25
thread-voting
N=20000
Elo: 44.05 +- 13.26 [0.00, 5.00]
Games: 1134 W: 378 L: 235 D: 521
Ptnml(0-2): 16, 92, 231, 189, 39
thread-voting
N=40000
Elo: 13.48 +- 6.73 [0.00, 5.00]
Games: 3482 W: 977 L: 842 D: 1663
Ptnml(0-2): 34, 374, 807, 475, 51
thread-voting
N=20000
Elo: 29.95 +- 10.40 [0.00, 5.00]
Games: 1628 W: 516 L: 376 D: 736
Ptnml(0-2): 19, 152, 348, 260, 35
debug
N=5000
Elo: -0.00 +- 0.00 (95%) [N=40000]
Games: 1182 W: 400 L: 400 D: 382
Ptnml(0-2): 0, 0, 591, 0, 0
don't overflow depth calc in TT
debug
N=20000
Elo: -0.00 +- 0.00 (95%) [N=100000]
Games: 752 W: 230 L: 230 D: 292
Ptnml(0-2): 0, 0, 376, 0, 0
debug
debug *
N=20000
Elo: -0.00 +- 0.00 [0.00, 5.00]
Games: 746 W: 219 L: 219 D: 308
Ptnml(0-2): 0, 0, 373, 0, 0
debug
thread-voting
5.0+0.05
Elo: 0.97 +- 4.44 [0.00, 5.00]
Games: 5714 W: 1298 L: 1282 D: 3134
Ptnml(0-2): 7, 657, 1512, 675, 6
time-management-tune
40.0+0.40
Elo: -7.05 +- 13.29 [0.00, 5.00]
Games: 542 W: 110 L: 121 D: 311
Ptnml(0-2): 1, 58, 163, 49, 0
tune results
c236c7fe
4.0+0.04
Elo: 0.80 +- 2.64 [-5.00, 0.00]
Games: 17442 W: 4197 L: 4157 D: 9088
Ptnml(0-2): 83, 1935, 4650, 1965, 88
quick non-reg of time control changes for single thread
thread-time-management
5.0+0.05
Elo: 5.69 +- 4.73 [-5.00, 0.00]
Games: 4946 W: 1131 L: 1050 D: 2765
Ptnml(0-2): 3, 534, 1322, 607, 7
thread time management
qsearch-in-check
20.0+0.20
Elo: -8.17 +- 5.88 [-5.00, 0.00]
Games: 3362 W: 760 L: 839 D: 1763
Ptnml(0-2): 5, 440, 871, 359, 6
qsearch in check
upcoming-repetition-detection
20.0+0.20
Elo: 0.88 +- 2.20 [0.00, 5.00]
Games: 23616 W: 5288 L: 5228 D: 13100
Ptnml(0-2): 41, 2717, 6235, 2771, 44
upcoming repetition detection
spsa-2026-06-30
40.0+0.40
Elo: 5.58 +- 3.78 [0.00, 5.00]
Games: 8038 W: 1853 L: 1724 D: 4461
Ptnml(0-2): 6, 897, 2095, 1004, 17
spsa 2026-06-30
master
40.0+0.40
Elo: 62.06 +- 10.10 (95%) [N=1024]
Games: 1024 W: 342 L: 161 D: 521
Ptnml(0-2): 0, 41, 256, 208, 7
Progtest
master
40.0+0.40
Elo: 10.64 +- 20.50 (95%) [N=1024]
Games: 294 W: 63 L: 54 D: 177
Ptnml(0-2): 0, 34, 70, 43, 0
Add quiescence futility pruning Elo | 4.32 +- 3.18 (95%) SPRT | 20.0+0.20s Threads=1 Hash=16MB LLR | 2.91 (-2.25, 2.89) [0.00, 5.00] Games | N: 11338 W: 2634 L: 2493 D: 6211 Penta | [13, 1270, 2971, 1393, 22]
quiescence-futility-pruning
20.0+0.20
Elo: 4.32 +- 3.18 [0.00, 5.00]
Games: 11338 W: 2634 L: 2493 D: 6211
Ptnml(0-2): 13, 1270, 2971, 1393, 22
quiescence futility pruning
ranger-net
8.0+0.08
Elo: 3.54 +- 2.80 [0.00, 5.00]
Games: 17288 W: 4311 L: 4135 D: 8842
Ptnml(0-2): 104, 2025, 4204, 2213, 98
ranger net
ranger-net
N=25000
Elo: 3.59 +- 2.83 [0.00, 5.00]
Games: 25428 W: 7689 L: 7426 D: 10313
Ptnml(0-2): 592, 3003, 5352, 3084, 683
ranger net
spsa-2026-06-30
8.0+0.08
Elo: 1.51 +- 4.71 [0.00, 5.00]
Games: 5764 W: 1377 L: 1352 D: 3035
Ptnml(0-2): 23, 686, 1441, 707, 25
spsa 2026-06-30
spsa
20.0+0.20
Tuning 47 Parameters
2682/25000 Iterations
42924/400000 Games Played
time-management-tune
40.0+0.40
Tuning 12 Parameters
305/25000 Iterations
4892/400000 Games Played
spsa
20.0+0.20
Tuning 57 Parameters
216/25000 Iterations
3462/400000 Games Played