FuryBench Testing Framework

home toIntegralPlentyChessObsidianCaissaQuanticadeSerendipitySpaghetSchoenemannClarityIlluminaPawnocchioVineAnura
No logs for event exist
Active : 11 Machines / 166 Threads / 190.04 MNPS
Priority 0
Integral Fury
razoring_all_node
diff10.0+0.10
ELO    1.10 +- 1.55 (95%)
LLR    0.84 (-2.25, 2.89) [0.00, 3.00]
GAMES  N: 48268 W: 12063 L: 11910 D: 24295
PENTA  [102, 5585, 12604, 5744, 99]
Caissa Witek
master
New neural net (eval-ml-8-152B-spsa) LTC: ELO 6.09 +- 3.21 (95%) SPRT 40.0+0.40s Threads=1 Hash=64MB LLR 2.95 (-2.94, 2.94) [0.00, 3.00] GAMES N: 10950 W: 2664 L: 2472 D: 5814 PENTA [11, 1188, 2894, 1362, 20] https://furybench.com/test/7437/ STC: ELO 5.97 +- 2.73 (95%) SPRT 8.0+0.08s Threads=1 Hash=16MB LLR 2.95 (-2.94, 2.94) [0.00, 2.00] GAMES N: 17164 W: 4182 L: 3887 D: 9095 PENTA [67, 1926, 4330, 2163, 96] https://furybench.com/test/7436/ Net was created by SPSA tuning previous net (eval-ml-8-152B-permuted) for 6769 iterations (108304 games) at 6+0.06 time control: https://furybench.com/tune/7435/ Only buckets 1-7 were tuned.
diff60.0+0.60
ELO    39.91 +- 6.79 (95%)
GAMES  N: 2632/40000 W: 770 L: 469 D: 1393
PENTA  [2, 197, 629, 474, 14]
Illumina Mergener
remove-redundant-check
Remove redundant move legality check from qsearch
diff8.0+0.08
ELO    -0.85 +- 2.31 (95%)
LLR    -0.05 (-2.25, 2.89) [-3.00, 0.00]
GAMES  N: 25230 W: 5577 L: 5639 D: 14014
PENTA  [150, 3079, 6221, 3013, 152]
Quanticade DarkNeutrino
noisy-triple
diff60.0+0.60
ELO    -0.26 +- 2.27 (95%)
LLR    -1.02 (-2.25, 2.89) [0.00, 3.00]
GAMES  N: 19930 W: 4674 L: 4689 D: 10567
PENTA  [5, 2198, 5576, 2179, 7]
Priority -1
Quanticade DarkNeutrino
master
25K node datagen.
diffN=25000
ELO    -0.11 +- 0.15 (95%)
GAMES  N: 14710098/30000000 W: 5160123 L: 5164758 D: 4385217
Quanticade DarkNeutrino
hse-tuned
crash smp diff20.0+0.20
ELO    -0.00 +- 3.18 (95%)
LLR    -0.36 (-2.25, 2.89) [0.00, 3.00]
GAMES  N: 9688 W: 2217 L: 2217 D: 5254
PENTA  [3, 1010, 2816, 1014, 1]
Illumina Mergener
no-gives-check-condition-in-se
No check extensions during singular extensions
diff8.0+0.08
ELO    -4.37 +- 9.06 (95%)
LLR    -0.73 (-2.25, 2.89) [0.00, 5.00]
GAMES  N: 1668 W: 348 L: 369 D: 951
PENTA  [11, 211, 412, 188, 12]
Finished
Caissa Witek
master
New neural net (eval-ml-8-152B-spsa) LTC: ELO 6.09 +- 3.21 (95%) SPRT 40.0+0.40s Threads=1 Hash=64MB LLR 2.95 (-2.94, 2.94) [0.00, 3.00] GAMES N: 10950 W: 2664 L: 2472 D: 5814 PENTA [11, 1188, 2894, 1362, 20] https://furybench.com/test/7437/ STC: ELO 5.97 +- 2.73 (95%) SPRT 8.0+0.08s Threads=1 Hash=16MB LLR 2.95 (-2.94, 2.94) [0.00, 2.00] GAMES N: 17164 W: 4182 L: 3887 D: 9095 PENTA [67, 1926, 4330, 2163, 96] https://furybench.com/test/7436/ Net was created by SPSA tuning previous net (eval-ml-8-152B-permuted) for 6769 iterations (108304 games) at 6+0.06 time control: https://furybench.com/tune/7435/ Only buckets 1-7 were tuned.
diff10.0+0.10
ELO    29.15 +- 4.59 (95%)
GAMES  N: 6284/40000 W: 1814 L: 1288 D: 3182
PENTA  [25, 550, 1478, 1052, 37]
Quanticade DarkNeutrino
pv-triple
diff60.0+0.60
ELO    -2.30 +- 2.67 (95%)
LLR    -2.27 (-2.25, 2.89) [0.00, 3.00]
GAMES  N: 14212 W: 3237 L: 3331 D: 7644
PENTA  [1, 1588, 4024, 1490, 3]
Caissa Witek
eval-ml-8-152B-spsa
New neural net (eval-ml-8-152B) LTC: ELO 7.32 +- 3.54 (95%) SPRT 60.0+0.60s Threads=1 Hash=128MB LLR 2.98 (-2.94, 2.94) [0.00, 3.00] GAMES N: 9072 W: 2217 L: 2026 D: 4829 PENTA [8, 990, 2348, 1183, 7] https://furybench.com/test/7428/ STC: ELO 1.58 +- 1.19 (95%) SPRT 10.0+0.10s Threads=1 Hash=16MB LLR 3.01 (-2.94, 2.94) [0.00, 2.00] GAMES N: 91412 W: 22159 L: 21744 D: 47509 PENTA [420, 10869, 22705, 11300, 412] https://furybench.com/test/7427/ 25k fixed nodes (local test): Elo: 22.73 +/- 5.65, nElo: 33.53 +/- 8.31 LOS: 100.00 %, DrawRatio: 41.37 %, PairsRatio: 1.41 Games: 6720, Wins: 2292, Losses: 1853, Draws: 2575, Points: 3579.5 (53.27 %) Ptnml(0-2): [141, 678, 1390, 903, 248], WL/DD Ratio: 1.80
diff40.0+0.40
ELO    6.09 +- 3.21 (95%)
LLR    2.95 (-2.94, 2.94) [0.00, 3.00]
GAMES  N: 10950 W: 2664 L: 2472 D: 5814
PENTA  [11, 1188, 2894, 1362, 20]
Caissa Witek
eval-ml-8-152B-spsa
New neural net (eval-ml-8-152B) LTC: ELO 7.32 +- 3.54 (95%) SPRT 60.0+0.60s Threads=1 Hash=128MB LLR 2.98 (-2.94, 2.94) [0.00, 3.00] GAMES N: 9072 W: 2217 L: 2026 D: 4829 PENTA [8, 990, 2348, 1183, 7] https://furybench.com/test/7428/ STC: ELO 1.58 +- 1.19 (95%) SPRT 10.0+0.10s Threads=1 Hash=16MB LLR 3.01 (-2.94, 2.94) [0.00, 2.00] GAMES N: 91412 W: 22159 L: 21744 D: 47509 PENTA [420, 10869, 22705, 11300, 412] https://furybench.com/test/7427/ 25k fixed nodes (local test): Elo: 22.73 +/- 5.65, nElo: 33.53 +/- 8.31 LOS: 100.00 %, DrawRatio: 41.37 %, PairsRatio: 1.41 Games: 6720, Wins: 2292, Losses: 1853, Draws: 2575, Points: 3579.5 (53.27 %) Ptnml(0-2): [141, 678, 1390, 903, 248], WL/DD Ratio: 1.80
diff8.0+0.08
ELO    5.97 +- 2.73 (95%)
LLR    2.95 (-2.94, 2.94) [0.00, 2.00]
GAMES  N: 17164 W: 4182 L: 3887 D: 9095
PENTA  [67, 1926, 4330, 2163, 96]
Integral Fury
se_margin_history
diff40.0+0.40
ELO    4.04 +- 2.52 (95%)
LLR    2.89 (-2.25, 2.89) [0.00, 3.00]
GAMES  N: 16866 W: 4196 L: 4000 D: 8670
PENTA  [21, 1760, 4680, 1946, 26]
Caissa Witek
spsa-l3-tunables
diff6.0+0.06
TUNE   231 parameters
ITER   6,769 / 20,000
GAMES  108,304 / 320,000
Caissa Witek
net_1536
Neural net: Increase accumulator size to 1536
diff20.0+0.10
ELO    1.19 +- 1.95 (95%)
LLR    0.68 (-2.94, 2.94) [0.00, 2.00]
GAMES  N: 26918 W: 6776 L: 6684 D: 13458
PENTA  [9, 2900, 7548, 2994, 8]
Integral Fury
se_margin_history
diff10.0+0.10
ELO    2.12 +- 1.65 (95%)
LLR    2.90 (-2.25, 2.89) [0.00, 3.00]
GAMES  N: 42382 W: 10622 L: 10363 D: 21397
PENTA  [72, 4836, 11133, 5061, 89]
Caissa Witek
net_1536
Neural net: Increase accumulator size to 1536
diff10.0+0.10
ELO    1.58 +- 1.19 (95%)
LLR    3.01 (-2.94, 2.94) [0.00, 2.00]
GAMES  N: 91412 W: 22159 L: 21744 D: 47509
PENTA  [420, 10869, 22705, 11300, 412]
Integral Fury
failed_pc_lmr
diff10.0+0.10
ELO    -1.16 +- 2.54 (95%)
LLR    -1.71 (-2.25, 2.89) [0.00, 3.00]
GAMES  N: 17976 W: 4414 L: 4474 D: 9088
PENTA  [35, 2113, 4768, 2021, 51]
Caissa Witek
net_1536
Neural net: Increase accumulator size to 1536
diff60.0+0.60
ELO    7.32 +- 3.54 (95%)
LLR    2.98 (-2.94, 2.94) [0.00, 3.00]
GAMES  N: 9072 W: 2217 L: 2026 D: 4829
PENTA  [8, 990, 2348, 1183, 7]
Integral Fury
failed_pc_lmr
diff10.0+0.10
ELO    0.00 +- 0.00 (95%)
LLR    0.00 (-2.25, 2.89) [0.00, 3.00]
GAMES  N: 0 W: 0 L: 0 D: 0
PENTA  [0, 0, 0, 0, 0]
Quanticade DarkNeutrino
se-bs
search.c: capture for SE and pv for triple extensions margins
diff10.0+0.10
ELO    -24.49 +- 22.48 (95%)
LLR    -0.31 (-2.25, 2.89) [0.00, 3.00]
GAMES  N: 270 W: 55 L: 74 D: 141
PENTA  [3, 39, 68, 24, 1]
Integral Fury
inc_nmp_base_red
diff10.0+0.10
ELO    -1.72 +- 2.54 (95%)
LLR    -2.26 (-2.25, 2.89) [0.00, 3.00]
GAMES  N: 18570 W: 4632 L: 4724 D: 9214
PENTA  [49, 2235, 4815, 2131, 55]
Integral Fury
funky_nmp_tt_idea
diff10.0+0.10
ELO    -10.34 +- 5.37 (95%)
LLR    -2.27 (-2.25, 2.89) [0.00, 3.00]
GAMES  N: 4202 W: 974 L: 1099 D: 2129
PENTA  [12, 570, 1057, 455, 7]