10000 games - net_only_quiet_no_prune_no_tt_40m_d3_2 vs net_master_no_prune_no_tt_40m_d3_2 diff

  
  
    Elo: 30.80 ± 5.0 (95%) LOS: 100.0%
      
Total: 9400 W: 2768 L: 1937 D: 4695
Ptnml(0-2): 220, 946, 1724, 1403, 407
nElo: 43.39 ± 7.1 (95%) PairsRatio: 1.55

Diff (0 comments) View on GitHub

Details

id 5f844da38ea73fb8ddf83ea9
rescheduled_from 5f844d138ea73fb8ddf83ea6
new_tag net_only_quiet_no_prune_no_tt_40m_d3_2 (net trained with constant lr)
new_signature 4092574
new_options Hash=16 Use NNUE=true
resolved_new fed186d037
new_net nn-8f7566bef177.nnue
base_tag net_master_no_prune_no_tt_40m_d3_2 (net trained with constant lr)
base_signature 3914428
base_options Hash=16 Use NNUE=true
resolved_base ff6db03f59
base_net nn-8f743bcd055d.nnue
num_games 10000
tc 10+0.1
new_tc 10+0.1
threads 1
book noob_3moves.epd
book_depth 8
auto_purge False
priority 0
itp 50.00
throughput 100
username Sopel
tests_repo https://github.com/Sopel97/Stockfish
adjudication True
info same setting as https://tests.stockfishchess.org/tests/view/5f8431618ea73fb8ddf83e96 but trained with constant lr instead of decresing lr to see if it had an impact.
events /actions?run_id=5f844da38ea73fb8ddf83ea9
raw statistics /tests/stats/5f844da38ea73fb8ddf83ea9

Actions

Base branch not same as Stockfish master


Stats

chi^266.36
dof60
p-value26.71%

Time

start time2020-10-12 12:35:47
last updated2020-10-12 15:35:44

Tasks (0 active workers with 0 cores)

Idx Worker Info Last Updated Played Pentanomial [0‑2] Crashes Time Residual