SPRT tuning_count_pruning vs master diff

Diff (0 comments) View on GitHub

Details

id 51a6e2038f0c3e6ac564d845
new_tag tuning_count_pruning (more aggresive pruning by glinscott that may work )
new_signature 4144462
new_options Hash=128 OwnBook=false
resolved_new f2e270438d
new_net
base_tag master (Shrink engine UCI name)
base_signature 4821467
base_options Hash=128 OwnBook=false
resolved_base abb40777bf
base_net
sprt elo0: -1.50 alpha: 0.05 elo1: 4.50 beta: 0.05 state: rejected (BayesElo)
num_games 128000
tc 15+0.05
new_tc 15+0.05
threads 1
book 8moves_GM.pgn
book_depth 8
auto_purge False
priority -1
itp 100.00
throughput 100
username uriblass
tests_repo https://github.com/uriblass/Stockfish
adjudication True
info old idea about tuning that now may work because of better move order.
events /actions?run_id=51a6e2038f0c3e6ac564d845
raw statistics /tests/stats/51a6e2038f0c3e6ac564d845

Actions

Base branch same as Stockfish master


Stats

chi^234.73
dof40
p-value70.61%

Time

start time2013-05-30 05:22:11
last updated2013-05-30 20:47:59

Tasks (0 active workers with 0 cores)

Idx Worker Info Last Updated Played Wins Losses Draws Crashes Time Residual