OpenBench Testing Framework

AuthorClaude
Creation1781635093
Dev Sourcehttps://github.com/adamtwiss/coda
Dev Sha5fbad2fc3708fd431d749623cdb9500896fde081
Dev Branchexperiment/psqt-s800-tuned
Dev Bench2790916
Dev Networkft1024-l1-32-psqt-s800
Dev OptionsThreads=1 Hash=64
Dev Time10.0+0.10
Base Sourcehttps://github.com/adamtwiss/coda
Base Sha1e553abf46e89166fc75d8012c437d33de956bdd
Base Branchexperiment/psqt-inference-on-main
Base Bench2466418
Base Networkft1024-l1-32-baseline-s800
Base OptionsThreads=1 Hash=64
Base Time10.0+0.10
Opening BookUHO_Lichess_4852_v1.epd
Upload PGNsFALSE
Syzygy WDLOPTIONAL
Syzygy ADJ.OPTIONAL
Win ADJ.movecount=3 score=500
Draw ADJ.movenumber=20 movecount=10 score=10
ScalingBASE 250000 NPS

experiment/psqt-s800-tun... vs experiment/psqt-inferenc...

ApproveRestartDelete
Copy Stat Block
Elo   | -10.91 +- 4.85 (95%)
SPRT  | 10.0+0.10s Threads=1 Hash=64MB
LLR   | -2.95 (-2.94, 2.94) [-1.50, 1.50]
Games | N: 6022 W: 1404 L: 1593 D: 3025
Penta | [52, 826, 1428, 669, 36]
#UserUpdatedGamesWonLostDrawnTimeCrash