MattBench MattBench Workload n9x.co

Configuration

AuthorNocturn9x
Variantstandard
Creation1774916517
Dev Sourcehttps://github.com/nocturn9x/heimdall
Dev SHA1df950bb9247661343c61d9119e3b70e2f662317
Dev Branchclaude-fixes
Dev Bench5370815
Dev Avg NPS—
Dev Networklaevateinn-v2-verbatim
Dev OptionsThreads=1 Hash=16
Dev Time8.0+0.08
Base Sourcehttps://github.com/nocturn9x/heimdall
Base SHAd8ccfa570cd4ae22ab75854e549d47d6a182eb81
Base Branchmaster
Base Bench5370815
Base Avg NPS—
Base Networklaevateinn-v2-verbatim
Base OptionsThreads=1 Hash=16
Base Time8.0+0.08
Opening BookUHO_Lichess_4852_v1.epd
Upload PGNsFALSE
Syzygy WDLDISABLED
Syzygy ADJ.OPTIONAL
Win ADJ.movecount=8 score=800
Draw ADJ.movenumber=32 movecount=10 score=0
ScalingBOTH 830000 NPS

claude-fixes vs master

Name  | claude-fixes
Elo | -4.53 +- 3.72 (95%)
SPRT | 8.0+0.08s Threads=1 Hash=16MB
LLR | -2.31 (-2.25, 2.89) [-5.00, 0.00]
Games | N: 8738 W: 2023 L: 2137 D: 4578
Penta | [40, 1036, 2330, 924, 39]

LLR history

Hover or touch to inspect samples. Use the left and right arrow keys to move between samples, Home and End to jump, and Escape to dismiss.

Favour promising tests within their priority, using between half and twice the configured throughput.

# User Games Wins Draws Losses LL LD DD / WL DW WW Pairs Time Crash

Errors

TimeWorkerUserErrorLogs
1774928334 #2295 Tecci [Heimdall] claude-fixes build failed Full log