Skip to content

Lozza 2.1

Choose a tag to compare

View all tags
@op12no2 op12no2 released this 18 Oct 18:37
· 28 commits to main since this release
2.1
dbcb390

Evaluation retuned with a new gradient descent (GD) Texel tuner, non-linear mobility split into mobility/tightness/tension, and Lichess support (lozzaBot).

Changes

  • Non-linear mobility.
  • Split up mobility into mobility, tightness and tension.
  • Add Lichess support.
  • Fixate on one square in Q search from depth -12.
  • Retune using gd tuner.
  • Add eval feature extraction code (for gd tuner) which is removed on release.
  • Extract imbalance as a separate eval term.
  • Handle old node versions WRT stdin.resume(). I think.
  • Optimise pruning to pre makeMove().
  • (2.0a) Fix timeouts.
  • (2.0a) Add USEPAWNHASH - useful when testing.
  • (2.0a) Set mob offsets to 0 while buggy.

Rating

CCRL 40/40: 2621 · CCRL Blitz: 2622

Usage

Use lozza.js with Node.js in any UCI chess GUI (Arena, Cute Chess, etc.), or in a web page/web worker.

Binaries

While Lozza is intended for web use, I can supply binaries on request. Please email me op12no2@gmail.com noting the target platform; thanks.

Assets 3
Loading