Wednesday, 29 July 2026 @ 16:01 UTC~1013,752154,727

Palamedes

Nikolaos Papahristou·nikpapa.palamedes

Play backgammon and other variants against award winning AI

Palamedes is a project dedicated to offering expert playing computer opponents at backgammon and several other variants. Currently users of Palamedes can play the variants Portes, Plakoto, Fevga collectively called Tavli (Greek backgammon) and Narde. For position evaluation, Palamedes uses neural networks trained by a form of temporal difference learning and millions of self-play games. Palamedes also offers simple ways of storing and loading games, as well as simple analyzing tools.

winget install --id nikpapa.palamedes --exact --source winget

Latest 0.55.0·September 1, 2025

Release Notes

Development:

  • Updated Qt to 6.9.2 Core:
  • Algorithms: added multithreading to NegaBatch and IDDFS search algorithms
  • Algorithms: removed stats.currentMove and added movesCompleted. This statistic is better suited for multi-threaded search GUI:
  • Start position can now be saved/loaded to/from .tavli files
  • Fixed visual glitches when AI has finished thinking before computer dice has finished animated
  • Added pips inside board
  • NeoAnalysis: user can expand multiple moves at once (with right click)
  • NeoAnalysis: fix boldness of Diff column
  • NeoAnalysis: board move is now added inside search and not expanded after analysis has finished
  • Advanced: removed old analysis window

Installer type: nullsoft

x86—BEEC84B24455988ADEC6093169FCF5036E75A079E3C03AC09215D1F12D464F80

Details

Homepage
https://nikpapa.com/Palamedes
License
Freeware
Publisher
Nikolaos Papahristou
Privacy Policy
https://nikpapa.com/Palamedes/Palamedes_Privacy_Policy.html
Copyright
Copyright (c) Nikolaos Papahristou. All rights reserved.
Moniker
palamedes

Tags

backgammonboardgamegamesvideogame

Older versions (4)

0.53.3
x64—5AC090325D2B95E39BFE40C665228496047841ADE956170D2DF438DEB685949A
0.53.2
x64—51AEA8767F4F54F7E0DC01853241EFD38C851F307433E20F860C6D02103EEBF3
0.53.1
x64—AFD9E254C6772BAF4C7DB29A476893269E6E3D7AA097E811CB1D60CA1C21D045
0.47
x86—814E9B7D174ED72CECC92C5471038D73160F70B398D39A684DC7389CE4624EC4