Automate Chess with WinBoard/UCI

Project: automate chess processes using WinBoard/UCI and orchestrate local AI agents for massive custom variant tournaments. Main skill: Master Windows PowerShell, LM Studio, use of Winboard and Fairy Stockfish. The project involves creating a system that can efficiently manage and execute large-scale chess tournaments with custom rules: Core Objective Build a local orchestration system that can: Generate and run massive batches of custom tournaments. Use custom armies (often asymmetrical) with variable piece sets. Start games from a FEN database generated from all defined piece combinations. Export all games to PGN files, with filenames linked to the corresponding FEN database. Evaluate and validate a handicap formula based on piece-count asymmetry and relative piece values (in centipawns). Technical Scope The system should automate: Tournament creation and launch (WinBoard/UCI workflow). Generate Fen data for each armies combinations. Save Fen Data using Letters and their corresponding Betza Notation: For example White plays with extra Q and Black plays with Hawk and Unicorn. The generated Data will be saved on file named Q vs HH then add Betza Notation for each extra piece, Q = Q etc Engine setup and execution (Fairy-Stockfish compatible). FEN database ingestion as tournament starting positions. Match scheduling across many custom army configurations. PGN output management with consistent naming conventions. PGN should have the same name than the FEN Data. If FEN Data is big (for example superior than 500 positions, the data will be split to multiples of 500, then the data will be remaned incrementally 1 etc. Local execution on Windows (PowerShell-friendly automation). Tournament Design Logic Each side can be assigned custom piece pools (example ranges: 1 to 20 custom pieces available to chose from per side). The armies can be 1 to 4 pieces, For each generated army: Every custom piece has a predefined relative value in centipawns (cp). Armies are frequently asymmetrical (example: White has 1 very strong piece vs Black has 3 weaker pieces). In many scenarios, White’s single piece may have a higher raw value than Black’s total piece value. Black may receive a handicap adjustment because it plays with more pieces. Handicap Model to Validate (Main Experimental Goal) The primary experiment is to validate whether the handicap correction below is appropriate: 1 extra piece: +100 cp handicap 2 extra pieces: +250 cp handicap (100 + 150) 3 extra pieces: +350 cp handicap (250 + 100) 4 or more extra pieces: capped at +350 cp This handicap is applied to the side with more pieces (typically Black in asymmetrical setups), and the project’s main scientific objective is to test if this model produces balanced and meaningful outcomes across large tournament samples. Expected Outcome A robust local pipeline that can repeatedly: design custom tournaments, run them at scale, store clean PGN datasets, and generate evidence to confirm or reject the handicap formula.