[NeurIPS 2023] An online fine-tuning game solver for 7x7 Killall-Go
-
Updated
Oct 30, 2023 - C++
[NeurIPS 2023] An online fine-tuning game solver for 7x7 Killall-Go
A* search algorithm implementation for solving 8-puzzle problem. Includes code modularization, performance optimization and solvability checking features.
Standalone C++23 solver and interactive CLI for undirected vertex geography using exact matching-based evaluation.
Game solving algorithms (MCTS, Negamax...)
Solved games with their result, method and primary source, and the open problems one step past them. CC0 data.
Offline-first Royal Game of Ur with search, self-play ML, and solved-game-distilled Oracle AI in React, Rust/WASM, and Cloudflare Workers.
Crazy Stackers IA is an automatic solver for the Crazy Stackers game, created using the libGDX framework and leveraging the DLV2 logic programming system. This project aims to provide an efficient solution to the game using advanced AI techniques.
Antichess on the xiangqi board, measured before it was designed: the rules, every engine game, every proof certificate, and a checker that replays them. It is a draw.
Minimal reproduction for the 4x4 / 6-piece Chinese Checkers Table 1 row
Strongly solving Gobblet Gobblers: a from-scratch Rust retrograde tablebase, a probe API, and a move-coloring explorer.
Tic Tac Chec, Don Green's 4x4 chess-movement alignment game, strongly solved: a draw under perfect play, with the audited tablebase and a live explorer.
Three-player Chinese Checkers, strongly solved on the 37-cell board: the packed tablebase, the kingmaker analysis, and a live explorer.
Retrograde-analysis engine for small combinatorial games, plus a browser explorer. Six men's morris strongly solved: a draw over 42,372,745 positions.
Gekitai solver with independently verified strategies and an animated browser player
Strongly solving Dōbutsu Shōgi: a from-scratch Rust tablebase, a probe API, and a lichess-style explorer.
Strongly solve Gobblet Gobblers by retrograde analysis — 0.37 s on a B200, verified against an independent solver
Shogi4 (4x4 shogi): validated strong-solution engine for the public-domain 4×4 drop-shogi. Rules recovered from the official Android package; 2.1B-position reduced solve; full distributed solve designed and calibrated, not run.
Rules engine, run records, and tablebase experiments for Micro Shogi (4x5 drop-shogi).
To associate your repository with the game-solving topic, visit your repo's landing page and select "manage topics."