# Zeta36/chess-alpha-zero

Chess reinforcement learning by AlphaGo Zero methods.

Repository: https://github.com/Zeta36/chess-alpha-zero
Canonical: https://ross.abutalabs.com/products/chess-alpha-zero
Language: Jupyter Notebook
License: MIT
License Family: permissive
Topics: reinforcement-learning, keras, alphago-zero, tensorflow, chess
Last push: 2023-03-24T22:22:17+00:00

## Health v2 (maintenance only)
Score: 32/100 (v2, computed 2026-09-03T02:20:16.233290+00:00)
- activity 0, release rhythm 35, longevity 100
- inputs: {"age_days": 3209, "days_push": 1258, "days_rel": null, "gap_med": null, "n_releases_24m": 0}
- flags: no_releases
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 2224, forks 472 (observed 2026-08-28T04:06:27.863536+00:00)

## What it is
A research/educational project implementing chess reinforcement learning using AlphaGo Zero methods (self-play, MCTS, residual neural networks) in Python with Keras and TensorFlow. It includes notebooks and a supervised-learning-trained model, though full self-play training is impractical on a single machine.

## Use cases
- learn how alphago zero works by studying code
- train a neural network to play chess
- implement mcts with a policy-value network
- run a chess ai demo in a jupyter notebook
- experiment with self-play reinforcement learning
- reproduce deepmind's alphazero approach on chess

## When to choose
- you want to study or learn AlphaZero-style RL algorithms hands-on
- you want a readable reference implementation in Keras/TensorFlow
- you want to experiment with MCTS plus neural networks on chess

## When to avoid
- you need a strong, production-ready chess engine (use Stockfish or Leela Chess Zero)
- you want to complete full self-play training on a single machine (too computationally expensive)
- you need actively maintained, up-to-date dependencies (the project is largely historical)

## Facets
- artifact type: learning-resource
- maturity: maintenance
- function: reinforcement-learning, machine-learning, deep-learning
- domain: reinforcement-learning, machine-learning
- platform: python, cross-platform
- tags: alphago-zero, chess, mcts, keras, tensorflow, self-play, jupyter-notebooks, game-development, gpu

## Member repositories
- Zeta36/chess-alpha-zero (main) score 32

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:06:27.863536+00:00.
- Health v2: computed from the inputs above; adoption is never an input.
- Inferred fields (summary, facets, guidance): AI-extracted, prompt v1, taxonomy v1, on 2026-08-30T02:44:57.242550+00:00, confidence not recorded.
  - readme: https://github.com/Zeta36/chess-alpha-zero (fetched 2026-08-28T04:06:27.863536+00:00, sha ae870079cfa6)
- Data as of 2026-08-30T08:39:29.467469+00:00.