Approved Milestones

Assign tasks to these milestones logically. Each milestone must be fully tested (including performance benchmarks) before moving to the next.

View current milestones on GitHub

---

✅ v0.1 - Foundation & Core Types

• Status: Completed 🏆
• Scope: Project setup (SBT 2.x / Scala 3), configuration, mise setup.
• Key Deliverables:
  • Implementation of basic Opaque Types (Bitboard, Square, Piece, Color).
  • Basic FEN parsing and serialization.
  • DFEN extension (7th field for dice pool, multiple en-passant targets).

✅ v0.2 - Move Generation (Classic)

• Status: Completed 🏆
• Scope: Fast legal and pseudo-legal move computation.
• Key Deliverables:
  • Bitwise operations and precomputed attack tables (Magic Bitboards).
  • Pawn, knight, king, and sliding piece move generation.
  • Perft (Performance Test) framework integration to verify move correctness.

✅ v0.3 - Dice Chess Mechanics

• Status: Completed 🏆
• Scope: Integrating physical probability layers of Dice Chess.
• Key Deliverables:
  • Dice roll representations.
  • Filtering pseudo-legal moves based on dice outcomes.
  • Game state management with random events.
  • Maximum Micro-moves Rule enforcement.
  • Turn lifecycle: roll → generate moves → apply micro-moves → endTurn.

✅ v0.4 - Basic Bot & Gameplay

• Status: Completed 🏆
• Scope: Validating game mechanics and state transitions with primitive bots.
• Key Deliverables:
  • Implementation of 7 primitive bots: Random, Checkmate-Aware, Greedy, Cautious Greedy, Aggressive, Monte-Carlo, Expectimax.
  • Integration with Scala.js for browser-based execution.
  • Dedicated Svelte/Vite PWA test harness for human vs. engine testing.
  • JVM Battle Arena for bot-vs-bot win-rate validation.

✅ v0.5 - Evaluation & Heuristics

• Status: Completed 🏆
• Scope: Understanding board values statically and efficiently caching calculations.
• Key Deliverables:
  • Static evaluation function (Material balance, Piece-Square Tables).
  • Zobrist Hashing and Transposition Tables (TT) for caching board states.
  • King Capture Probability analysis (exact enumeration of 216 dice outcomes).

✅ v0.6 - Expectimax Search Engine

• Status: Completed 🏆
• Scope: Deep probabilistic search with chance nodes.
• Key Deliverables:
  • Deep Expectimax search implementation with chance nodes (216 ordered rolls / 56 unique multisets).
  • Star1/Star2 pruning for chance nodes to reduce search tree.
  • Time management subsystem (TimeManager policy + TimeBudgetedSearch honouring).
  • Rao-Blackwellized Monte-Carlo pre-roll equity estimator.
  • Structured concurrency with Virtual Threads (Ox) for parallel chance-node evaluation.

✅ v0.7 - Advanced Features

• Status: Completed 🏆
• Scope: Production-ready features and integrations.
• Key Deliverables:
  • ONNX model integration for learned evaluation (OnnxEvalSearch, OnnxExpectimaxSearch).
  • Opening book support for common positions (OpeningBook, OpeningBookBot, OpeningBookParser).
  • Doubling cube logic and draw offer handling (DrawOfferLogic).
  • JMH benchmarks for performance validation.

🚀 v1.0 - Production & Optimization

• Scope: Deployment optimization and infrastructure operations.
• Key Deliverables:
  • GraalVM Native Image compilation for fast startup.
  • Dockerfile optimization for containerized deployment.
  • CI/CD pipeline improvements (release automation, publishing).
  • Deployment configurations for Oracle Cloud Free Tier (Ampere ARM64).