Formal Specification
Javanese Tiger Game / Macanan — Formal Specification
Indonesia · board-abstract · Confidence: high
← Back to game pageImplementation Suitability
Suitability: engine-readyFidelity: directGap-fills: 4
# Macanan — AI-optimized specification
## General idea
A two-player asymmetric abstract strategy game on an undirected planar graph where one Tiger agent attempts to capture prey via jump-captures while a Herd agent drops and maneuvers 16 to 24 pieces to trap and immobilize the tiger. The mechanics are discrete and deterministic, with zero hidden information.
## Formal model
- **Abstraction taken:** None required (`fidelity_class: direct`). The traditional graph topology (23 nodes) and discrete piece movement are modeled directly.
- **Player/agent count:** 2 agents: (Tiger) and (Herd).
- **Turn model:** Strictly sequential alternating turns. moves first.
## Formal rules
R1. Board graph: with nodes comprising a grid with diagonals (9 nodes) joined to a 4-node apex triangle and extended perimeter transversals.
R2. Piece initialization:
- Tiger piece starts at apex node .
- Herd player has pieces in hand; board has 0 herd pieces.
R3. Turn sequence — Phase 1 (Drop & Hunt):
- (Herd) action: Place 1 piece from reserve onto any vacant node . .
- (Tiger) action:
- Regular move: Shift from node to adjacent vacant node where .
- Jump capture: If and along a collinear straight line segment, contains a herd piece, and is vacant: Tiger moves , and the piece at is removed from play.
R4. Turn sequence — Phase 2 (Maneuver & Hunt):
- When , moves any on-board herd piece to an adjacent vacant node along an incident edge. Herd pieces cannot jump or capture.
- continues regular moves and jump captures.
R5. Multiple captures [gap-fill]: Tiger is restricted to a single capture per turn to preserve game balance.
R6. Win/Loss conditions:
- Herd Victory: wins if (Tiger) has zero legal moves and zero legal captures on its turn.
- Tiger Victory: wins if active on-board herd pieces [gap-fill].
- Draw: [gap-fill] If the same board state repeats 3 times (threefold repetition) or 100 turns elapse without a capture, the game is declared a draw.
## Edge cases
- Tie at the end condition: Resolved via threefold repetition or 100-turn non-capture limit yielding a draw.
- A player with no legal move:
- If Tiger has no legal move Herd wins immediately (R6).
- If Herd has no legal move on Phase 2 Herd forfeits, Tiger wins [gap-fill].
- Repetition / cycles: Bounded by threefold repetition rule and maximum non-capture turn counter.
- Simultaneous claims on the same resource or position: Not applicable; nodes have single-occupancy and turns are strictly sequential.
- Player elimination mid-round: Captured herd pieces are permanently removed from state; turn alternating structure is invariant.
- Resource exhaustion: Reserve pieces deplete monotonically in Phase 1 to zero, triggering Phase 2.
- Termination guarantee: Enforced by the non-capture move counter limit and discrete finite state space.
## State representation
- Persistent state:
- `board`: array of length 23, values .
- `reserve_herd`: integer .
- `captured_herd`: integer .
- `active_player`: `0` (Tiger) or `1` (Herd).
- `halfmove_clock`: integer (tracks turns since last capture).
- `history_hashes`: set of previous board state hashes.
- Per-turn ephemeral state:
- `selected_move`: `(from_node, to_node, captured_node_opt)`.
- Observable information: Fully observable (no hidden information or randomness).
## Action space
- `drop_piece(target_node)`: Precondition: Phase 1, target node is EMPTY.
- `move_piece(from_node, to_node)`: Precondition: `(from_node, to_node)` , `to_node` is EMPTY.
- `jump_capture(from_node, jumped_node, land_node)`: Precondition: collinear edge alignment, `jumped_node` has HERD, `land_node` is EMPTY.
## Playability notes
- Single-player-vs-AI: Engine-ready for Minimax/Alpha-Beta search and MCTS without modifications.
- Group games (n>4): Not applicable; inherently a 2-agent asymmetric zero-sum game.
- Physical-skill games: Purely combinatorial abstract strategy; no physical skill abstraction needed.
- What is lost in the adaptation: None; the mathematical structure of the traditional game is 100% preserved.
## Implementation hints
Ideal for zero-sum tree search benchmark. Board representation can be bitboards or an adjacency matrix over 23 vertices. First milestone: Implement graph adjacency matrix and legal move generator for both Tiger and Herd.
Download as JSON
Download the full game corpus entry for this game as structured JSON.
Download JSONCite this page
Javanese Tiger Game / Macanan (Indonesia). (n.d.). Games of Countries. https://gamesofcountries.com/games/id-indonesia/macanan
@misc{gamesofcountries-id-indonesia-macanan,
title = {Javanese Tiger Game / Macanan (Indonesia)},
year = {n.d.},
publisher = {Games of Countries},
url = {https://gamesofcountries.com/games/id-indonesia/macanan}
}