Skip to main content
Formal Specification

Fox and Geese (Asymmetric Tafl Hunt Game) — Formal Specification

United Kingdom · board-abstract · Confidence: medium

← Back to game page

Implementation Suitability

Suitability: engine-readyFidelity: directGap-fills: 2
# Fox and Geese — AI-optimized specification ## General idea A deterministic asymmetric hunt game on a 33-point cross graph where 1 Fox duels 13 Geese. The Geese attempt to coordinate an encircling wall to trap and immobilize the Fox, while the Fox maneuvers to execute leaping checker-like captures to thin the flock below the critical entrapment mass. ## Formal model - **Abstraction taken:** None (`fidelity_class: direct`). Fully discrete mathematical state space. - **Board Graph:** 33 nodes `V = {0..32}` arranged on a 7x7 grid minus four 2x2 corner cutouts. Edges connect orthogonal neighbors, and diagonal lines connect points across the central 3x3 square. - **Player/agent count:** 2 agents: `Agent_Fox` (controls 1 piece) and `Agent_Geese` (controls 13 pieces). - **Turn model:** Strictly alternating sequential turns; `Agent_Geese` plays first. ## Formal rules R1. **Initial Setup (13 Geese Variant):** - `Geese` occupy nodes `0..12` (the entire top arm and horizontal midline). - `Fox` is placed at the center node `16`. R2. **Geese Turn:** - On its turn, `Agent_Geese` selects one goose at node `u` and moves it to an adjacent vacant node `v` where `(u, v) in E` and `state[v] == 0`. - Geese cannot jump or capture. R3. **Fox Turn:** - `Agent_Fox` selects either a simple step or a jumping capture: - Simple Step: Move Fox from `u` to adjacent vacant node `v` (`(u, v) in E`, `state[v] == 0`). - Jump Capture: If node `w` contains a goose, `(u, w) in E`, and collinear point `v` directly beyond `w` is vacant (`state[v] == 0`), Fox leaps to `v`. The goose at `w` is permanently removed (`geese_remaining -= 1`). - Multi-Jump: If landing on `v` immediately presents further valid jump captures, Fox may execute additional consecutive leaps in the same turn. R4. **Victory Conditions:** - **Geese Victory:** Occurs if on Fox's turn, `Agent_Fox` has zero legal moves (no legal steps and no legal captures). - **Fox Victory:** Occurs immediately if `geese_remaining <= 5` (mathematically insufficient to enclose the Fox). R5. [gap-fill] Draw Conditions:** - Threefold Repetition: The match is a draw if the identical state (Fox position, Geese bitmask, active player) occurs 3 times. - Move Limit: If 60 consecutive turns elapse without a goose capture, the game is declared a draw. ## Edge cases - **Tie at the end condition:** Handled by R5 (threefold repetition or 60-move non-capture rule). - **A player with no legal move:** Handled by R4: if Fox has no legal move, Geese win immediately. If a individual goose is blocked, Geese player must move any other unblocked goose; if all geese are immobilized, [gap-fill] Fox wins. - **Repetition / cycles:** Bounded by R5 (threefold repetition). - **Simultaneous claims on the same resource:** Not applicable; strictly alternating turns. - **Player elimination mid-round:** Geese pieces are removed on capture; piece reduction triggers Fox victory when reaching `<= 5`. - **Resource exhaustion:** Non-renewable geese pieces; board is finite. - **Termination guarantee:** Guaranteed by monotone decrease of geese count towards `<= 5` combined with R5 repetition limits. ## State representation - **Persistent state:** `fox_pos: int in [0..32]`, `geese_bitmask: int (33-bit integer)`, `geese_remaining: int (0..13)`, `active_player: enum [FOX, GEESE]`, `halfmove_clock: int`, `position_history: set of hash`. - **Per-turn ephemeral state:** `multi_jump_path: list of int`. - **Observable information:** Perfect complete information game (zero hidden information). ## Action space - `move_goose(src: int, dst: int)`: Action for Geese. - `step_fox(dst: int)`: Fox single-step movement. - `jump_fox(landing_path: list of int)`: Fox capture leap(s). ## Playability notes - **Single-player-vs-AI:** Highly suitable for Minimax with Alpha-Beta pruning, MCTS, or Deep RL. Asymmetric dynamics provide distinct tactical play styles. - **Group games reduction:** Inherently a 1v1 duel. - **Physical-skill games:** Not applicable (pure deterministic abstract board game). - **What is lost in the adaptation:** Zero fidelity loss; graph state maps 1:1 with the historical board. ## Implementation hints - Represent the entire board state in a single 64-bit integer (`fox_pos: 6 bits`, `geese_bitmask: 33 bits`). - Fast bitwise operations for legal jump detection and adjacent neighbors. - First milestone: Write state machine and legal move generator, simulating 10,000 random-action games to verify termination and win conditions.

Download as JSON

Download the full game corpus entry for this game as structured JSON.

Download JSON
Cite this page

Fox and Geese (Asymmetric Tafl Hunt Game) (United Kingdom). (n.d.). Games of Countries. https://gamesofcountries.com/games/gb-united-kingdom/fox-and-geese

@misc{gamesofcountries-gb-united-kingdom-fox-and-geese,
  title = {Fox and Geese (Asymmetric Tafl Hunt Game) (United Kingdom)},
  year = {n.d.},
  publisher = {Games of Countries},
  url = {https://gamesofcountries.com/games/gb-united-kingdom/fox-and-geese}
}