Skip to main content
Formal Specification

Hungarian Tag-and-Rescue Hide-and-Seek (Ipiapacs) — Formal Specification

Hungary · hide-seek · Confidence: low

← Back to game page

Implementation Suitability

Suitability: adaptation-requiredFidelity: discretizedGap-fills: 2
# Ipiapacs — AI-optimized specification ## General idea A multi-agent hide-and-seek simulation on an obstacle grid combining partial-observability search with a deterministic sprint race to a central sanctuary cell. The seeker balances exploration distance against the vulnerability of leaving the home base unguarded against stealth infiltrators. ## Formal model - **Abstraction taken:** The environment is discretized into a 20×20 2D grid containing obstacles (vision and movement blocking). The Home Base is located at `(10, 10)`. Agents possess 2D coordinates, movement speeds (cells per tick), and directional field-of-view (FOV) cones. - **Player/agent count:** 1 Seeker `S` and `N` Hiders `H_1..H_N` (standard `N = 4`). - **Turn model:** Tick-based simultaneous movement (tick = 0.5s equivalent). ## Formal rules R1. Setup Phase: Seeker `S` is locked at Base `(10, 10)` for `T_count = 20 ticks` with vision disabled (`FOV = None`). Hiders `H_i` move freely to choose initial hiding coordinates `(x, y)`. R2. Search Phase: Seeker `S` vision is enabled (360-degree FOV, range = 8 cells, raycasting blocked by obstacle cells). R3. Discovery Check: If a straight line-of-sight exists between `S` and `H_i` with `dist(S, H_i) <= 8`, Hider `H_i` is marked `Spotted`. R4. Base Race Resolution: - Once `H_i` is spotted, or chooses to initiate a self-rescue run: - Both `S` and `H_i` compute paths to Base `(10, 10)`. - At each movement tick, agents move up to their speed limit toward Base. - If `S` reaches Base cell first and executes `tag_out(H_i)`: `H_i` status becomes `Captured`. - If `H_i` reaches Base cell first and executes `rescue_self()`: `H_i` status becomes `Safe`. R5. Stealth Self-Rescue: Any unspotted hider `H_j` may move toward Base at any time. If `H_j` enters Base cell without `S` in the same cell, `H_j` executes `rescue_self()` and becomes `Safe`. R6. Collective Liberation: If the last remaining non-captured/non-safe hider `H_last` reaches Base cell first and executes `liberate_all()`, all `Captured` hiders transition to `Safe`, and Seeker `S` fails the round. R7. Round End & Seeker Assignment: - Round terminates when all hiders are either `Captured` or `Safe`. - If at least 1 hider is `Captured` and no liberation occurred, the first captured hider becomes Seeker for the next round. - If 0 hiders are `Captured` (or collective liberation succeeded), the current Seeker remains Seeker for the next round. R8. [gap-fill] Search Time Limit: Round automatically terminates after `T_max = 300 ticks`. Any hiders still hidden and uncaptured at `T_max` are automatically deemed `Safe`. ## Edge cases - Tie at the end condition: If Seeker and Hider step onto the Base cell simultaneously on the exact same tick, the Seeker takes precedence (R4 tag succeeds) [gap-fill]. - A player with no legal move: In grid navigation, if an agent is cornered, the agent can always `hold_position()`. - Repetition / cycles: Bounded by R8 round time limit (`T_max = 300 ticks`). - Simultaneous claims on the same resource or position: Base cell can hold multiple agents; claims are resolved instantaneously by arrival tick order. - Player elimination mid-round: Captured hiders are moved to a holding queue and cease taking movement actions until next round. - Resource exhaustion (pieces, tokens, space, time): Static grid; finite agent counts; clock strictly decreases. - Termination guarantee: Guaranteed by the hard tick bound `T_max`. ## State representation - Persistent state: `current_seeker_id`, `seeker_penalties[agent_id]`, `round_number`. - Per-player state: `pos (x, y)`, `speed`, `status (Hiding, Spotted, Racing, Captured, Safe)`. - Ephemeral state: `tick_timer`, `captured_queue (ordered list)`, `grid_map (20x20 boolean obstacles)`. - Observable information: - Seeker observes: Grid map, visible hiders in FOV, Base status, captured queue. - Hider observes: Grid map, Seeker position if in FOV, Base distance, own status. ## Action space - Seeker: `move(dx, dy)`, `tag_out(target_id)`. - Hider: `move(dx, dy)`, `sprint_to_base()`, `rescue_self()`, `liberate_all()`, `hold_position()`. ## Playability notes - Single-player-vs-AI: Highly compelling stealth gameplay as human hider or pursuit-evasion AI benchmark as seeker. - Group games (n>4): Perfectly matches multi-agent predator-prey scenarios with asymmetric incentives. - Physical-skill games: Pure spatial navigation and pathfinding strategy. - What is lost: The physical breathless crouching in bushes and genuine acoustic listening cues. ## Implementation hints Medium complexity. Use standard A* pathfinding over a 20×20 grid. The Seeker's decision policy requires balancing two sub-goals: exploring frontier cells vs. patrolling within sprint distance of the Base. Suggested first milestone: implement grid raycasting FOV and 1v1 sprint races before adding multi-hider stealth rescue logic.

Download as JSON

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

Download JSON
Cite this page

Hungarian Tag-and-Rescue Hide-and-Seek (Ipiapacs) (Hungary). (n.d.). Games of Countries. https://gamesofcountries.com/games/hu-hungary/ipiapacs

@misc{gamesofcountries-hu-hungary-ipiapacs,
  title = {Hungarian Tag-and-Rescue Hide-and-Seek (Ipiapacs) (Hungary)},
  year = {n.d.},
  publisher = {Games of Countries},
  url = {https://gamesofcountries.com/games/hu-hungary/ipiapacs}
}