Skip to main content
Formal Specification

Syrian Hide and Seek — Formal Specification

Syria · hide-seek · Confidence: medium

← Back to game page

Implementation Suitability

Suitability: adaptation-requiredFidelity: discretizedGap-fills: 2
# Al-Tamimeh — AI-optimized specification ## General idea A chase-and-seek game discretized to a grid or 2D node map where one "seeker" agent tries to spot and tag other "hider" agents. Strategic depth is in pathfinding, line-of-sight visual angles, and the timing of when to sprint out of cover to reach the home base. ## Formal model - **Abstraction taken:** The continuous physical environment (courtyard/streets) is discretized into a 2D grid graph. Movement is step-based per turn. Vision is calculated by raycasting line-of-sight from the seeker's position. - **Player/agent count:** 1 seeker agent, 2 to 11 hider agents. - **Turn model:** Real-time-ticked (simultaneous turn execution per state tick, with speed differentials represented by movement point allocations per tick). ## Formal rules R1. Setup: A 2D grid of size XxY with impassable walls (cover). The home base node *mirsa* is at a fixed coordinate. All hiders start at the *mirsa*. R2. [gap-fill] Blinding phase: The seeker is locked at the *mirsa* node for T ticks (counting phase) and has a vision radius of 0. During these T ticks, hiders can move up to M nodes per tick to hide behind walls. R3. Search phase: After T ticks, the seeker's vision radius is set to normal (e.g., 10 cells) and 120-degree field of view based on their facing direction. R4. Tick loop: Every tick, each non-captured hider and the seeker simultaneously declare their next move action (one of: up, down, left, right, stay, or change facing direction). R5. Speed: Seeker moves at 2 nodes per tick. Hiders move at 1 node per tick. [gap-fill: speed differential to model physical chase dynamics]. R6. Raycasting: At the end of each tick, if a hider is within the seeker's field of view and has no wall blocking the straight line-of-sight (LOS) ray, the hider is "spotted". R7. Tagging/Tamm: 7.1. Once spotted, a race begins. The seeker must reach the *mirsa* node and call `tammam(hider_id)`. 7.2. Alternatively, the spotted hider can reach the *mirsa* node and call `halal()`. 7.3. Whichever agent reaches and interacts with the *mirsa* node first resolves that hider's state. If the seeker wins, the hider is marked "captured". If the hider wins, they are marked "saved". R8. Win / Loss / Draw: 8.1. If all hiders are captured, the seeker wins. 8.2. If any hider is saved, they win individually. 8.3. The game ends after a maximum tick limit (e.g. 500 ticks), at which point any non-captured active hider is declared "saved". ## Edge cases - Tie at the end condition: If a hider and the seeker reach the *mirsa* at the exact same tick, the hider's action is resolved first, rendering them "saved" (modeled as runner's advantage). - A player with no legal move: Not applicable; an agent can always execute a `stay` or change facing direction action. - Repetition / cycles: Bounded by the absolute game tick limit (R8.3), preventing infinite cycling. - Simultaneous claims on the same resource: If multiple hiders attempt to move into the same node, a random hider succeeds, and the other is deflected to their previous node. - Player elimination mid-round: When a hider is captured, they are removed from the active grid, and turn order remains unaffected. - Resource exhaustion: None. - Termination guarantee: Guaranteed by the maximum tick limit (R8.3). ## State representation - Persistent state: Grid layout (dimensions, wall locations, *mirsa* coordinates), seeker position and facing angle, dictionary of hider positions, hider statuses (hiding, spotted, captured, saved), current game tick, tick limit. - Per-turn ephemeral state: Simultaneous movement intentions of all active agents. - Observable information: Seeker sees the entire grid within their line-of-sight cone; hiders see the positions of the seeker and other hiders at all times (representing spatial awareness and hearing), but can only see grid layout within their own vision range. ## Action space - `move(direction)`: Moves 1 cell in a cardinal direction. - `turn(angle)`: Changes facing direction (seeker only). - `interact_mirsa()`: Triggers the *mirsa* action when standing on the base node. ## Playability notes - Single-player-vs-AI: Highly suitable. The human plays as a hider or seeker, with simple heuristic-based pathfinding (A* search) and raycasting for AI opponents. - Group games (n>4): Generalizes cleanly. Since all hiders act simultaneously, adding more agents only increases simulation cost, not turn latency. - Physical-skill games: Physical running and stealth are modeled by the speed parameter, obstacle geometry, and raycast line-of-sight. - What is lost: The physical excitement of real hiding, the adrenaline of sprinting, and the actual sensory cues of sound/shadow. ## Implementation hints - Moderate complexity due to simultaneous turns and line-of-sight calculations. Use simple 2D Bresenham's line algorithm for fast raycasting. - First milestone: Create the grid and the simultaneous movement engine, then implement the line-of-sight detection, and finally add the *mirsa* racing and capturing logic.

Download as JSON

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

Download JSON
Cite this page

Syrian Hide and Seek (Syria). (n.d.). Games of Countries. https://gamesofcountries.com/games/sy-syria/al-tamimeh

@misc{gamesofcountries-sy-syria-al-tamimeh,
  title = {Syrian Hide and Seek (Syria)},
  year = {n.d.},
  publisher = {Games of Countries},
  url = {https://gamesofcountries.com/games/sy-syria/al-tamimeh}
}