Skip to main content
Formal Specification

Frost Chaser / Freeze Tag — Formal Specification

Serbia · chase-tag · Confidence: low

← Back to game page

Implementation Suitability

Suitability: engine-readyFidelity: discretizedGap-fills: 2
# Ledeni čiča — AI-optimized specification ## General idea A multi-agent pursuit-evasion game with a cooperative unfreezing mechanic played on a 2D discrete grid. A single Chaser agent attempts to freeze all N−1 Runner agents into static obstacles, while active Runners coordinate diversions and close-quarters touches to unfreeze immobilized teammates before the entire team is frozen. ## Formal model - **Abstraction taken:** The continuous playground is modeled as a 2D grid G of size 20×20 with toroidal or hard bounded perimeter. Time progresses in discrete synchronous ticks Δt. - **Player/agent count:** N agents (Agent 0 = Chaser, Agents 1..N−1 = Runners). - **Turn model:** Synchronous simultaneous movement and interaction per tick Δt. ## Formal rules R1. Initialization: Grid 20×20. Agent 0 (Chaser) spawns at (10,10) with velocity vchaser=1.2 cells/tick. Runners 1..N−1 spawn at random perimeter coordinates with velocity vrunner=1.0 cells/tick. Status initialized `frozen[i] = False` for all i∈[1..N−1]. R2. Synchronous Movement per Tick: - Chaser selects movement vector d→0. - Each unfrozen Runner i (`frozen[i] == False`) selects movement vector d→i. - Frozen Runners cannot move: d→j=(0,0). R3. Tag & Freeze Check: - If Chaser arrives within Euclidean distance Dtag≤1.0 of an unfrozen Runner i: - Runner i status updates: `frozen[i] = True`. Runner i is immediately immobilized at current grid cell. R4. Unfreeze & Rescue Check: - If an unfrozen Runner i (`frozen[i] == False`) moves within distance Drescue≤1.0 of a frozen Runner j (`frozen[j] == True`), and Chaser is not within tag distance of either: - Runner j status updates: `frozen[j] = False` (rescued). R5. Win Conditions: - **Chaser Victory:** If all N−1 Runners are frozen simultaneously (`all(frozen[1..N-1]) == True`), the Chaser wins. `score[Chaser] += 1`. - **Runners Victory:** If at least 1 Runner remains unfrozen after Tmax=300 ticks, the Runners win. R6. [gap-fill] Next Chaser Selection: In the next round, the first Runner who was frozen becomes the new Chaser. ## Edge cases - Tie at the end condition: Evaluated cleanly by R5 (either all are frozen, or time expires; outcomes are mutually exclusive). - A player with no legal move: Not applicable; unfrozen agents always have 4 directional movement choices plus STAY. - Repetition / cycles: Bounded by R5's hard limit of 300 ticks. - Simultaneous claims on the same resource or position: - If Chaser and an unfrozen Runner reach a frozen Runner on the exact same tick: the Chaser tags the rescuer (`frozen[rescuer] = True`); no unfreeze occurs. - Player elimination mid-round / turn order repair: [gap-fill] If an agent disconnects, they are removed from the active grid; remaining frozen/unfrozen status checks adjust to the new subset count. - Resource exhaustion: Grid and agent entities are invariant. - Termination guarantee: Guaranteed by the finite 300-tick round cap. ## State representation - Persistent state: `scores: dict[AgentID, int]`, `chaser_history: list[AgentID]`. - Per-tick state: `agent_positions: dict[AgentID, tuple[float, float]]`, `is_frozen: dict[AgentID, bool]`, `first_frozen_id: Optional[AgentID]`, `tick: int`. - Observable information: Fully observable 2D state for all agents. ## Action space - `move(dx: float, dy: float)`: Direction vector where ‖d→‖≤vmax. - `rescue(target_id: int)`: Explicit unfreezing intent when in range. ## Playability notes - Single-player-vs-AI: Highly dynamic; player as Chaser tests multi-agent trapping heuristics, or as lone survivor executes evasive rescue loops against aggressive AI Chaser. - Group games (n>4): Standard multi-agent benchmark for cooperative multi-agent RL (e.g. StarCraft / Predator-Prey variants). - Physical-skill games: Sprinting agility and evasive maneuvers mapped to relative velocity parameters (vchaser>vrunner). - What is lost in the adaptation: The somatic suspense of standing perfectly motionless as an ice statue while a chaser brushes past is replaced by a frozen boolean flag. ## Implementation hints Standard continuous 2D plane or discretized gridworld environment. First milestone: implement predator-prey tag state without rescues, then add unfreezing condition and verify equilibrium between chaser speed advantage and runner rescue rate.

Download as JSON

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

Download JSON
Cite this page

Frost Chaser / Freeze Tag (Serbia). (n.d.). Games of Countries. https://gamesofcountries.com/games/rs-serbia/ledeni-cica

@misc{gamesofcountries-rs-serbia-ledeni-cica,
  title = {Frost Chaser / Freeze Tag (Serbia)},
  year = {n.d.},
  publisher = {Games of Countries},
  url = {https://gamesofcountries.com/games/rs-serbia/ledeni-cica}
}