Skip to main content
Formal Specification

Taiwanese Cops and Robbers (Save the Nation, Save the People) — Formal Specification

Taiwan · chase-tag · Confidence: low

← Back to game page

Implementation Suitability

Suitability: engine-readyFidelity: discretizedGap-fills: 3
# Guanbing Zhuo Qiangdao — AI-optimized specification ## General idea A team-based pursuit, evasion, and prison-rescue game modeling Taiwanese "Cops and Robbers" (*Koaⁿ-peng lia̍h chha̍t* / *Kiù-kok kiù-bîn*). Officers patrol an environment to chase and tag Robbers into a designated prison, while free Robbers use stealth, speed, and sacrificial decoys to infiltrate the prison zone and touch the captive chain to trigger a mass breakout ("Save the Nation!"). ## Formal model - **Abstraction taken:** Open continuous physical playgrounds are mapped to a 2D discrete grid graph G=(V,E) containing obstacles, cover cells, and a designated Prison Base zone. Movement speeds are modeled via movement points (MP) per discrete turn. - **Player/agent count:** 2 teams: Team Officers (O1…Ok) and Team Robbers (R1…Rk), with k∈[2..6] agents per team. - **Turn model:** Simultaneous tick-based phase: both teams submit movement/action vectors simultaneously; movements update, and contact/rescue events resolve deterministically. ## Formal rules R1. **Grid Environment:** - Graph G is an M×N grid (20×20 standard). - Specific subgrid P⊂V of size 2×2 is designated as the `Prison Base`. - Obstacle cells block line-of-sight and movement. - Each Officer has speed MPO=2 cells/tick. Each Robber has speed MPR=2 cells/tick (with a 1-tick sprint bonus MPR=3 once per 10 ticks). R2. **Initial Setup:** - Officers spawn at Prison Base P. - [gap-fill] Robbers have a 3-tick head start to scatter and occupy any non-prison cells before Officers can move. R3. **Capture Mechanics:** - If an Officer occupies or enters the same cell as an un-imprisoned Robber (or moves through an adjacent cell without cover), that Robber is tagged. - A tagged Robber is immediately placed into the `Prison Base` P and acquires state `Imprisoned`. - While imprisoned, Robbers may form a connected chain: the first prisoner must stay within P, and each subsequent prisoner may occupy an adjacent cell connected to the chain. R4. **Rescue Mechanics (*Save the Nation!* / 救國):** - A free Robber executes a rescue by entering an adjacent cell to any member of the captive chain (or the Prison Base P) without being tagged by an Officer on that same resolution tick. - Upon successful rescue: - All imprisoned Robbers are instantly liberated (`Free` state). - All liberated Robbers gain 2 ticks of immunity during which Officers cannot capture them, allowing them to scatter. R5. **Officer Guarding Constraint (*Anti-Camp Rule*):** - Officers are prohibited from remaining statically inside the 1-cell perimeter buffer of P for more than 2 consecutive ticks unless pursuing an active Robber within line-of-sight. R6. [gap-fill] **Win Conditions:** - **Officers Win:** All Robbers are simultaneously in the `Imprisoned` state (no free Robbers remain on the board). - **Robbers Win:** At least one Robber remains un-imprisoned when the global round limit (Tmax=100 ticks) expires. R7. [gap-fill] **Out-of-Bounds Disqualification:** Any Robber who leaves the grid boundary M×N is immediately classified as `Imprisoned` at P. ## Edge cases - **Tie at the end condition:** Not applicable; binary win/loss determined by whether all Robbers are imprisoned at Tmax or before. - **A player with no legal move:** Agents can always choose the `stay` action (0 movement) or move to an adjacent free cell. - **Repetition / cycles:** Infinite looping around obstacles is bounded by the strict Tmax=100 tick termination limit. - **Simultaneous claims on the same resource or position:** If an Officer tags a Robber on the exact tick the Robber reaches the prison chain, priority is given to the rescue if the Robber had initiative, otherwise capture resolves first. - **Player elimination mid-round:** Imprisoned Robbers remain in game state forming the rescue chain and are revived upon breakout. - **Resource exhaustion:** Grid graph and agent counts remain invariant. - **Termination guarantee:** Guaranteed by monotonic progress of discrete time ticks t∈[0,100]. ## State representation - **Persistent state:** - `grid_topology`: 2D obstacle map and prison coordinates P. - `officers`: list of `{id, x, y, mp_current}`. - `robbers`: list of `{id, x, y, state: [Free, Imprisoned], immunity_ticks}`. - `tick_counter`: integer in [0..100]. - **Per-tick ephemeral state:** - `intended_moves`: mapping of `agent_id` →(Δx,Δy). - `prisoner_chain`: ordered list of coordinates connected to P. - **Observable information:** Fully observable in full-visibility mode, or fog-of-war where visibility is bounded by line-of-sight raycasting. ## Action space - `move(direction)`: Step 1 or 2 cells in cardinal directions {North, South, East, West, Stay}. - `sprint(direction)`: Consume stamina to move 3 cells (Robber only, cooldown 10 ticks). ## Playability notes - **Single-player-vs-AI:** Highly playable; 1 human Officer commanding AI squad vs. AI Robber swarm (or vice versa) using multi-agent pathfinding (MAPF) and minimax / MCTS. - **Group games (n>4):** Reduces directly to 2-player control where Player 0 controls the Officer team and Player 1 controls the Robber team. - **Physical-skill games:** Physical sprinting stamina is discretized into MP allowances and cooldown-based sprint abilities. - **What is lost in the adaptation:** Physical hiding behind foliage, agility feints, and loud verbal shouting of "Save the Nation!". ## Implementation hints - Discretized 2D grid allows standard A* and Dijkstra pathfinding for Officer pursuit heuristics. - Prisoner chain is verified with simple breadth-first search (BFS) connectivity to P. - Suggested first milestone: Implement a 2 Officer vs. 3 Robber grid simulator with capture, chain formation, and rescue triggers.

Download as JSON

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

Download JSON
Cite this page

Taiwanese Cops and Robbers (Save the Nation, Save the People) (Taiwan). (n.d.). Games of Countries. https://gamesofcountries.com/games/tw-taiwan/guanbing-zhuo-qiangdao

@misc{gamesofcountries-tw-taiwan-guanbing-zhuo-qiangdao,
  title = {Taiwanese Cops and Robbers (Save the Nation, Save the People) (Taiwan)},
  year = {n.d.},
  publisher = {Games of Countries},
  url = {https://gamesofcountries.com/games/tw-taiwan/guanbing-zhuo-qiangdao}
}