Skip to main content
Formal Specification

The Can (Post-tag hide-and-seek) — Formal Specification

Sweden · hide-seek · Confidence: high

← Back to game page

Implementation Suitability

Suitability: adaptation-requiredFidelity: discretizedGap-fills: 6
# Burken — AI-optimized specification ## General idea A high-stakes spatial hide-and-seek and base-defense game played in an obstacle-rich arena. A central Seeker guards a home landmark ("the can"), searches for concealed Hiders, and races back to base to "dunk" (eliminate) spotted opponents. Hiders balance stealth and speed, either sprinting to base to secure personal safety ("Egendunk") or executing a daring jailbreak ("Fridunk") to resurrect all captured teammates. ## Formal model - **Abstraction taken:** Continuous 3D terrain and foliage occlusion are mapped to an M×M 2D grid containing discrete obstacles (walls/trees) that block line-of-sight (LOS) rays calculated via Bresenham raycasting. The physical count is modeled as a countdown during which the Seeker has zero visibility. Sprinting to base is modeled as simultaneous path-length traversal over discrete ticks. - **Player/agent count:** N agents (N≥3; benchmark N=6: 1 Seeker S, 5 Hiders H1…H5). - **Turn model:** Phase 1 (Counting) advances via a 100-step counter; Phase 2 (Active Search & Race) operates on synchronous discrete time ticks (t=1…Tmax⁡). ## Formal rules R1. **Setup:** Arena is an M×M grid with a designated home base cell B=(x0,y0) and a prison zone adjacent to B. Obstacle cells 𝒪 block movement and LOS. One agent is assigned `SEEKER` at B; remaining agents are `HIDER`. R2. **Phase 1 (The Count & Hide):** - S is locked at B for C=20 setup ticks [gap-fill] with observation set ∅. - Hiders simultaneously execute movement actions v→∈{(0,0),(±1,0),(0,±1)} up to speed sH=1 to conceal themselves behind obstacles. R3. **Phase 2 (Active Search):** - On each tick t, all active agents simultaneously choose `move(dx, dy)` or `interact_base()`. - LOS Check: If an unoccluded straight-line ray exists between S and Hider Hi with dist(S,Hi)≤Rvis=15 cells [gap-fill], Hi is `SPOTTED`. R4. **Dunking a Spotted Hider:** - Once Hi is spotted, both S and Hi may race toward B. - If S reaches B and executes `dunk(H_i)` before Hi reaches B, Hi becomes `CAPTURED` and is moved to the prison zone. R5. **Self-Dunk (*Egendunk*):** - If an active Hider Hi reaches B and executes `dunk_self()` before S dunks them, Hi becomes permanently `SAFE` for the remainder of the round. R6. **Free-Dunk / Jailbreak (*Fridunk*):** - If an active (uncaptured, un-safed) Hider Hi reaches B while S is not at B and executes `fridunk()`, all currently `CAPTURED` prisoners are instantly liberated. - All liberated prisoners enter state `HIDING`, S is teleported back to B, and a mini-count of Cmini=5 ticks [gap-fill] is initiated where S is blinded while prisoners scatter. R7. **Round End & Role Transfer:** - The round terminates when every Hider is either `SAFE` or `CAPTURED`. - The first Hider captured in the round becomes the Seeker for the subsequent round. If all hiders reach safety, the current Seeker remains Seeker [gap-fill]. R8. **Timeout Limit:** - [gap-fill] If Tmax⁡=200 ticks elapse with active hiders remaining unfound, the round ends; all remaining hiders are ruled `SAFE`. ## Edge cases - Tie at the end condition: If Seeker S and Hider Hi reach B on the exact same tick, the tie resolves in favor of the Hider (safe) [gap-fill]. - A player with no legal move: Agents blocked by obstacles or boundaries may only choose unblocked orthogonal directions or (0,0). - Repetition / cycles: Infinite evasion loops or repetitive jailbreaks are bounded by the hard tick ceiling Tmax⁡=200. - Simultaneous claims on the same resource or position: The base cell B can support simultaneous occupancy; action priority resolves: `fridunk()` > `dunk_self()` > `dunk(target)`. - Player elimination mid-round: Captured players sit in prison until freed by a `fridunk()`; if a captured player disconnects, prison count decreases. - Resource exhaustion (pieces, tokens, space, time): Spatial limits are fixed by the M×M grid. Time is bounded by Tmax⁡. - Termination guarantee: The monotonically increasing tick counter t→Tmax⁡=200 guarantees absolute termination. ## State representation - **Persistent state:** - `round_number`: integer. - `seeker_id`: agent ID. - `scores`: map of agent ID →{successful_captures,self_dunks,fridunks,times_captured}. - **Per-turn ephemeral state:** - `tick`: integer ∈[1,Tmax⁡]. - `phase`: `COUNTING`, `SEARCHING`, or `MINI_COUNT`. - `positions`: map of agent ID →(x,y)∈{1…M}2. - `hider_status`: map of Hi→{HIDDEN,SPOTTED,SAFE,CAPTURED}. - `capture_order`: list of captured agent IDs (to track the first capture). - **Observable information:** - Seeker observes: own position, positions of Hiders with valid LOS, and base state (empty / occupied). During counting, observation is empty. - Hiders observe: own position, Seeker position (if in LOS), base position B, and prison occupancy. ## Action space - `move(dx, dy)`: orthogonal or diagonal step (‖v→‖≤1). - `dunk(target_id)`: executed by Seeker at base B to capture a spotted target. - `dunk_self()`: executed by Hider at base B to achieve safety. - `fridunk()`: executed by Hider at base B to liberate all prisoners. ## Playability notes - Single-player-vs-AI: Highly dynamic and engaging — player as Seeker managing patrol routes and base tethering vs stealth AI, or player as Hider coordinating high-risk rescue sprints. - Group games (n>4): Reduces to 2 agents (1 Seeker vs 1 Hider), where the game becomes a pure duel of stealth, line-of-sight baiting, and base-sprinting (the `fridunk` rule becomes inactive with only 1 hider). - Physical-skill games: Physical sprinting speed and peripheral vision are parameterized by movement rate (sH,sS) and raycast visibility cones. - What is lost in the adaptation: The somatic suspense of holding one's breath behind a physical bush, ambient audio cues (twigs snapping), and the physical kick of a metal tin can. ## Implementation hints Moderate complexity. Requires an efficient 2D grid raycasting function (e.g. Bresenham or DDA) to evaluate visibility between Seeker and all Hiders each tick. Keep base interaction atomic at the end of each tick. First milestone: build a 20×20 grid with 5 obstacle blocks and test Seeker line-of-sight spotting against static hiders.

Download as JSON

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

Download JSON
Cite this page

The Can (Post-tag hide-and-seek) (Sweden). (n.d.). Games of Countries. https://gamesofcountries.com/games/se-sweden/burken

@misc{gamesofcountries-se-sweden-burken,
  title = {The Can (Post-tag hide-and-seek) (Sweden)},
  year = {n.d.},
  publisher = {Games of Countries},
  url = {https://gamesofcountries.com/games/se-sweden/burken}
}