Skip to main content
Formal Specification

Heaven and Hell (Hopscotch) — Formal Specification

Luxembourg · dexterity-throw · Confidence: high

← Back to game page

Implementation Suitability

Suitability: engine-readyFidelity: probabilistic-proxyGap-fills: 1
# Himmel an Häll — AI-optimized specification ## General idea A discrete sequential grid-navigation and dexterity game where agents toss a marker stone into sequentially numbered court squares (1 through K), then hop through the single and double cells of the court avoiding lines, hazard cells (*Häll*), and the marker cell, retrieve the marker while balancing, and return to the baseline to advance their target stage. ## Formal model - **Abstraction taken:** The 2D chalk court is represented as an array of 8 discrete cells C=[c1,c2,c3,(c4,c5),c6,cHäll,cHimmel]. Stone tosses and hopping precision are parameterized by success probability functions. - **Player/agent count:** N agents (N≥1, standard 2 to 4). - **Turn model:** Sequential turn rotation; an agent continues attempting their current stage until a fault occurs. ## Formal rules R1. Setup: Court cells 1..7 initialized, with cell 6 containing hazard Häll and cell 7 as Himmel (Heaven). Each agent i has progress level Li=1 (target square). R2. Toss Phase: Active agent i executes `toss_stone(L_i)`. - Toss success probability Ptoss=f(skill,Li). - If toss fails (lands on line or wrong square): turn ends immediately, turn passes to agent i+1(modN). R3. Outward Hop Phase: Agent hops through cells from 1 to 7: - Must skip cell Li (cannot step into target cell). - Single cell: single-foot landing. Paired cells (c4,c5): dual-foot landing. - Hazard cell cHäll: must never be landed in (Pfault=1.0). - Reaching cell 7 (cHimmel): safe zone, agent turns 180 degrees. R4. Return Hop & Retrieval Phase: - Agent hops back towards cell 1 in reverse order. - In cell adjacent to Li, agent executes `retrieve_stone()`. - Agent hops over cell Li to baseline. R5. Stage Progression: If all hops and retrieval succeed with zero faults, Li←Li+1. R6. Win condition: First agent to successfully complete stage Li=6 (all numbered squares) and return safely wins the game. R7. [gap-fill] Max round cap: If no agent completes stage 6 within 30 full rotations, the agent with the highest progress level Li wins. ## Edge cases - Tie at the end condition: Resolved by higher stage level Li; if identical, fewest lifetime faults breaks tie (R7). - A player with no legal move: Not applicable — tossing the stone into target Li is always legal. - Repetition / cycles: Bounded by progress progression (Li strictly advances up to 6) and max rotation cap (R7). - Simultaneous claims on the same resource or position: Not applicable — agents take turns sequentially; only 1 agent on the court at any time. - Player elimination mid-round; turn order repair: If an agent withdraws, they are removed from rotation; turn passes i+1(modN). - Resource exhaustion: Not applicable — one persistent stone per agent. - Termination guarantee: Provably terminates in at most 30×N turns due to rotation cap. ## State representation - Persistent state: Progress array [L1,…,LN], fault counters [faults1,…,faultsN], current turn index i, global round counter. - Per-turn ephemeral state: Toss outcome ∈{success,fault}, hop trajectory log, stone location. - Observable information: Fully observable public game. ## Action space - `toss_stone(target_cell)`: Attempt toss into square Li. - `hop_sequence(path)`: Array of hop movements across court cells. - `retrieve_stone()`: Execute balance pickup at cell Li. ## Playability notes - Single-player-vs-AI: Natural — single player attempts to clear all 6 stages with minimal total faults or races against AI ghost trajectories. - Group games (n>4): Supported without rules modifications; turn queue rotates. - Physical-skill games: Motor balance modeled via probabilistic success checks P(hop)=1.0−fatigue×0.05. - What is lost: The physical muscle-burn of one-legged hopping, chalk dust on knees, and windy outdoor conditions. ## Implementation hints Discrete state-machine with simple probabilistic skill resolver. First milestone: implement linear single-player stage advancement from L=1 to L=6 with failure reset.

Download as JSON

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

Download JSON
Cite this page

Heaven and Hell (Hopscotch) (Luxembourg). (n.d.). Games of Countries. https://gamesofcountries.com/games/lu-luxembourg/himmel-an-haell

@misc{gamesofcountries-lu-luxembourg-himmel-an-haell,
  title = {Heaven and Hell (Hopscotch) (Luxembourg)},
  year = {n.d.},
  publisher = {Games of Countries},
  url = {https://gamesofcountries.com/games/lu-luxembourg/himmel-an-haell}
}