Skip to main content
Formal Specification

Philippine Hopscotch / House-Building Hopscotch — Formal Specification

Philippines · dexterity-throw · Confidence: low

← Back to game page

Implementation Suitability

Suitability: engine-readyFidelity: probabilistic-proxyGap-fills: 4
# Piko — AI-optimized specification ## General idea A turn-based sequential hopscotch and territory acquisition game on an 8-compartment graph. Agents alternate throwing a marker (*pamato*) into target compartments (1..8), hopping through the graph avoiding the target compartment and opponents' owned cells (*houses*), and executing a backward blind throw to permanently claim houses. The game concludes when all cells are claimed or jump traversals become impossible. ## Formal model - **Abstraction taken:** Court is modeled as a connected graph G=(V,E) where nodes V={1,2,3,4,5,6,7,8} represent compartments with foot landing constraints: {1,4,7}→single-foot, {2,3} and {5,6}→paired straddle, {8}→double-foot rest. Throwing precision and hopping balance are parameterized stochastic actions. - **Player/agent count:** 2 to 4 agents (P0..Pk−1). - **Turn model:** Sequential rotation across rounds. ## Formal rules R1. Setup: - Compartment ownership initialized: owner[v]=NONE for all v∈[1..8]. - Each agent starts at stage Si=1 and `houses_owned = 0`. R2. Throwing Action: - On active turn, agent Pi executes `throw_pamato(target_box = S_i)`: - Success probability: P(throw_land)=min⁡(0.95,max⁡(0.20,1.0−(0.08×Si−0.05×skill))). - If throw fails (lands off-box or on line): turn ends immediately (`fault = true`), Si remains unchanged, turn passes to next agent. R3. Hopping Traversal: - If throw succeeds, agent executes graph traversal from entrance to node 8 and back: - For each node v∈V: - If v==Si (contains pamato) or owner[v]∉{NONE,Pi}: node v must be skipped via jump. - If jump distance exceeds max jump reach (d>2 nodes) [gap-fill]: traversal fails. - Balance check sampled for traversal: P(balance)=max⁡(0.10,1.0−(0.05×number_of_jumps)). - If traversal succeeds: - If Si<8, advance Si←Si+1, and agent takes an immediate consecutive turn [gap-fill]. - If Si==8, agent transitions to `HOUSE_CLAIM_PHASE`. - If traversal fails: turn terminates, Si unchanged, turn passes. R4. House Claim Phase (*Pagtatayo ng Bahay*): - Agent executes `blind_throw()`: - Samples a landing box b∼Uniform(1..8) with success probability P(clean_landing)=0.60. - If b is currently unowned (owner[b]==NONE): - owner[b]←Pi, `houses_owned[P_i] += 1`. - Si resets to 1 for the next house cycle [gap-fill]. R5. Match Termination: - Match concludes when all 8 compartments are claimed, or when maximum turn limit Tmax⁡=100 is reached. - The agent with the maximum `houses_owned` wins. ## Edge cases - Tie at the end condition: If top agents have equal houses owned, tie is broken by highest current stage Si, then by sudden-death blind throw [gap-fill]. - A player with no legal move: If opponent houses block any legal topological jump across the court (gap >2 nodes), the blocked player must pass their turn. - Repetition / cycles: Bounded by hard turn horizon Tmax⁡=100 and monotonic cell ownership transitions (NONE→Pi). - Simultaneous claims on the same resource or position: Sequential turn order prevents simultaneous claims. - Player elimination mid-round: Not applicable; all players remain active until match termination. - Resource exhaustion: Court has exactly 8 claimable nodes. - Termination guarantee: Enforced by the finite turn cap Tmax⁡ and bounded house allocations. ## State representation - Persistent state: - `court_owners`: array of 8 integers (storing owner ID or -1). - `agent_stages`: array of size k (current target box 1..8). - `houses_count`: array of size k. - `active_agent`: integer 0..k−1. - `turn_counter`: integer 0..100. - Observable information: Fully observable public state. ## Action space - `throw_pamato(target_box)`: parameter 1..8. - `traverse_course()`: automated execution step evaluating topological reach and balance. - `blind_throw()`: executed at stage 8 to claim house. ## Playability notes - Single-player-vs-AI: Natural and strategic; agent evaluates whether to claim central nodes (nodes 4, 2, 3) to block opponent hop trajectories. - Group games (n>4): Reduces cleanly to 2 or 3 agents; 4 agents creates high territorial congestion. - Physical-skill games: Single-foot balance and throwing precision are abstracted into calibrated Bernoulli sampling. - What is lost in the adaptation: Physical foot precision near drawn chalk boundaries and body fatigue from continuous hopping. ## Implementation hints Graph data structure with 8 nodes and node-jump validation. First milestone: Implement single-agent stage progression (1 to 8) without opponents to verify throwing and traversal success distributions.

Download as JSON

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

Download JSON
Cite this page

Philippine Hopscotch / House-Building Hopscotch (Philippines). (n.d.). Games of Countries. https://gamesofcountries.com/games/ph-philippines/piko

@misc{gamesofcountries-ph-philippines-piko,
  title = {Philippine Hopscotch / House-Building Hopscotch (Philippines)},
  year = {n.d.},
  publisher = {Games of Countries},
  url = {https://gamesofcountries.com/games/ph-philippines/piko}
}