Skip to main content
Formal Specification

Ampali (Lefkadian Wooden Boules) — Formal Specification

Greece · dexterity-throw · Confidence: high

← Back to game page

Implementation Suitability

Suitability: engine-readyFidelity: probabilistic-proxyGap-fills: 2
# Ampali — AI-optimized specification ## General idea A discrete, frame-based target bowling and aerial knocking game on a 2D rectangular corridor. Agents take sequential throws aiming large wooden boules at a smaller target jack (balino). The game models two distinct tactical choices: continuous low-risk rolling toward the target, and high-risk lofted aerial strikes designed to blast opposing boules out of scoring contention. ## Formal model - **Abstraction taken:** Court modeled as a continuous 2D plane [0,L]×[0,W] (L=20.0m, W=3.5m). Ball rolling and aerial impacts are discretized into stochastic 2D landing coordinates (x,y) with 2D Gaussian dispersion and standard 2D rigid-body elastic displacement upon impact. - **Player/agent count:** 2 teams (Team 0 and Team 1), 2 agents per team; each team controls 4 total boules per frame. - **Turn model:** Dynamic sequential turn order per frame governed by the standard boules priority rule (the team furthest from target must throw next). ## Formal rules R1. Frame Setup: - Target *balino* coordinate 𝐩bal=(xbal,ybal) sampled uniformly from [12.0..18.0]×[1.0..2.5]. - Team 0 boule pool B0=4; Team 1 boule pool B1=4. - Played boules list 𝒫=∅. R2. Throw Action Selection: - Active agent chooses throw type ∈{roll,smash}, aim target coordinate 𝐩aim, and power F∈[1..5]. - **Roll Execution:** - Landing point: 𝐩=𝐩aim+𝝐roll, where 𝝐∼𝒩(𝟎,σroll2𝐈/skill). - **Smash Execution:** - Landing point: 𝐩=𝐩aim+𝝐smash (σsmash>σroll). - If landing point is within collision threshold r≤0.20m of an existing boule 𝐩k∈𝒫: - Target boule 𝐩k is displaced along impact vector by distance ddisp=0.8×F. R3. Dynamic Priority / Next Throw: - Compute best distance to balino: d0=min𝐛∈𝒫0⁡‖𝐛−𝐩bal‖, d1=min𝐛∈𝒫1⁡‖𝐛−𝐩bal‖. - If d0<d1 and B1>0: Team 1 throws next. - If d1<d0 and B0>0: Team 0 throws next. - If a team has 0 boules remaining, the other team throws all remaining boules consecutively. R4. Frame Scoring: - Once B0=0 and B1=0: - Identify winning team W∈{0,1} possessing the minimum distance dmin⁡=min⁡(d0,d1). - Let dloser=min𝐛∈𝒫loser⁡‖𝐛−𝐩bal‖. - Points scored by team W: Count of boules 𝐛∈𝒫W satisfying ‖𝐛−𝐩bal‖<dloser. - `match_score[W] += points`. R5. Win Condition: - First team whose `match_score` reaches or exceeds 11 points wins the match. R6. [gap-fill] Boundary Out-of-Bounds: Any boule whose coordinates exceed court boundaries (x<0,x>L,y<0,y>W) is marked `dead` and removed from scoring calculations for that frame. ## Edge cases - Tie at the end condition: If both teams' closest boules are at identical distances within measurement precision (|d0−d1|≤0.001m), [gap-fill] the frame is declared void (0 points awarded to both teams) and re-played. - A player with no legal move: Not applicable — throwing action is always available while pool B>0. - Repetition / cycles: Bounded — each throw strictly decrements the finite boule pool B0+B1; each frame terminates in exactly 8 throws. - Simultaneous claims on the same resource or position: Handled by continuous Euclidean coordinate representations and rigid-body impact displacement. - Player elimination mid-round; turn order repair: Not applicable (no eliminations; boules rotate among team members). - Resource exhaustion: Exactly 8 boules per frame; match score accumulates monotonically toward 11. - Termination guarantee: Non-zero expected points per frame (𝔼[points]≥1.0) ensures finite frames to reach 11 points. ## State representation - Persistent state: - `match_score`: Array of 2 integers `[score_0, score_1]`. - `frame_number`: Integer counter. - Per-frame state: - `balino_pos`: `(x, y)`. - `boules`: List of structs `[team_id, x, y, is_active]`. - `boules_remaining`: Array of 2 integers `[b0, b1]`. - Observable information: Fully observable. ## Action space - `throw_boule(throw_type, target_x, target_y, power)`: - `throw_type`: ∈{ROLL,SMASH}. - `target_x, target_y`: Floats within court limits. - `power`: Float ∈[1.0,5.0]. ## Playability notes - Single-player-vs-AI: Engine-ready. Excellent decision-making challenge balancing tactical rolling (pointing) against offensive aerial smashing (shooting). - Group games (n>4): Extends naturally to 2v2 (2 boules per player) or 3v3 (2 boules per player). - Physical-skill games: Physical arm release and gravel roll dynamics converted into stochastic landing distributions. - What is lost: Sensory feel of heavy seasoned olive wood, sound of clashing boules, and kafenion wine debates. ## Implementation hints - Implement as a sequential turn loop parameterized by the dynamic priority rule R3. - AI state evaluation requires computing Voronoi / distance rankings from the balino coordinate.

Download as JSON

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

Download JSON
Cite this page

Ampali (Lefkadian Wooden Boules) (Greece). (n.d.). Games of Countries. https://gamesofcountries.com/games/gr-greece/ampali

@misc{gamesofcountries-gr-greece-ampali,
  title = {Ampali (Lefkadian Wooden Boules) (Greece)},
  year = {n.d.},
  publisher = {Games of Countries},
  url = {https://gamesofcountries.com/games/gr-greece/ampali}
}