Skip to main content
Formal Specification

Traditional Rattan Ball Circle — Formal Specification

Malaysia · ball-team · Confidence: low

← Back to game page

Implementation Suitability

Suitability: engine-readyFidelity: discretizedGap-fills: 2
# Sepak Raga Bulatan — AI-optimized specification ## General idea A fully cooperative multi-agent rally-optimization game where 6 to 10 agents arranged in a circular topology select kicking styles and passing trajectories to maximize unbroken continuous airborne touches while avoiding ball-grounding faults or individual touch-limit violations. ## Formal model - **Abstraction taken:** The continuous flight trajectory of the rattan ball and full-body physical acrobatics are discretized into turn-by-turn pass and strike choices. Each strike choice carries a success probability P(success) dependent on agent skill, incoming pass quality, and strike difficulty. - **Player/agent count:** N cooperative agents (N∈[6..10], standard N=6). - **Turn model:** Sequential state transitions driven by the ball's location and active receiver. ## Formal rules R1. Agent topology: N agents occupy fixed angular positions θi=2πi/N for i∈[0..N−1] along a circle of radius R=3.5 meters. R2. State tracking: State contains `active_agent`, `ball_position` (r,θ,z), `ball_velocity` (vr,vθ,vz), `consecutive_touches` k∈[0..3], `rally_length` L∈ℕ, and `bonus_points` B∈ℕ. R3. Legal action set: On receiving the ball, the active agent i chooses an action tuple (M,j): - Strike maneuver M∈{Sepak Sila,Sepak Kura,Sepak Paha,Menanduk,Sepak Kuda (acrobatic)}. - Target recipient j∈[0..N−1]∖{i} (or self j=i if k<3). R4. Execution probability: - Pass success probability: Psuccess=Skilli(M)⋅Qualityincoming⋅(1−Difficulty(M)). - If successful: Rally length L←L+1; if M=Acrobatic, B←B+2. If j=i, k←k+1; else k←0. Next active agent becomes j. - If failed: Ball grounds (z=0); the rally terminates immediately. R5. [gap-fill] Round structure: The team receives R=3 independent attempts (rallies). Total score = maxr∈[1..R]⁡(Lr+Br). R6. [gap-fill] Termination guarantee: Since Psuccess<1.0 for all actions, expected rally length is strictly finite. Maximum rally cap is enforced at Lmax=1000 steps. ## Edge cases - Tie at the end condition: In a multi-team tournament, ties are broken by total cumulative touches across all 3 attempts. - A player with no legal move: Not applicable; an agent with the ball always has ≥5 distinct passing actions. - Repetition / cycles: Passing back and forth between two agents is bounded by the k≤3 self-touch rule and pass dispersion noise. - Simultaneous claims on the same resource: Ball coordinates dictate exactly one unique active agent per turn. - Player elimination mid-round: Not applicable; fully cooperative squad. - Resource exhaustion: None. - Termination guarantee: Stochastically bounded by failure probability >0 at every step and hard-capped at 1000 steps. ## State representation - Persistent state: `rally_count` (1..3), `best_score` (integer), agent skill profiles `Skill[N][M]`. - Ephemeral state: `active_agent_id` (0..N−1), `self_touches` (0..3), `current_rally_touches` (integer), `current_bonus` (integer). - Observable information: Fully observable by all agents. ## Action space - `strike_and_pass(maneuver_id, target_agent_id)`: Where maneuver_id∈[0..4] and target_agent_id∈[0..N−1]. ## Playability notes - Single-player-vs-AI: Natural; player controls one or all circle positions in cooperative pursuit of a high-score benchmark. - Group games (n>4): Native multi-agent environment (N=6); ideal for cooperative MARL (Multi-Agent Reinforcement Learning) and communication protocol development. - Physical-skill games: Acrobatic coordination is parameterized via skill matrices and pass accuracy distributions. - What is lost: The exact three-dimensional biomechanics of jumping reverse kicks (*sepak gunting*) is abstracted into discrete categorical actions. ## Implementation hints Extremely clean transition graph. Can be implemented in under 100 lines of standard Python/C++ with zero external physics engine dependencies. First milestone: evaluate random policy baseline vs. greedy passing policy to establish expected rally length.

Download as JSON

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

Download JSON
Cite this page

Traditional Rattan Ball Circle (Malaysia). (n.d.). Games of Countries. https://gamesofcountries.com/games/my-malaysia/sepak-raga

@misc{gamesofcountries-my-malaysia-sepak-raga,
  title = {Traditional Rattan Ball Circle (Malaysia)},
  year = {n.d.},
  publisher = {Games of Countries},
  url = {https://gamesofcountries.com/games/my-malaysia/sepak-raga}
}