Skip to main content
Formal Specification

Madurese Bull Racing / Kerap Sape — Formal Specification

Indonesia · animal-contest · Confidence: high

← Back to game page

Implementation Suitability

Suitability: adaptation-requiredFidelity: probabilistic-proxyGap-fills: 3
# Karapan Sapi — AI-optimized specification ## General idea A high-velocity two-agent head-to-head sprint racing contest where agents control paired draft animals harnessed to a skidding sled along a 100-metre 1D track. Agents manage a continuous trade-off between acceleration whip pacing, bilateral bull synchronization, and sled balance stability to reach the finish line first without losing control. ## Formal model - **Abstraction taken:** Physical bovine biomechanics, sled friction, and dirt surface dynamics are abstracted into a 1D position state xk∈[0.0,100.0] m with velocity vk(t), drift offset yk(t)∈[−1.5,+1.5] m, and jockey balance metric Bk(t)∈[0.0,1.0]. - **Player/agent count:** 2 competing agents per heat (P0,P1). - **Turn model:** Simultaneous tick execution (Δt=50 ms). ## Formal rules R1. Setup: Both bull pairs start at track position x=0.0,v=0.0,B=1.0,y=0.0. Track distance Xfinish=100.0 m. R2. Agent control actions: - Every tick, each agent chooses an action a∈{WHIP_SURGE,STEER_CENTER,BRACE_BALANCE}. R3. Dynamic state updates: - `WHIP_SURGE`: Acceleration av=+2.5 m/s2, but balance decays ΔB=−0.05 and lateral drift noise Δy∼𝒩(0,0.152) [gap-fill]. - `STEER_CENTER`: Lateral correction toward y=0 (Δy=−0.5×y), neutral acceleration (av=0.0), balance recovery ΔB=+0.02. - `BRACE_BALANCE`: Rapid balance recovery ΔB=+0.08, slight braking deceleration av=−1.0 m/s2 [gap-fill]. R4. Risk & failure triggers: - If balance Bk≤0.0, the jockey falls from the sled (`status = JOCKEY_FALL`); velocity immediately decays to 0, and the agent is disqualified [gap-fill]. - If lateral drift |yk|>2.0 m, the bull pair collides with the track boundary fence (`status = CRASH`); the agent is disqualified. R5. Win condition: The first agent whose position xk≥Xfinish with active jockey status wins the heat. ## Edge cases - Tie at the end condition: If both bull pairs cross Xfinish on the exact same tick with position difference <1 cm, a re-run heat is conducted. - A player with no legal move: Not applicable; choosing a valid jockey action is always available. - Repetition / cycles: Bounded by monotonic positive acceleration and finite track distance Xfinish=100.0 m. - Simultaneous claims on the same resource or position: Both agents race in isolated parallel lanes; no physical collision between opposing bull pairs is modeled. - Player elimination mid-round: If an agent crashes or suffers a jockey fall, the surviving opponent wins immediately upon crossing the finish line. - Resource exhaustion: Not applicable. - Termination guarantee: Velocity v≥5.0 m/s ensures crossing within <20 seconds (400 ticks). Hard cutoff at Tmax⁡=500 ticks guarantees termination. ## State representation - Persistent state: - `jockeys`: array of 2 objects `{agent_id, x, y, velocity, balance, status: ACTIVE|JOCKEY_FALL|CRASH}`. - `current_tick`: integer 0..500. - Per-tick ephemeral state: - `actions`: array of selected actions `[a_0, a_1]`. - Observable information: Fully observable (all positions, velocities, balance gauges, and lane offsets are public). ## Action space - `control_jockey(action)`: `action` ∈{WHIP_SURGE,STEER_CENTER,BRACE_BALANCE}. ## Playability notes - Single-player-vs-AI: Highly playable as an arcade/tactical sprint management simulator against an AI jockey. - Group games (n>4): Easily extended into multi-heat knockout tournament brackets (e.g. 8-agent bracket of 1v1 heats). - Physical-skill games: Physical balance and equine control reflexes are abstracted into continuous balance and drift gauges. - What is lost in the adaptation: The thunderous roar of the crowd, saronen music, and the physical mud spray from bull hooves are abstracted into telemetry. ## Implementation hints Lightweight 1D/2D physics simulation with PID-like control dynamics. First milestone: Implement single-agent speed/balance tradeoff model with 100m finish detection before adding competitive 1v1 AI.

Download as JSON

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

Download JSON
Cite this page

Madurese Bull Racing / Kerap Sape (Indonesia). (n.d.). Games of Countries. https://gamesofcountries.com/games/id-indonesia/karapan-sapi

@misc{gamesofcountries-id-indonesia-karapan-sapi,
  title = {Madurese Bull Racing / Kerap Sape (Indonesia)},
  year = {n.d.},
  publisher = {Games of Countries},
  url = {https://gamesofcountries.com/games/id-indonesia/karapan-sapi}
}