Skip to main content
Formal Specification

Taiwanese Hoop Rolling — Formal Specification

Taiwan · dexterity-small · Confidence: low

← Back to game page

Implementation Suitability

Suitability: engine-readyFidelity: probabilistic-proxyGap-fills: 3
# Gun Tiehuan — AI-optimized specification ## General idea A traditional locomotion and balancing dexterity game modeling Taiwanese hoop rolling (*kún-chhia-kho͘*). Players guide a rolling metal ring across a track using a hooked wire pusher rod, balancing forward thrust against lateral wobble, steering through slalom gates, and preventing hoop falls (*chhia-tó*). ## Formal model - **Abstraction taken:** Continuous gyroscopic precession, rod-rim contact friction, and rolling ground contact are modeled as a 2D kinematic vehicle with rotational inertia, velocity v, roll angle (tilt) θ, and steering heading ϕ. - **Player/agent count:** 1 to 6 players competing in timed or head-to-head sprint heats. - **Turn model:** Real-time ticked simulation (Ttick=0.1 s) with simultaneous agent control inputs. ## Formal rules R1. **Track & Course:** - Track is a directed 2D spline path of length L=50.0 m with width W=2.0 m, populated by K slalom obstacle checkpoints. - Each racer starts at position s=0 with an upright hoop. R2. **Kinematic State Variables:** - Forward velocity: v∈[0,8.0] m/s. - Lateral tilt / lean: θ∈[−45∘,+45∘] (0∘=vertical). - Heading: ϕ∈[−π,π]. - Distance along track: s∈[0,L]. R3. **Agent Action Inputs (Per Tick):** - `push_thrust`: F∈[0,50] N (transmits forward acceleration). - `steer_torque`: τ∈[−10,+10] N⋅m (steers hoop heading ϕ). - `balance_correction`: β∈[−1.0,+1.0] (counteracts roll tilt θ). R4. [gap-fill] **Gyroscopic Stability & Fall Dynamics:** - Stability threshold: A minimum critical velocity vcrit=1.2 m/s is required for gyroscopic self-righting. - Tilt evolution per tick: θ(t+Δt)=θ(t)+(ghrimsin⁡θ(t)−kgyro⋅v(t)⋅ϕ˙(t)−kβ⋅β(t))Δt+ϵθ where ϵθ∼𝒩(0,σ2/skill). - **Hoop Fall (*Chhia-tó*):** If |θ(t)|≥40∘ or v(t)<0.2 m/s while s>0, the hoop falls. The player incurs a 2.0-second time penalty to right the hoop at the current position before resuming. R5. **Course Navigation & Checkpoints:** - Racers must pass each slalom checkpoint on the designated side. - Missing a slalom marker incurs a 5.0-second penalty added to the final time. R6. [gap-fill] **Win Condition:** The racer whose hoop crosses the finish line s≥L with the lowest elapsed time (including penalties) wins the match. R7. [gap-fill] **Time Limit & Disqualification:** Any racer failing to complete the course within Tmax=120 seconds is declared DNF (Did Not Finish). ## Edge cases - **Tie at the end condition:** If two racers finish with identical elapsed times (within 0.01 s), tie is broken by fewest recorded hoop drops; if still tied, a 15-meter sprint shoot-off is run. - **A player with no legal move:** Not applicable; input controls (F,τ,β) are continuous bounded vectors available on every tick. - **Repetition / cycles:** Racers stalling in place are bounded by the 120-second global cutoff (R7). - **Simultaneous claims on the same resource or position:** Parallel designated lanes eliminate direct physical collisions; in open-course modes, spatial avoidance is evaluated with elastic contact repulsive forces. - **Player elimination mid-round:** If an agent incurs 3 hoop drops in a single run, the run terminates immediately with a DNF. - **Resource exhaustion:** Track length is finite; time strictly advances. - **Termination guarantee:** Guaranteed by monotonic time progression toward Tmax=120 s. ## State representation - **Persistent state:** - `course_spline`: list of 2D waypoints and slalom gate coordinates. - `racers`: array of `{id, skill_rating, total_drops, finish_time, status}`. - **Per-tick ephemeral state:** - `elapsed_time`: float. - `racer_states`: array of `{s, x, y, v, theta, phi, penalty_timer}`. - **Observable information:** Fully observable track layout, opponent positions, and own vehicle telemetry. ## Action space - `control_hoop(thrust, steer, balance)`: - *Parameters:* F∈[0.0,50.0],τ∈[−10.0,10.0],β∈[−1.0,1.0]. ## Playability notes - **Single-player-vs-AI:** Highly playable; an AI agent solves an optimal control trajectory problem (e.g. Model Predictive Control / PID controller) balancing maximum forward speed against roll stability margins. - **Group games (n>4):** Scales naturally to multiplayer parallel heats or leaderboard time trials. - **Physical-skill games:** Hook contact dynamics are formalized into coupled differential equations for forward drive and roll tilt. - **What is lost in the adaptation:** Tactile vibrations of the wire hook buzzing against iron welds, dirt road bumps, and physical running fatigue. ## Implementation hints - Implement as a 2D vehicle dynamic simulation using standard Euler or Runge-Kutta 4th-order integration. - State vector per racer is 6 floats `[x, y, s, v, theta, phi]`. - Suggested first milestone: Implement single-agent straight-line rolling with velocity-dependent tilt stabilization and fall detection.

Download as JSON

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

Download JSON
Cite this page

Taiwanese Hoop Rolling (Taiwan). (n.d.). Games of Countries. https://gamesofcountries.com/games/tw-taiwan/gun-tiehuan

@misc{gamesofcountries-tw-taiwan-gun-tiehuan,
  title = {Taiwanese Hoop Rolling (Taiwan)},
  year = {n.d.},
  publisher = {Games of Countries},
  url = {https://gamesofcountries.com/games/tw-taiwan/gun-tiehuan}
}