Skip to main content
Formal Specification

Egg-and-Spoon Races — Formal Specification

Cyprus · dexterity-small · Confidence: low

← Back to game page

Implementation Suitability

Suitability: engine-readyFidelity: probabilistic-proxyGap-fills: 3
# Avgodromies — AI-optimized specification ## General idea A simultaneous discrete-time race simulation where agents control acceleration against an accumulating dynamic balance instability metric. Higher forward velocities generate higher destabilizing oscillations; if cumulative wobble exceeds a spoon-containment threshold, the egg drops, requiring stationary ticks to retrieve and reset before moving forward. ## Formal model - **Abstraction taken:** 1D linear track of length D (e.g. D=25 meters). Micro-balancing of the wrist/jaw is modeled as a 1D continuous stability state wt∈[0,1] (wobble index), updated via stochastic differential transitions conditioned on agent speed choice. - **Player/agent count:** N racing agents (A1,…,AN), where 2≤N≤8. - **Turn model:** Simultaneous tick-based simulation. All agents submit speed/stabilization actions simultaneously each tick. ## Formal rules R1. **Initial State:** Track distance D. For each agent i: position xi=0, wobble index wi=0.0, egg state EggStatei=BALANCED, and recovery timer Ri=0. R2. **Action Selection:** Each tick, an agent with EggStatei==BALANCED chooses an action ai∈{SLOW_STABILIZE,CRUISE,SPRINT}: - SLOW_STABILIZE: Advance Δx=1, Wobble update Δw=−0.3+𝒩(0,0.05). - CRUISE: Advance Δx=2, Wobble update Δw=+0.1+𝒩(0,0.1). - SPRINT: Advance Δx=3, Wobble update Δw=+0.35+𝒩(0,0.15). R3. **Wobble Clamping & Drop Threshold:** - Update wobble: wi←max⁡(0.0,wi+Δw). - If wi≥1.0 (critical threshold): - EggStatei←DROPPED. - DropCoordi←xi. - Ri←3 (agent must spend 3 recovery ticks picking up the egg) [gap-fill]. - No advance occurs on this tick (Δx=0). - Otherwise: xi←xi+Δx. R4. **Recovery Dynamics:** - If EggStatei==DROPPED: Agent's only legal action is `retrieve_egg()`. - Recovery timer decrements: Ri←Ri−1. - When Ri==0: EggStatei←BALANCED, wi←0.0, and agent resumes from DropCoordi. R5. **Finish Condition:** - The race finishes at the end of the tick where at least one agent satisfies xi≥D and EggStatei==BALANCED. - The winner is the agent with largest xi≥D. R6. [gap-fill] If multiple agents cross D simultaneously with balanced eggs, the agent with lower wobble index wi wins the tiebreaker. R7. [gap-fill] A maximum race limit Tmax⁡=100 ticks guarantees termination; if reached without crossing D, the agent with largest xi wins. ## Edge cases - **Tie at the end condition:** Broken by lowest wobble index wi in R6; exact ties share victory. - **A player with no legal move:** Balanced agents have 3 movement options; dropped agents have `retrieve_egg()`. - **Repetition / cycles:** Positions xi are non-decreasing; dropping the egg only delays advancement without reversing progress. - **Simultaneous claims on the same resource or position:** Parallel independent lanes — no collision or blocking between agents. - **Player elimination mid-round:** Not applicable — all agents race until finish line or timeout. - **Resource exhaustion:** Eggs cannot be permanently broken in standard mode; wobble resets upon recovery. - **Termination guarantee:** Guaranteed because xi increases with P>0 and Tmax⁡ in R7 bounds total elapsed ticks. ## State representation - **Persistent state:** Distance D, agent roster size N. - **Per-tick state:** For each agent i: position xi∈[0,D+10], wobble wi∈[0.0,1.0], egg status EggStatei∈{BALANCED,DROPPED}, recovery counter Ri∈[0..3], tick counter t. - **Observable information:** Fully observable — all positions, wobble meters, and egg states are public knowledge. ## Action space - `move(speed_mode)`: Valid when EggState==BALANCED, with speed_mode∈{SLOW_STABILIZE,CRUISE,SPRINT}. - `retrieve_egg()`: Valid and required when EggState==DROPPED. ## Playability notes - **Single-player-vs-AI:** Excellent risk-management balancing game where player regulates speed throttle to stay just below the wobble threshold. - **Group games (n>4):** Scales seamlessly to N agents due to decoupled parallel lanes. - **Physical-skill games:** Physical vestibular and motor stabilization is mapped to the internal wobble accumulator wt. - **What is lost:** The micro-tactile feel of a wooden spoon handle and physical egg rotation is collapsed into scalar variable wt. ## Implementation hints Treat agent policy optimization as a classic optimal control problem (or Markov Decision Process) with a stochastic penalty zone at w≥1.0. The optimal policy switches between `SPRINT` and `SLOW_STABILIZE` based on current w. First milestone: implement single-agent wobble accumulator and test threshold triggers before running multi-agent races.

Download as JSON

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

Download JSON
Cite this page

Egg-and-Spoon Races (Cyprus). (n.d.). Games of Countries. https://gamesofcountries.com/games/cy-cyprus/avgodromies

@misc{gamesofcountries-cy-cyprus-avgodromies,
  title = {Egg-and-Spoon Races (Cyprus)},
  year = {n.d.},
  publisher = {Games of Countries},
  url = {https://gamesofcountries.com/games/cy-cyprus/avgodromies}
}