Formal Specification
Koara (Spinning Propeller Toy) — Formal Specification
Australia · dexterity-throw · Confidence: low
← Back to game pageImplementation Suitability
Suitability: adaptation-requiredFidelity: probabilistic-proxyGap-fills: 2
# Koara — AI-optimized specification
## General idea
A single launch technique repurposed across several distinct objectives — accuracy, distance, height, or a cooperative catch-and-step-back rally — so the interesting design question is which scoring wrapper sits on top of one shared launch mechanic.
## Formal model
- **Abstraction taken:** the spin-and-release launch becomes a single `launch(technique)` action producing a flight outcome (height, distance, landing accuracy) drawn from a distribution parameterized by technique and skill.
- **Player/agent count:** any number, acting independently, or in pairs for the cooperative rally variant.
- **Turn model:** single committed action per attempt; sequential across attempts in a competition.
## Formal rules
R1. A player performs `launch(technique)`, where technique parameterizes grip and release timing.
R2. The launch resolves to `(height, distance, landing_position)` drawn from `f(technique, skill) + noise`. [gap-fill] no explicit height/distance/accuracy model is given; it is added because the source's variant games (accuracy, golf, height) each need a resolvable outcome.
R3. Scoring depends on the chosen mode: accuracy mode scores a hit if `landing_position` is within the target zone; golf mode counts launches needed to reach the goal; height mode compares peak height; partner-rally mode alternates `launch()` and `catch()` between two players, stepping back one unit after each success, until a catch fails.
R4. The player or pair with the best result under the chosen mode's win condition wins.
## Edge cases
- **Tie at the end condition:** equal results (e.g. same height, or both reaching the golf goal in the same count) are possible; [gap-fill] resolved by a decider attempt, since the source gives no explicit tie-break.
- **Player with no legal move:** not applicable — a launch attempt is always available.
- **Repetition / cycles:** the partner-rally mode repeats launch/catch indefinitely until a miss; this is the intended structure, not a defect, and terminates with probability 1 since catch success is less than certain.
- **Simultaneous claims:** not applicable — no shared resource is contested.
- **Mid-round elimination and turn-order repair:** not applicable.
- **Resource exhaustion:** not applicable in most modes; golf mode is bounded by the goal eventually being reached almost surely.
- **Termination guarantee:** guaranteed for single-attempt modes; for the rally mode, guaranteed because catch success probability is under 1, so an eventual miss occurs with probability 1.
## State representation
Persistent: chosen mode, target/goal position (if any).
Per-player: skill parameter, cumulative score or rally distance.
Per-turn ephemeral: this attempt's technique choice and resolved outcome.
Information set: fully observable — everyone sees each launch's outcome.
## Action space
- `launch(technique)` — precondition: none.
- `catch()` — precondition: rally mode, a propeller is currently airborne toward the actor.
## Playability notes
- **Single-player-vs-AI:** natural as a high-score/personal-best challenge in any of the four scoring modes.
- **Group games (n>4):** scales without change — players do not interact except in the partner-rally mode, which is inherently 2-player.
- **Physical-skill model:** the technique-to-outcome function (R2) is the entire model; this is a clean dexterity proxy with several selectable scoring lenses on the same underlying skill roll.
- **What is lost:** the object's regional material history (folded leaves vs. carved wood vs. Cherbourg's decorated cardboard-and-straw form) has no gameplay analogue and is preserved only in the entry's Origin section, not the spec.
## Implementation hints
Build the shared `launch()` outcome model once, then implement the four scoring modes as thin wrappers over it — do not duplicate the physics/probability logic per mode. The partner-rally mode is the most interesting to implement since it has genuine escalating risk (stepping back after each success).
Download as JSON
Download the full game corpus entry for this game as structured JSON.
Download JSONCite this page
Koara (Spinning Propeller Toy) (Australia). (n.d.). Games of Countries. https://gamesofcountries.com/games/au-australia/koara
@misc{gamesofcountries-au-australia-koara,
title = {Koara (Spinning Propeller Toy) (Australia)},
year = {n.d.},
publisher = {Games of Countries},
url = {https://gamesofcountries.com/games/au-australia/koara}
}