Formal Specification
Mallanpara Mock-Warfare Tournament — Formal Specification
Australia · stick-team · Confidence: low
← Back to game pageImplementation Suitability
Suitability: adaptation-requiredFidelity: discretizedGap-fills: 4
# Prun — AI-optimized specification
## General idea
Continuous mutual missile exchange between two lines, where ammunition circulates rather than depletes — anything thrown at you becomes yours to throw back. The attested skill is not accuracy but the last-moment evasion, and above all the clean catch.
## Formal model
- **Abstraction taken:** the two lines become ranks on a grid; a throw resolves as a hit check against the target's chosen response. Missile flight is a one-tick delay.
- **Player/agent count:** two teams of five to eight.
- **Turn model:** ticked and simultaneous — the source is explicit that play is continuous.
## Formal rules
R1. Two teams face each other across an open space, 20–30 units apart.
R2. Each tick, a player holding a missile may `throw(target)`; a player without one may `retrieve()` or reposition.
R3. A targeted player may `dodge()` or `catch()` on the tick the missile arrives.
R4. A dodge reduces hit probability; a catch, if successful, both avoids contact and gains the missile. A failed catch counts as a contact. [gap-fill] No probabilities are given.
R5. A clean catch is the highest-skill outcome and is scored as such. [gap-fill] The source calls it skilful but attests no scoring; a catch is worth more than a dodge here by construction.
R6. Missiles landing in the field may be retrieved by either team and thrown back — ammunition circulates and is never consumed.
R7. Play is continuous; there are no turns.
R8. [gap-fill] A contact does not eliminate the player. The source does not say, and elimination would end a continuous tournament quickly.
R9. [gap-fill] The event ends by agreement; impose a tick cap for a bounded artifact.
## Edge cases
- **Tie at the end condition:** with no attested scoring (R5), nothing can properly tie. Under imposed contact-counting, equal counts stand as a draw.
- **Player with no legal move:** never — retrieving, repositioning and dodging are always available.
- **Repetition / cycles:** because ammunition circulates (R6) and contacts do not eliminate (R8), the exchange can continue without limit. Only R9's cap ends it. This is faithful: a prun was an event, not a race to a score.
- **Simultaneous claims:** two players reaching one grounded missile resolve to the nearer, then lower index. Mutual simultaneous contacts both count.
- **Mid-round elimination and turn-order repair:** nobody is eliminated (R8), so no turn-order repair arises.
- **Resource exhaustion:** missiles are explicitly *not* exhausted. Nothing else is consumed.
- **Termination guarantee:** none from the attested rules; R9 supplies it and is flagged as imposed.
## State representation
Persistent: line positions, separation, missile count in circulation, tick cap.
Per-player: team, cell, missiles held, response chosen this tick, contacts taken, catches made.
Per-turn ephemeral: missiles in flight with target and arrival tick; missiles lying in the field.
Information set: fully observable — the throw is visible in flight, which is what makes last-moment evasion possible.
## Action space
- `throw(target_id)` — precondition: actor holds a missile.
- `dodge()` / `catch()` — precondition: a missile is arriving at the actor this tick.
- `retrieve(missile_id)` — precondition: missile is grounded within reach.
- `move(direction)`.
## Playability notes
- **Single-player-vs-AI:** good. The catch-versus-dodge choice under an incoming missile is a clean risk decision: catching gains ammunition and status but fails more often than dodging.
- **Group games (n>2):** already two teams; reduces to small sides with the exchange intact.
- **Physical-skill model:** hit probability by distance and response; catch success as a separate, lower probability.
- **What is lost:** everything that made it a tournament. The war cries, the clan identification, the dispute settlement and the display before onlookers are the *purpose*; the spec models only the missile exchange. Note also that the traditional missiles are not described in the source — only the Commission's foxtail substitute — so the projectile itself is undocumented.
## Implementation hints
Make catching strictly better in payoff and strictly worse in probability than dodging, or R5 does nothing and every agent dodges. Do not add elimination: it is unattested and it converts an open-ended tournament into a quick knockout, which misrepresents the event.
Download as JSON
Download the full game corpus entry for this game as structured JSON.
Download JSONCite this page
Mallanpara Mock-Warfare Tournament (Australia). (n.d.). Games of Countries. https://gamesofcountries.com/games/au-australia/prun
@misc{gamesofcountries-au-australia-prun,
title = {Mallanpara Mock-Warfare Tournament (Australia)},
year = {n.d.},
publisher = {Games of Countries},
url = {https://gamesofcountries.com/games/au-australia/prun}
}