Formal Specification
Rotten Egg — Formal Specification
Estonia · chase-tag · Confidence: high
← Back to game pageImplementation Suitability
Suitability: adaptation-requiredFidelity: discretizedGap-fills: 8
# Mädamuna (Rotten Egg) — AI-optimized specification
## General idea
A dynamic spatial ball-tagging and elimination game. A central player tosses a ball skyward and calls out an opponent's name while all other players scatter; if the named player drops the catch, they freeze the fleeing group with a "Stopp!" command and attempt to strike a frozen target with a thrown ball to assign penalty points ("rotten eggs") until a threshold triggers elimination or forfeit.
## Formal model
- **Abstraction taken:** The continuous 3D vertical ball toss, radial sprint scattering, and physical dodge maneuvers are discretized into a 2D polar/Cartesian grid. Dodge attempts are modeled as an evasion posture choice evaluated against throw accuracy.
- **Player/agent count:** agents ().
- **Turn model:** Sequential rounds comprising 3 phases: Toss & Scatter Freeze & Advance Throw & Penalty.
## Formal rules
R1. Setup:
- All agents gather at the origin .
- One agent is designated as the active Thrower .
- Penalty point map initialized for all agents .
R2. Phase 1 — Toss & Callout:
- submits `toss_ball(target_id)` specifying a target agent .
- While the ball is airborne ( ticks), all agents except sprint radially outward at velocity [gap-fill: airborne duration standardized to 3 ticks].
- attempts a clean catch with probability [gap-fill: base ].
- If caught: immediately calls a new name and re-tosses without stopping other runners.
- If dropped: The ball hits the ground at and claims the ball.
R3. Phase 2 — Freeze & Approach:
- On claiming a dropped ball, executes `call_stop()`.
- All fleeing agents must immediately freeze their coordinates.
- is granted 3 discrete approach steps toward any chosen frozen target , reducing distance by [gap-fill: approach steps fixed at 3.0m total].
R4. Phase 3 — Throw and Penalty Resolution:
- submits `throw_at(victim_id, power)`.
- submits a defensive reaction: `DODGE_LEFT`, `DODGE_RIGHT`, `DUCK`, or `ATTEMPT_CATCH`.
- Hit probability is evaluated as [gap-fill: hit resolver formula].
- Outcome:
- If hit succeeds: receives *mädamuna* point ().
- If throw misses or executes a successful `ATTEMPT_CATCH`: receives *mädamuna* point.
R5. Round Transition and Role Assignment:
- The agent who received the penalty point becomes the new for the subsequent round.
- All agents return to .
R6. Termination and Defeat:
- When any agent accumulates *mädamuna* penalty points, that agent loses the match and receives a penalty forfeit [gap-fill: threshold set to 3 points].
- The match terminates, and the agent with the fewest penalty points is declared the winner.
## Edge cases
- Tie at the end condition: If multiple agents hit the penalty threshold on the same round, all tied agents lose collectively [gap-fill: collective loss on tied threshold].
- A player with no legal move: Not applicable; agents always have valid scatter, call, step, or dodge actions.
- Repetition / cycles: Bounded because exactly 1 penalty point is awarded per round, guaranteeing termination in at most rounds [gap-fill: strict monotonic penalty accumulation bounds cycles].
- Simultaneous claims on the same resource or position: Not applicable; exactly one ball is contested, and only the named target may retrieve it.
- Player elimination mid-round; turn order repair: The recipient of the penalty point deterministically anchors the next round toss, ensuring unambiguous continuity.
- Resource exhaustion: Points accumulate monotonically until the threshold is breached.
- Termination guarantee: Exactly 1 point is generated every round with no possible null round; the game strictly terminates in at most rounds [gap-fill: termination guarantee via strict point emission].
## State representation
- Persistent state:
- `penalty_points`: Map of Agent ID .
- `current_thrower`: Agent ID currently holding .
- `round_number`: Integer .
- Per-round ephemeral state:
- `agent_positions`: Map of Agent ID .
- `frozen_status`: Boolean array indicating frozen movement.
- `ball_state`: `AIRBORNE`, `GROUND`, `HELD`.
- Observable information: Penalty totals, current thrower identity, and agent frozen coordinates are public.
## Action space
- `toss_ball(target_id)`: Select target agent to receive ball.
- `scatter(angle)`: Choose radial sprint heading during airborne phase.
- `call_stop()`: Trigger freezing of all agent positions.
- `advance_steps(victim_id)`: Move up to 3 steps toward target.
- `throw_at(victim_id, power)`: Execute ball throw at frozen target.
- `react(reaction_type)`: Choose `DODGE_LEFT`, `DODGE_RIGHT`, `DUCK`, or `ATTEMPT_CATCH`.
## Playability notes
- Single-player-vs-AI: Highly natural party/arcade game. Human balances sprint heading, target selection, and timing against AI dodgers.
- Group games (n>4): Generalizes seamlessly to 4–15 players; 1v1 reduction is degenerate (no scatter group dynamic), minimum viable agent count is 3.
- Physical-skill games: Throwing accuracy and torso dodging are abstracted into polar distances and reaction matchup matrices.
- What is lost in the adaptation: The sudden physical panic of hearing your own name shouted, the laughter of frozen contorted body postures, and the linguistic humor of assigned nicknames.
## Implementation hints
Model as a three-state round loop. First milestone: Implement Phase 2 & 3 freeze-and-throw resolution before integrating Phase 1 airborne sprint angles.
Download as JSON
Download the full game corpus entry for this game as structured JSON.
Download JSONCite this page
Rotten Egg (Estonia). (n.d.). Games of Countries. https://gamesofcountries.com/games/ee-estonia/madamuna
@misc{gamesofcountries-ee-estonia-madamuna,
title = {Rotten Egg (Estonia)},
year = {n.d.},
publisher = {Games of Countries},
url = {https://gamesofcountries.com/games/ee-estonia/madamuna}
}