Formal Specification
Meggars / Irish Pitching Horseshoes — Formal Specification
Ireland · dexterity-throw · Confidence: low
← Back to game pageImplementation Suitability
Suitability: adaptation-requiredFidelity: probabilistic-proxyGap-fills: 1
# Meggars / Cluiche na gCrúb — AI-optimized specification
## General idea
A target-throwing contest where two players pitch two heavy forged iron horseshoes (*meggars*) at an iron pin (*hob*) centered in a clay box 30 feet away. The core strategic challenge lies in executing high-rotation pitch arcs to encircle the pin for 3-point ringers while leveraging cancellation scoring to negate an opponent's successful throws.
## Formal model
- **Abstraction taken:** Horseshoe flight trajectory is modeled as a 2D landing coordinate `(x, y)` relative to the central hob `(0, 0)` on the clay box plane, combined with an orientation angle `θ ∈ [0, 2π)`. Error variance follows a 2D Gaussian distribution parameterized by thrower skill. A shoe is evaluated as a `Ringer` if distance `d < r_pin + r_shoe` and the shoe's open throat spans the pin without collision; `Leaner` if resting against the pin; or `Close` if `d ≤ 15 cm`.
- **Player/agent count:** 2 discrete agents (`P1`, `P2`).
- **Turn model:** Alternating frames (ends) where both agents pitch 2 shoes each, evaluated using cancellation scoring until a player reaches 21 points.
## Formal rules
R1. Frame initialization: The iron hob is at `(0, 0)`. Pitching distance `D = 9.0m`. Scores initialized: `score[P1] = 0`, `score[P2] = 0`. First thrower `P_lead` determined by coin toss.
R2. On an agent's end turn:
- `P_lead` throws 2 shoes consecutively: `pitch_shoe(turn_turns: {1.25, 1.75}, aim_bias)`.
- `P_follow` throws 2 shoes consecutively: `pitch_shoe(turn_turns, aim_bias)`.
R3. Outcome evaluation per shoe:
- Landing coordinate `(X, Y) ~ N((target_x, target_y), σ² * I)` where `σ = f(accuracy_skill)`.
- Euclidean distance `d = sqrt(X² + Y²)`.
- If `d ≤ 3.5 cm` and throat angle test passes: shoe is a `Ringer` (value = 3).
- Else if `3.5 < d ≤ 5.0 cm` and vertical angle test passes: shoe is a `Leaner` (value = 2).
- Else if `d ≤ 15.0 cm`: shoe is `Close` (value = 1, distance = `d`).
- Else: shoe is `Dead` (value = 0).
R4. Cancellation scoring resolution:
- Let ringers of `P1` be `R1` and `P2` be `R2`.
- If `R1 > R2`: `P1` scores `3 * (R1 - R2)` points; all non-ringer shoes are canceled.
- If `R2 > R1`: `P2` scores `3 * (R2 - R1)` points; all non-ringer shoes are canceled.
- If `R1 == R2` and `R1 > 0`: ringers cancel completely; scoring falls to leaners/closest shoes.
- If no uncanceled ringers exist:
* Player with the single closest shoe within 15 cm scores 1 point (or 2 points if both of their shoes are closer than the opponent's closest shoe).
* If one player has a `Leaner` and opponent does not, `Leaner` awards 2 points.
R5. Lead update & victory:
- The player who scored points in the current end becomes `P_lead` for the next end. If no points were scored (0–0 tie in end), `P_lead` remains unchanged.
- The first player to reach or exceed `score[P_i] ≥ 21` wins the match.
- [gap-fill] Turn limit: If 25 ends elapse without a winner reaching 21, the agent with `argmax(score[P_i])` is declared the winner.
## Edge cases
- Tie at the end condition: In sequential ends, the first player to cross 21 points wins. If tied at end 25, a sudden-death end is played.
- A player with no legal move: Not applicable; an agent always pitches 2 shoes toward the hob.
- Repetition / cycles: Governed by the 25-end match cap.
- Simultaneous claims on the same resource or position: Physical displacement of shoes already in the clay box is resolved sequentially: incoming shoes can knock an opponent's shoe away or push it closer.
- Player elimination mid-round: Forfeit awards immediate match victory to the opponent.
- Resource exhaustion: Exactly 2 shoes per player are reused each end.
- Termination guarantee: Monotonically progressing end counter capped at 25 ends ensures termination.
## State representation
- Persistent state: Scores `score[P1]`, `score[P2]`, end counter (1..25), active `P_lead` ID.
- Per-player state: Skill variance `σ`, throw records.
- Per-turn ephemeral state: Pitched shoe array `[(agent_id, shoe_index, X, Y, outcome_type)]`, end score breakdown.
- Observable information: Full visibility of clay pit layout, exact coordinates of landed shoes, and current match score.
## Action space
- `pitch_shoe(turn_turns: [1.0..2.0], aim_bias_x, aim_bias_y)`: Execute horseshoe pitch with target trajectory and aerodynamic rotation rate.
## Playability notes
- Single-player-vs-AI: Highly playable; AI agent optimizes pitch selection between aiming directly for ringers versus defensive blocking throws.
- Group games (n>4): Reduced to 2 players or 2 pairs alternating throws.
- Physical-skill games: Aerodynamic shoe rotation and clay impact dynamics modeled via parameterized 2D physics.
- What is lost: The distinctive metallic ring of heavy iron on iron and the earthy smell of wet forge clay.
## Implementation hints
Implement the shoe classification logic as a deterministic bounding cylinder test around `(0, 0)`. Cancellation scoring resolves as a clean priority comparator: `Ringers -> Leaners -> Euclidean Distance`.
Download as JSON
Download the full game corpus entry for this game as structured JSON.
Download JSONCite this page
Meggars / Irish Pitching Horseshoes (Ireland). (n.d.). Games of Countries. https://gamesofcountries.com/games/ie-ireland/meggars
@misc{gamesofcountries-ie-ireland-meggars,
title = {Meggars / Irish Pitching Horseshoes (Ireland)},
year = {n.d.},
publisher = {Games of Countries},
url = {https://gamesofcountries.com/games/ie-ireland/meggars}
}