Formal Specification
Toad in the Hole (Sussex Pub Disc-Pitching Game) — Formal Specification
United Kingdom · dexterity-throw · Confidence: high
← Back to game pageImplementation Suitability
Suitability: engine-readyFidelity: probabilistic-proxyGap-fills: 0
# Toad in the Hole — AI-optimized specification
## General idea
A turn-based pub pitching game where agents toss brass discs ("toads") at a sloping lead-topped target table with a central 2-point hole and 1-point surface beds. Players race to reach exactly 31 points under a strict "bust" penalty (over-scoring cancels all turn gains) and an asymmetric opening handicap (2-3-4 throws) designed to balance first-turn initiative.
## Formal model
- **Abstraction taken:** The lead-topped table (24x24 inches) viewed from the 8-foot oche is modeled as a 2D impact probability space:
- `Hole Zone` (center radius `r < 1.125` in) -> 2 points.
- `Lead Table Zone` (`r in [1.125, 12.0]` in, excluding rim bounce) -> 1 point.
- `Miss / Off-Table Zone` -> 0 points.
Lead impact provides high energy absorption (damping coefficient `e = 0.15`), preventing significant disc roll.
- **Player/agent count:** 2 agents (`P1`, `P2`).
- **Turn model:** Sequential turns with variable throw counts per round: Turn 1 (P1: 2 toads), Turn 2 (P2: 3 toads), subsequent turns (4 toads each).
## Formal rules
R1. **Setup:** Scores initialize at `Score_P1 = 0, Score_P2 = 0`.
R2. **Turn Sequence & Throw Allocations:**
- Round 1, Player 1: throws `min(2, 31 - Score_P1)` toads.
- Round 1, Player 2: throws `min(3, 31 - Score_P2)` toads.
- Round `k >= 2`, active player:
- Points needed `rem = 31 - Score_active`.
- If `rem <= 3`: Allocated throws = `rem`.
- If `rem > 3`: Allocated throws = 4.
R3. **Pitch Action:** For each allocated toad, the agent selects:
- `pitch(aim_x: float, aim_y: float, spin_rate: float)`
R4. **Outcome Resolution per Pitch:**
- Impact point `(x, y)` sampled from 2D Gaussian around `(aim_x, aim_y)` with variance `sigma_skill`.
- If `sqrt(x^2 + y^2) < Radius_Hole`: Outcome = 2 points.
- Else if `abs(x) < Width_Table/2` and `abs(y) < Depth_Table/2`: Outcome = 1 point.
- Else: Outcome = 0 points.
R5. **Bust Rule & Score Commit:**
- Turn points `Pts_turn = sum(outcomes)`.
- If `Score_active + Pts_turn > 31`: **Bust!** `Score_active` remains unchanged (0 points earned this turn).
- If `Score_active + Pts_turn <= 31`: `Score_active = Score_active + Pts_turn`.
R6. **Victory Condition:** As soon as `Score_active == 31`, that agent wins immediately and the match terminates.
## Edge cases
- **Tie at the end condition:** Not applicable; turns are sequential and exactly 31 is hit by one agent first.
- **A player with no legal move:** Not applicable; `pitch()` is always legal.
- **Repetition / cycles:** Bounded because the game state `(Score_P1, Score_P2)` in `[0..31]^2` progresses with positive probability of reaching 31 on every turn where `rem <= 4`.
- **Simultaneous claims on the same resource:** Not applicable; sequential turns.
- **Player elimination mid-round:** Not applicable; 2-player closed match.
- **Resource exhaustion:** Exactly 4 toads reused each round.
- **Termination guarantee:** Guaranteed by finite Markov decision process where absorbing state `31` is reachable from all states with transition probability `P > 0`.
## State representation
- **Persistent state:** `Score_P1: int (0..31)`, `Score_P2: int (0..31)`, `turn_number: int`, `active_agent: int (1 or 2)`.
- **Per-turn ephemeral state:** `toads_thrown_this_turn: int`, `turn_points_accumulated: int`, `allocated_throws: int`.
- **Observable information:** Fully observable state and scoreboard.
## Action space
- `pitch(aim_x, aim_y)`: Continuous 2D aim vector on the lead table surface.
## Playability notes
- **Single-player-vs-AI:** Excellent pub game implementation. High endgame drama when an agent is on 29 or 30 points and must decide whether to aim for a risky 2-point hole or safe 1-point surface tap.
- **Group games reduction:** Team play is an aggregate series of four 1v1 legs.
- **Physical-skill games:** Modeled via 2D Gaussian landing error distribution.
- **What is lost in the adaptation:** The physical weight of the cast brass toad and the solid metallic "thud" against the lead bed.
## Implementation hints
- Very low computational requirements (sub-millisecond evaluation).
- State space is small: `32 x 32 = 1024` persistent states.
- First milestone: Write a Markov chain solver to find the optimal aim policy for any score from 0 to 30.
Download as JSON
Download the full game corpus entry for this game as structured JSON.
Download JSONCite this page
Toad in the Hole (Sussex Pub Disc-Pitching Game) (United Kingdom). (n.d.). Games of Countries. https://gamesofcountries.com/games/gb-united-kingdom/toad-in-the-hole
@misc{gamesofcountries-gb-united-kingdom-toad-in-the-hole,
title = {Toad in the Hole (Sussex Pub Disc-Pitching Game) (United Kingdom)},
year = {n.d.},
publisher = {Games of Countries},
url = {https://gamesofcountries.com/games/gb-united-kingdom/toad-in-the-hole}
}