Formal Specification
Ring Guessing Deduction Game — Formal Specification
Tunisia · guessing-hidden · Confidence: medium
← Back to game pageImplementation Suitability
Suitability: engine-readyFidelity: directGap-fills: 1
# Laab el-Khatem — AI-optimized specification
## General idea
A hidden-information deduction and bluffing game. One team secretly places a ring in 1 of 8 closed fists (2 fists per agent across 4 agents). The opposing detector iteratively eliminates empty hands or makes a direct identification accusation. Eliminating the ring hand by mistake awards points to the hiders; pinpointing the ring transfers possession.
## Formal model
- **Abstraction taken:** Hand selection and bluffing cues are formalized as a discrete hidden-assignment game over 8 hands `Hand_0..Hand_7`. Bluffing noise and detection accuracy are modeled via bayesian belief updates.
- **Player/agent count:** 2 teams of 4 agents (Team 0 and Team 1).
- **Turn model:** Round-based. Phase 1: secret ring assignment. Phase 2: iterative elimination by detector agent.
## Formal rules
R1. Setup: Hiding team has 4 agents, each providing 2 hands, forming `Hands = [0..7]`. Scores `Score_0 = 0`, `Score_1 = 0`.
R2. Phase 1 (Secret Hiding):
- Hiding captain selects `RingHand in {0..7}` uniformly or strategically.
- Each hand `k` produces a observable nervousness signal: `Signal_k = (k == RingHand ? TrueSignal : FalseSignal) + Noise`.
R3. Phase 2 (Iterative Elimination / Accusation):
- The detecting captain observes all active hands and repeatedly chooses either:
- `eliminate_empty(k)`: Order hand `k` to open.
- If `k != RingHand`: Hand `k` is safely removed from active set `ActiveHands`. Detecting captain continues turn.
- If `k == RingHand`: Detector committed a fatal elimination error! Round ends immediately; Hiding team scores `+count(ActiveHands) points`.
- `accuse_ring(k)`: Final direct accusation that hand `k` holds the ring.
- If `k == RingHand`: Detector is correct! Round ends; Detecting team scores `+1 point` and wins hiding rights for next round.
- If `k != RingHand`: Accusation failed; Hiding team scores `+count(ActiveHands) points`.
R4. Role Swap: If the detector correctly finds the ring via `accuse_ring()`, teams swap hiding/detecting roles for the next round.
R5. Win Condition: Points accumulate across rounds. First team to reach `TargetScore = 21` points wins.
## Edge cases
- Tie at the end condition: [gap-fill] If both teams reach 21 points in the same round, sudden-death single rounds continue until one team leads.
- A player with no legal move: Not applicable; eliminate and accuse actions are always valid on active hands.
- Repetition / cycles: Bounded strictly by `|ActiveHands| <= 8` per round.
- Simultaneous claims on the same resource or position: Sequential hand queries resolve all actions unambiguously.
- Player elimination mid-round; turn order repair: Not applicable; full team participates in every round.
- Resource exhaustion: Hands set monotonically decrements from 8 to 1.
- Termination guarantee: Enforced by positive point accumulation toward the 21-point threshold.
## State representation
- Persistent state: `scores: Array[2][int]`, `hiding_team_id: int in {0, 1}`.
- Ephemeral state: `ring_hand: int in [0..7]`, `active_hands: Set[int]`, `signals: Array[8][float]`.
- Observability: Imperfect information; `ring_hand` is known strictly to the Hiding Team; detecting captain observes `active_hands` and probabilistic `signals`.
## Action space
- `hide_ring(hand_id)`: Hiding captain action.
- `eliminate_empty(hand_id)`: Detector action.
- `accuse_ring(hand_id)`: Detector final accusation.
## Playability notes
- Single-player-vs-AI: Highly compelling Bayesian deduction / poker-style bluffing game against AI hiders.
- Group games (n>4): Naturally scales to any team size `M` (yielding `2*M` candidate hands).
- Physical-skill games: Micro-expression detection abstracted into noisy signal distributions.
- What is lost: The live psychological tension of locking eyes across a traditional Bedouin rug.
## Implementation hints
Bayesian belief update engine over 8 discrete states. First milestone: implement secret assignment and deterministic elimination/scoring rules before adding noisy perceptual signal vectors.
Download as JSON
Download the full game corpus entry for this game as structured JSON.
Download JSONCite this page
Ring Guessing Deduction Game (Tunisia). (n.d.). Games of Countries. https://gamesofcountries.com/games/tn-tunisia/laab-el-khatem-tunisia
@misc{gamesofcountries-tn-tunisia-laab-el-khatem-tunisia,
title = {Ring Guessing Deduction Game (Tunisia)},
year = {n.d.},
publisher = {Games of Countries},
url = {https://gamesofcountries.com/games/tn-tunisia/laab-el-khatem-tunisia}
}