Formal Specification
The Reconciliation (Guessing Slap Game) — Formal Specification
Egypt · guessing-hidden · Confidence: medium
← Back to game pageImplementation Suitability
Suitability: engine-readyFidelity: discretizedGap-fills: 2
# El-Solh — AI-optimized specification
## General idea
A hidden-identity deduction and social deception game. An anonymous striker delivers a tap to a blinded Seeker while all suspect agents generate behavioral bluffing signals. The Seeker processes observable behavioral telemetry to deduce and accuse the actual striker.
## Formal model
- **Abstraction taken:** Nonverbal human facial cues and body language are formalized into a vector of discrete behavioral features (eye contact jitter, respiration rate, posture deflection, vocal tremor) emitted by suspect agents, with noisy perturbation.
- **Player/agent count:** 1 Seeker vs. Suspects (standard formalization: 1 Seeker vs 3 Suspects; reduced 2-agent form: 1 Seeker vs 1 Striker + 2 automated decoys).
- **Turn model:** Discrete round structure (Selection → Strike → Signal Generation → Accusation → Resolution).
## Formal rules
R1. Setup: suspect agents form a line; 1 agent is the Seeker.
R2. Strike Phase: One suspect is chosen (or self-selects) to execute `strike()`.
R3. Signal Generation Phase:
- Each suspect generates an observable signal vector :
where models the agent's bluffing skill (higher skill reduces ).
R4. Accusation Phase:
- Seeker observes all without observing the striker selection step.
- Seeker outputs an accusation .
R5. Resolution:
- If : Accusation is `CORRECT`. Seeker earns +10 points; suspect becomes the new Seeker.
- If : Accusation is `INCORRECT`. Seeker earns 0 points (or -2 penalty); current Seeker remains for the next round.
R6. [gap-fill] Mercy Cap: If a Seeker fails 5 consecutive accusations, role rotation occurs automatically to the most recent striker.
R7. [gap-fill] Match Termination: The game ends after 10 total rounds. The agent with the highest accumulated score wins.
## Edge cases
- Tie at the end condition: If top scores are tied at round 10, tied agents play a single 1v1 tiebreaker round with 2 decoys.
- A player with no legal move: Not applicable; accusatory choice is always a valid index in .
- Repetition / cycles: Bounded by the 5-failure mercy cap (R6) and 10-round match limit (R7).
- Simultaneous claims on the same resource or position: One striker is selected per round via deterministic priority or random tiebreaker.
- Player elimination mid-round; turn order repair: Role swaps occur cleanly at round boundaries without player removal.
- Resource exhaustion (pieces, tokens, space, time): Rounds strictly expire at 10.
- Termination guarantee: Guaranteed by the finite 10-round limit.
## State representation
- Persistent state:
- `seeker_id`: agent identifier.
- `consecutive_seeker_failures`: integer.
- `agent_scores`: map of `agent_id` to integer.
- `round_number`: integer in `1..10`.
- Per-round ephemeral state:
- `true_striker`: agent identifier (hidden from Seeker).
- `observed_signals`: map of `suspect_id` to vector .
- Observable information: Signal vectors are public; true striker identity is hidden from Seeker.
## Action space
- Striker: `strike()` — execute the strike action.
- Suspect: `emit_bluff(intent)` — parameterize signal generation with bluffing strategy.
- Seeker: `accuse(suspect_id)` — select accused agent index.
## Playability notes
- Single-player-vs-AI: Natural; human player acts as Seeker reading AI signal distributions or as Striker attempting to spoof the AI.
- Group games (n>4): Parameterizes cleanly with any number of suspects .
- Physical-skill games: Physical slap sensation and nonverbal acting are abstracted into the feature vector and noise variance.
- What is lost in adaptation: The spontaneous comedic acting and peer laughter of live human social circles.
## Implementation hints
Statistical decision / Bayesian classification problem. Seeker agent can use Maximum Likelihood Estimation (MLE) or Logistic Regression over the signal features. First milestone: implement signal generation and Seeker bayesian estimation with constant bluff variance.
Download as JSON
Download the full game corpus entry for this game as structured JSON.
Download JSONCite this page
The Reconciliation (Guessing Slap Game) (Egypt). (n.d.). Games of Countries. https://gamesofcountries.com/games/eg-egypt/el-solh
@misc{gamesofcountries-eg-egypt-el-solh,
title = {The Reconciliation (Guessing Slap Game) (Egypt)},
year = {n.d.},
publisher = {Games of Countries},
url = {https://gamesofcountries.com/games/eg-egypt/el-solh}
}