Skip to main content
Formal Specification

Czech Knucklebones (Five Stones) — Formal Specification

Czechia · dexterity-small · Confidence: high

← Back to game page

Implementation Suitability

Suitability: engine-readyFidelity: probabilistic-proxyGap-fills: 1
# Drápaná — AI-optimized specification ## General idea A sequential dexterity progression game where agents advance through 5 distinct physical skill stages with 5 stones (singles, doubles, triples, quadruples, and hand-back flips). A turn ends on the first execution failure, and an agent resumes from their saved progress checkpoint on their subsequent turn. ## Formal model - **Abstraction taken:** Physical hand sweeps, pebble separation, and hand-flip catches are modeled as Bernoulli trials parameterized by stage difficulty Ds, stone clustering distance δ, and agent dexterity attribute θi∈[0,1]. - **Player/agent count:** N agents (2≤N≤6, standard N=2). - **Turn model:** Sequential round-robin rotation. Active agent continues performing stage attempts until a fault occurs or Stage 5 is completed. ## Formal rules R1. Setup: Stages S={1,2,3,4,5}. Each agent i maintains a persistent progress state (si,ki), initialized to (1,1), where si is current stage and ki is sub-step within the stage. R2. Stage definitions: - Stage 1 (Singles): 4 sub-steps (k∈{1,2,3,4}). Each sub-step requires sweeping 1 stone. Base success probability: P1(θ)=0.85+0.10θ. - Stage 2 (Doubles): 2 sub-steps (k∈{1,2}). Each requires sweeping 2 stones simultaneously. Base probability: P2(θ)=0.70+0.15θ. - Stage 3 (Triples): 2 sub-steps (k=1: sweep 3 stones, k=2: sweep 1 stone). Base probability: P3(θ)=0.60+0.20θ. - Stage 4 (Fours): 2 sub-steps (k=1: deposit 4 stones, k=2: sweep 4 stones). Base probability: P4(θ)=0.50+0.25θ. - Stage 5 (Hřbety / Back of Hand): 1 multi-catch step. Catching m∈{1..5} stones on back of hand, then flipping to palm. Success requires m≥1 and catching all m in palm. Base probability: P5(θ)=0.40+0.30θ. R3. Turn Execution: - Active agent attempts current sub-step (s,k). - If successful: (s,k) advances to the next sub-step/stage. Agent immediately attempts next sub-step. - If fault occurs: Agent's turn terminates immediately. State (s,k) is preserved. Turn passes to agent (i+1)(modN). R4. Win condition: The first agent to successfully complete Stage 5 wins the game immediately. R5. [gap-fill] Turn cap / Round limit: If no agent completes Stage 5 within 50 complete rotation rounds, the agent with the highest stage progress (si,ki) is declared the winner. ## Edge cases - Tie at the end condition: Single winner upon Stage 5 completion; turn cap tie broken by lowest agent ID. - A player with no legal move: Not applicable; an agent always attempts their current stage sub-step. - Repetition / cycles: Bounded by R5 maximum 50-round cap. - Simultaneous claims on the same resource: Not applicable; strictly sequential turns. - Player elimination mid-round: If an agent forfeits, they are removed; rotation continues. - Resource exhaustion: Fixed 5 stones invariant across all turns. - Termination guarantee: Strict monotonically non-decreasing progress checkpoints bounded by R5 round limit. ## State representation - Persistent state: `progress: list[tuple[int, int]]`, `active_agent: int`, `round_count: int`. - Ephemeral turn state: `current_attempt_outcome: bool`, `stones_on_back: int`. - Observable information: Fully observable (all progress checkpoints and outcomes are public). ## Action space - `attempt_substep()`: Trigger attempt for active agent's current stage checkpoint (s,k). ## Playability notes - Single-player-vs-AI: Natural; 1 human vs. 1 AI agent racing to finish Stage 5. - Group games (n>4): Supported up to 6 agents in round-robin order. - Physical-skill games: Multi-phase hand kinematic routines are mapped to parameterized stage success functions. - What is lost: The micro-tactile feel of stone weight, finger skin friction, and hand-tilt stabilization. ## Implementation hints Low complexity. Model as a clean progress-ladder state machine with a Bernoulli trial dispatcher per stage. First milestone: implement single-agent progression through stages 1 to 5 with fixed probabilities.

Download as JSON

Download the full game corpus entry for this game as structured JSON.

Download JSON
Cite this page

Czech Knucklebones (Five Stones) (Czechia). (n.d.). Games of Countries. https://gamesofcountries.com/games/cz-czechia/drapana

@misc{gamesofcountries-cz-czechia-drapana,
  title = {Czech Knucklebones (Five Stones) (Czechia)},
  year = {n.d.},
  publisher = {Games of Countries},
  url = {https://gamesofcountries.com/games/cz-czechia/drapana}
}