Skip to main content
Formal Specification

Five Stones / Knucklebones — Formal Specification

North Macedonia · dexterity-small · Confidence: low

← Back to game page

Implementation Suitability

Suitability: engine-readyFidelity: probabilistic-proxyGap-fills: 4
# Пиљци (Пет камења) — AI-optimized specification ## General idea A turn-based sequential dexterity game where players advance through a discrete ladder of six progressively difficult hand-eye coordination stages (Singles → Pairs → Triples → Fours → Gate → Emperor). In each attempt, a stone is tossed into the air while ground stones are manipulated and gathered before the airborne stone is caught. ## Formal model - **Abstraction taken:** The physical motor reflexes of throwing, grasping ground pebbles without collision, and catching falling stones are abstracted into a stage-dependent success probability Psuccess(Stage,Skill). A failed attempt terminates the active player's turn immediately, preserving stage progress for their next turn. - **Player/agent count:** 1 to 4 agents in sequential rotation. - **Turn model:** Strictly sequential turns; a successful stage completion allows immediate progression to the next stage in the same turn. ## Formal rules R1. **Stage Ladder:** The game consists of 6 discrete stages: - Stage 1: `Ones` (4 individual catches, difficulty D1=0.15). - Stage 2: `Twos` (2 paired catches, difficulty D2=0.25). - Stage 3: `Threes` (1 single + 1 triple catch, difficulty D3=0.35). - Stage 4: `Fours` (1 quadruple scoop, difficulty D4=0.45). - Stage 5: `Gate` (4 flick-and-catches through finger arch, difficulty D5=0.55). - Stage 6: `Emperor` (Back-of-hand catch and palm sweep, difficulty D6=0.65). R2. **Player State:** Each agent i maintains: - CurrentStagei∈[1,6] (initially 1). - Pointsi∈ℕ (initially 0). - Completedi∈{False,True}. R3. **Turn Execution:** Active agent i submits action `attempt_stage(s)` where s=CurrentStagei. - The success roll is sampled: X∼𝒰(0,1). - If X≥Ds−θskill,i: The stage attempt succeeds. R4. **Progression on Success:** - If s<6: CurrentStagei←s+1, and agent i immediately executes the next stage in the same turn. - If s=6 (`Emperor`): The number of stones caught K∈[1,5] is sampled from Binomial(5,1−D6+θskill). Agent i scores Pointsi←Pointsi+K, resets CurrentStagei←1, and ends their turn. R5. **Fault on Failure:** If X<Ds−θskill,i, agent i commits a fault, CurrentStagei remains at s, and the turn passes to player (i(modN))+1. R6. [gap-fill] Match Termination:** The first agent to accumulate ≥15 points from Emperor completions wins the game. R7. [gap-fill] Maximum Turns Cap:** If no agent reaches 15 points within 50 rounds, the agent with the highest accumulated score (or highest current stage) wins. ## Edge cases - Tie at the end condition: If multiple agents cross 15 points in the same round, the agent with higher total points wins [gap-fill]. - A player with no legal move: `attempt_stage` is always legally available for the active player. - Repetition / cycles: Bounded; stage progress is non-decreasing within a cycle, and the 50-round cap enforces hard termination. - Simultaneous claims on the same resource or position: Turns are strictly sequential; each agent uses their own set of 5 stones. - Player elimination mid-round: Not applicable; all agents retain their stage progress throughout. - Resource exhaustion: Stones are never consumed or destroyed. - Termination guarantee: [gap-fill] Guaranteed by the 50-round upper bound and non-zero stage completion probabilities. ## State representation - **Persistent state:** Number of players N, player skill ratings {θskill,i}, active turn index, scores {Pointsi}, stage pointers {CurrentStagei}. - **Per-turn ephemeral state:** Active stage s, toss outcome X, Emperor stone count K. - **Observable information:** Fully observable; all player scores, active stages, and historical toss outcomes are public. ## Action space - `attempt_stage(stage_id)`: Parameter `stage_id` must match `CurrentStage` of the active agent. ## Playability notes - Single-player-vs-AI: Natural; player competes against 1 to 3 AI agents in turn-taking stage races. - Group games (n>4): Plays up to 4 agents; larger groups naturally subdivide into independent 2-4 player circles. - Physical-skill games: Rapid finger flexion, airborne pebble catching, and finger-arch sliding are abstracted into scalar stage difficulties. - What is lost in the adaptation: The tactile feel of cool river pebbles and the physical dexterity of balancing stones on the back of knuckles. ## Implementation hints Very low complexity. Straightforward finite state machine tracking stage indices 1..6 and a score accumulator. First milestone: verify sequential stage advancement and turn handover upon simulated fault rolls.

Download as JSON

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

Download JSON
Cite this page

Five Stones / Knucklebones (North Macedonia). (n.d.). Games of Countries. https://gamesofcountries.com/games/mk-north-macedonia/piljci

@misc{gamesofcountries-mk-north-macedonia-piljci,
  title = {Five Stones / Knucklebones (North Macedonia)},
  year = {n.d.},
  publisher = {Games of Countries},
  url = {https://gamesofcountries.com/games/mk-north-macedonia/piljci}
}