Skip to main content
Formal Specification

Five Stones / Portuguese Knucklebones — Formal Specification

Portugal · dexterity-small · Confidence: low

← Back to game page

Implementation Suitability

Suitability: engine-readyFidelity: probabilistic-proxyGap-fills: 5
# Jogo dos Ossinhos — AI-optimized specification ## General idea A turn-based dexterity race through 5 escalating juggling phases with 5 stones. Agents must complete single scoops (1-by-1), double scoops (2-by-2), triple scoops (3+1), quadruple scoops (all 4), and a back-of-hand catch test to accumulate points towards a 50-point winning threshold. ## Formal model - **Abstraction taken:** Rapid hand-eye coordination (airborne throw, ground stone scoop without disturbing neighbor stones, and airborne catch) is modeled as a compound Bernoulli trial parameterized by phase scoop count and agent dexterity. - **Player/agent count:** 2 agents (1v1). - **Turn model:** Sequential alternating turns across 5 phases. ## Formal rules R1. Setup: Target win score `S_win = 50` points. Both agents start at `phase = 1`, `points = 0`. R2. On an agent's turn, the active agent attempts their current `phase in {1, 2, 3, 4, 5}`: - **Phase 1 (Ones):** 4 sequential attempts of (throw 1 -> pick 1 -> catch 1). Per-attempt success: `P(succ_1) = 0.90` [gap-fill]. Phase succeeds if all 4 attempts succeed. - **Phase 2 (Twos):** 2 sequential attempts of (throw 1 -> pick 2 -> catch 1). Per-attempt success: `P(succ_2) = 0.80` [gap-fill]. Phase succeeds if both attempts succeed. - **Phase 3 (Threes):** 1 attempt of (pick 3) followed by 1 attempt of (pick 1). Combined success: `P(succ_3) = 0.70` [gap-fill]. - **Phase 4 (Fours):** 1 attempt of (throw 1 -> sweep all 4 -> catch 1). Success: `P(succ_4) = 0.55` [gap-fill]. - **Phase 5 (Back-of-Hand):** Toss all 5 stones. Number of stones caught on back of hand `K_back ~ Binomial(5, 0.75)`. Secondary palm catch success: `K_caught ~ Binomial(K_back, 0.85)`. Points awarded = `K_caught * 10`. R3. Turn Progression: - If Phases 1..4 succeed: Agent advances to the next phase immediately on the same turn (or on next turn). - If Phase 5 succeeds (points scored > 0): Agent resets to Phase 1, keeping accumulated points. - If any phase fails: Turn ends immediately; agent stays at current phase for their next turn. R4. Win condition: The first agent to accumulate `>= 50` points wins the match. ## Edge cases - Tie at the end condition: First agent to reach 50 points wins immediately; turns are sequential so ties are impossible. - A player with no legal move: Not applicable; attempting the current phase is always valid. - Repetition / cycles: Score is monotonically non-decreasing; bounded by the 50-point win condition. - Simultaneous claims on the same resource or position: Turns are sequential. - Player elimination mid-round: If an agent disconnects, active agent completes turns to reach 50. - Resource exhaustion (pieces, tokens, space, time): Hard round limit of 30 turns per agent [gap-fill]. - Termination guarantee: Expected points per completed 5-phase cycle > 30 points, ensuring termination within ~15–20 turns. ## State representation - Persistent state: `scores: [int, int]`, `current_phases: [int, int]` (values 1..5), `active_agent: int`. - Per-turn ephemeral state: `phase_attempts: [bool]`, `stones_caught: int`, `points_earned: int`. - Minimal observable information: Fully observable public state. ## Action space - `execute_phase(dexterity_focus: Enum)`: where `dexterity_focus in {speed, balance, precision}`. ## Playability notes - Single-player-vs-AI: Engine-ready. Natural single-player dexterity benchmark or competitive race against an AI. - Group games (n>4): Extends naturally by iterating through agent roster. - Physical-skill games: Fine motor coordination and stone collision risks abstracted into stage-specific binomial/Bernoulli models. - What is lost in the adaptation: The tactile feel of river stones and muscle micro-adjustments. ## Implementation hints Low state complexity. Simple sequential phase advancement with binomial sampling. Suggested milestone 1: Implement the 5-phase FSM and simulate 100 matches to verify point accumulation dynamics.

Download as JSON

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

Download JSON
Cite this page

Five Stones / Portuguese Knucklebones (Portugal). (n.d.). Games of Countries. https://gamesofcountries.com/games/pt-portugal/jogo-dos-ossinhos

@misc{gamesofcountries-pt-portugal-jogo-dos-ossinhos,
  title = {Five Stones / Portuguese Knucklebones (Portugal)},
  year = {n.d.},
  publisher = {Games of Countries},
  url = {https://gamesofcountries.com/games/pt-portugal/jogo-dos-ossinhos}
}