Skip to main content
Formal Specification

Slab Stone Target Throwing — Formal Specification

Kosovo · dexterity-throw · Confidence: low

← Back to game page

Implementation Suitability

Suitability: engine-readyFidelity: probabilistic-proxyGap-fills: 3
# Loja e rrasave — AI-optimized specification ## General idea A turn-based team target-elimination contest. Two opposing teams take turns pitching stones at the opponent's standing stone slabs across a fixed distance. Each thrower selects an aim target and shot trajectory, balancing direct impact accuracy against the probability of multi-slab ricochets. The first team to topple all enemy slabs wins the round. ## Formal model - **Abstraction taken:** The 3D ballistics of throwing a flat stone across a 12-meter pitch are replaced by a probabilistic hit/topple function parameterized by agent skill and target selection. - **Player counts:** 2 teams (Team A, Team B), with K agents per team (default K=3). - **Turn model:** Alternating team volleys. In a team volley, all K agents on the active team throw one stone in sequence before turn passes to the opposing team. ## Formal rules R1. **Initial Setup:** Match target score Pointstarget=3. Match scores PointsA=0,PointsB=0. Starting team chosen uniformly at random [gap-fill]. R2. **Round Setup:** Both teams start with 3 standing slabs: SlabsA={1,2,3}, SlabsB={1,2,3}. R3. **Volley Execution:** The active team executes K sequential throws (one per agent j∈{1,…,K}). - Agent j selects a target slab index t∈Slabsopp. - Hit resolution: [gap-fill] Probability of direct knockdown is: Ptopple(t)=max⁡(0.15,0.40+0.05×Skill(j)) - If hit succeeds (ξ≤Ptopple), slab t is knocked down and removed from Slabsopp (Slabsopp←Slabsopp∖{t}). - [gap-fill] If t is knocked down, adjacent standing slabs have a 10% secondary ricochet topple probability. R4. **Mid-Volley Round End:** If at any point during a volley |Slabsopp|==0, the round ends immediately. The active team earns +1 point (Pointsactive←Pointsactive+1). R5. **Turn Pass:** If |Slabsopp|>0 after all K agents have thrown, the active team role toggles to the opposing team. R6. **Match Termination:** The first team to reach Points≥Pointstarget wins the match. ## Edge cases - **Tie at the end condition:** Not applicable; round ends immediately when one team's slabs are reduced to zero, making simultaneous victories impossible. - **A player with no legal move:** Not applicable; as long as the round is active, |Slabsopp|≥1, so a legal target slab always exists. - **Repetition / cycles:** Slabs are never restored during a round; each round strictly terminates in a finite number of volleys because P(topple)≥0.15>0. - **Simultaneous claims on the same resource or position:** Not applicable; throws are executed in strict sequential order. - **Player elimination mid-round / turn order repair:** If an agent disconnects or times out, their throw is scored as an automatic miss, and the turn advances to the next team member. - **Resource exhaustion:** Slab sets reset to {1,2,3} at the beginning of each new round. - **Termination guarantee:** Because P(topple)≥0.15>0 on every throw, the probability of a round lasting more than V volleys decays exponentially as (1−0.15)K⋅V→0. The game provably terminates with probability 1. ## State representation - **Persistent state:** - `points_A`, `points_B`: Integers ∈[0,3] - `active_team`: Enum {Team A,Team B} - **Per-round ephemeral state:** - `slabs_A`: Bitset / set of remaining slab IDs ⊆{1,2,3} - `slabs_B`: Bitset / set of remaining slab IDs ⊆{1,2,3} - `active_thrower_idx`: Integer ∈[1,K] - **Observable information:** Fully observable to all agents. ## Action space - `aim_and_throw(target_slab_id)`: - `target_slab_id`: Integer ∈Slabsopp. Precondition: `target_slab_id` must be an active standing slab. ## Playability notes - **Single-player-vs-AI:** Natural. A human player controls one team against an AI opponent adjusting aim choices based on remaining slab density. - **Group games (n>4):** Reduces cleanly to 2 decision agents (or 1v1 play with 1 throw per agent per volley) without modifying the mathematical game tree. - **Physical-skill games:** Modeled through the accuracy parameter Skill(j) and target prioritization. - **What is lost in the adaptation:** The physical tactile feedback of hefting a balanced flat stone and watching it skip or shatter on rocky terrain. ## Implementation hints - Very simple, fast state transitions; perfect for quick Monte Carlo evaluations or reinforcement learning baselines. - First milestone: Implement single-round 1v1 loop with fixed P(topple)=0.5 and verify round termination.

Download as JSON

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

Download JSON
Cite this page

Slab Stone Target Throwing (Kosovo). (n.d.). Games of Countries. https://gamesofcountries.com/games/xk-kosovo/loja-e-rrasave

@misc{gamesofcountries-xk-kosovo-loja-e-rrasave,
  title = {Slab Stone Target Throwing (Kosovo)},
  year = {n.d.},
  publisher = {Games of Countries},
  url = {https://gamesofcountries.com/games/xk-kosovo/loja-e-rrasave}
}