Skip to main content
Formal Specification

Carving Land (literally "cutting land"; traditional territory knife-throwing game) — Formal Specification

Norway · dexterity-throw · Confidence: low

← Back to game page

Implementation Suitability

Suitability: adaptation-requiredFidelity: probabilistic-proxyGap-fills: 7
# Kappe land — AI-optimized specification ## General idea Kappe land ("Carving Land") is a traditional outdoor territory-conquest throwing game. Players divide a circular field of damp earth into equal starting sectors and take turns flipping a pocket knife (or wooden peg) into opponents' lands. A successful stick allows the thrower to slice a line across the target territory and annex one of the sub-polygons, progressively shrinking opponents until they can no longer physically stand within their borders. ## Formal model - **Abstraction taken:** The continuous circular sandbox is represented as planar 2D polygon partitions of a unit circle C={(x,y)∈ℝ2|x2+y2≤1}. Knife throws are modeled as targeted coordinate samples (xt,yt) with blade orientation angle θt, resolved via a stochastic stick probability function. Territory slicing is computed via 2D geometric polygon clipping. The standing constraint is discretized into a minimum enclosed polygon area threshold Amin. - **Player/agent count:** 2 to 4 agents (standard model: 2-agent duel or 3-player circle). - **Turn model:** Strictly sequential, cyclic round-robin order. ## Formal rules R1. **Initial Setup:** The unit circle area Atotal=π is partitioned into N equal pie-shaped sectors P1,…,PN, one per player. R2. **Turn Order:** Players alternate taking turns in fixed cyclic order 1…N. R3. **Throw Declaration:** The active player i declares a target coordinate (x,y)∈Pj (where j≠i) and a cut angle θ∈[0,π) [gap-fill]. R4. **Stick Resolution:** The throw is successful with probability P(stick)=σ(αskill⋅skilli−βdist⋅d(Pi,(x,y))) [gap-fill]. If the throw fails, no territory changes and the turn passes. R5. **Territory Slicing & Annexation:** - If the throw succeeds, a line L(x,y,θ) is cast across polygon Pj, splitting Pj into two disjoint sub-polygons Pj,1 and Pj,2. - Player i chooses which sub-polygon Pj,k to annex [gap-fill: player i must choose the sub-polygon adjacent to their own territory to maintain contiguous land]. - Pi←Pi∪Pj,k, and Pj←Pj∖Pj,k. R6. **Elimination Threshold:** If after any turn the total area of an opponent's polygon falls below Amin=0.05⋅Atotal [gap-fill: modeling the minimum physical space required to balance on one foot], player j is eliminated (`OUT`). - Upon elimination of player j, player j's remaining territory is absorbed by the conqueror (player i) [gap-fill]. R7. **Victory Condition:** The last remaining player holding >(1.0−Amin)⋅Atotal wins the game immediately. ## Edge cases - **Tie at the end condition:** Not applicable; eliminations are strictly sequential, leaving exactly 1 surviving conqueror. - **A player with no legal move:** An eliminated player is skipped in turn rotation; active players always have a valid target sector while opponents exist. - **Repetition / cycles:** Territory annexation strictly reduces the target player's area (A(Pj) strictly decreases on successful cuts). To guard against endless missed throws by low-skill bots, [gap-fill] a maximum turn cap Tmax=100 plies is enforced; the player with the largest area wins at timeout. - **Simultaneous claims on the same resource or position:** Not applicable; turns and slicing operations are strictly sequential. - **Player elimination mid-round; turn order repair:** When player j is eliminated, their index is removed from the active turn queue; play continues seamlessly with (i+1)(mod|Vactive|). - **Resource exhaustion:** Total circular area π is invariant and conserved across all partitions. - **Termination guarantee:** Guaranteed by monotonic territorial shrinkage combined with the turn cap Tmax=100 [gap-fill]. ## State representation - **Persistent state:** Polygon list 𝒫=[P1,…,PN] (where each Pi is a set of 2D vertex loops), `active_players` (list of alive agent IDs), `turn_index` ∈[0,|Vactive|−1]. - **Per-turn state:** `declared_target` (x,y), `declared_angle` θ, `throw_result` ∈{STICK,FAIL}. - **Observable information:** Perfect public information. All territory boundaries and agent areas are fully observable. ## Action space - `aim_and_throw(target_player_id, target_x, target_y, angle)`: Continuous/discretized targeting parameters for knife throw. - `select_annexation_parcel(parcel_index)`: Binary choice of which sliced sub-polygon to absorb. ## Playability notes - **Single-player-vs-AI:** Excellent strategic geometry game. Player optimizes cut angles to maximize annexed area while severing opponent expansion vectors. - **Group games (n>4):** Best with 2 to 4 players. For N>4, polygon slicing geometry becomes overly fragmented; 2-agent duel is the cleanest reduction. - **Physical-skill games:** Knife-spin mechanics and ground-stick penetration are abstracted via logistic skill and distance functions. - **What is lost in the adaptation:** Physical danger of open blades, sensory damp soil texture, and the physical gymnastics of balancing on tiptoes. ## Implementation hints - Use standard 2D computational geometry libraries (e.g. Clipper / Shapely polygon clipping). - Ensure polygon intersection routines handle collinear edge clipping robustly. - Suggested first milestone: Implement 2-player sector slicing with deterministic knife stick coordinates, verifying area conservation (∑Area(Pi)=π).

Download as JSON

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

Download JSON
Cite this page

Carving Land (literally "cutting land"; traditional territory knife-throwing game) (Norway). (n.d.). Games of Countries. https://gamesofcountries.com/games/no-norway/kappe-land

@misc{gamesofcountries-no-norway-kappe-land,
  title = {Carving Land (literally "cutting land"; traditional territory knife-throwing game) (Norway)},
  year = {n.d.},
  publisher = {Games of Countries},
  url = {https://gamesofcountries.com/games/no-norway/kappe-land}
}