Skip to main content
Formal Specification

Court Hopscotch / Puck Kicking — Formal Specification

Pakistan · dexterity-small · Confidence: low

← Back to game page

Implementation Suitability

Suitability: engine-readyFidelity: probabilistic-proxyGap-fills: 1
# Stapoo / Kidi Kada — AI-optimized specification ## General idea A turn-based 2D grid-navigation and territorial-conquest game. Players toss a puck into sequential court chambers, complete a hopping traverse along a directed graph of single and double squares, and retrieve the puck without stepping on boundary lines. Conquering a square converts it into a private resting node that blocks opponents from entering. ## Formal model - **Abstraction taken:** The 8-square court is modeled as a topological directed graph `G = (V, E)` with node coordinates `V = {1..8, Samandar}`. Hopping balance and toss placement are resolved via continuous coordinate sampling relative to cell bounding boxes. - **Player/agent count:** 2 to 4 players (standard `N = 2`). - **Turn model:** Sequential turn rotation; an active player continues taking turns until a foul occurs. ## Formal rules R1. Setup: Court nodes `1..8` have bounding boxes `B_1..B_8`. Ownership map `owner[1..8] = None`. Each player `i` starts with `target_cell[i] = 1`, `homes_won[i] = 0`. R2. Marker Toss: - On active turn, player `i` tosses marker aiming at cell `c = target_cell[i]`. - Landing position is sampled: `(x_puck, y_puck) = center(B_c) + Gaussian(mean=0, std=sigma_toss(skill))`. - If `(x_puck, y_puck)` lies outside `B_c` or on boundary lines: Foul -> Turn ends; `target_cell[i]` remains unchanged. R3. Traversal Phase: - If toss is valid, player hops sequentially through `V \ {c}`: - For each single cell `v`: execute 1-foot balance check `P(pass) = skill_balance[i]`. - For double cells (4-5, 7-8): execute 2-foot straddle landing. - If cell `v` is owned by opponent (`owner[v] != None and owner[v] != i`): Player must hop over `v` without landing in it (higher balance check). - If any balance check fails or player steps on a line: Foul -> Turn ends immediately. R4. Marker Retrieval & Circuit Completion: - At resting node `Samandar`, player turns and traverses backward. - At cell adjacent to `c`, player executes `retrieve_marker()`, balances on one foot, and hops over baseline. - If successful without fouls: `target_cell[i] += 1`. Player retains turn to attempt the next square. R5. Home Conquering (*Ghar*): - When `target_cell[i] > 8`: Player attempts blind backward toss. - Target cell `c_home` is sampled uniformly across unowned cells. - If valid: `owner[c_home] = i`, `homes_won[i] += 1`, and `target_cell[i] = 1` for a new cycle. R6. Win condition: The game terminates when all 8 cells are owned. The player with the highest `homes_won` total wins. R7. [gap-fill] Max turns: If ownership is unresolved after 60 total turns, the player with the most owned cells (or higher target cell progress) wins. ## Edge cases - **Tie at the end condition:** Handled by target-cell progress tie-breaker in R7. - **A player with no legal move:** Marker toss is always legal. - **Repetition / cycles:** Bounded because successful traverses monotonically increase cell ownership until all 8 cells are claimed. - **Simultaneous claims on the same resource or position:** Not applicable; turns are strictly sequential. - **Player elimination mid-round:** Not applicable; players take turns until the court is fully conquered. - **Resource exhaustion:** 8 court cells are permanently claimed one by one. - **Termination guarantee:** Guaranteed by finite 8-cell capacity and max turn cap in R7. ## State representation - **Persistent state:** `owner[1..8] in {None, 1..N}`, `target_cell[1..N] in {1..9}`, `homes_won[1..N]`, `active_player_idx in {1..N}`, `turn_counter in {1..60}`. - **Per-traverse ephemeral state:** Marker position `(x_puck, y_puck)`, current hop node `v_current`, balance check results. - **Observable information:** Perfect information — court ownership, player progress, and marker landing positions are fully observable. ## Action space - `toss_marker(target_cell)`: Parameterized aim action. - `execute_traverse()`: Execute hopping sequence across graph nodes. - `blind_home_toss()`: Blind toss for home conquest. ## Playability notes - **Single-player-vs-AI:** Natural; human player balances risk and accuracy while racing AI traversal progress. - **Group games (n>4):** Sequential turn rotation extends cleanly up to 4 players. - **Physical-skill games:** Single-leg balance and blind throw accuracy are parameterized via probability distributions. - **What is lost:** The physical hopping on warm concrete and the playful argument over whether a stone barely touched the chalk line. ## Implementation hints Low complexity. The 8-node graph traversal with skip-nodes is a simple state-machine loop. First implement the node graph and owner blocking constraints before fine-tuning the toss accuracy parameters.

Download as JSON

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

Download JSON
Cite this page

Court Hopscotch / Puck Kicking (Pakistan). (n.d.). Games of Countries. https://gamesofcountries.com/games/pk-pakistan/stapoo

@misc{gamesofcountries-pk-pakistan-stapoo,
  title = {Court Hopscotch / Puck Kicking (Pakistan)},
  year = {n.d.},
  publisher = {Games of Countries},
  url = {https://gamesofcountries.com/games/pk-pakistan/stapoo}
}