Formal Specification
Pitching to the Peg (traditional Norwegian coin-pitching game; also known as kaste på stikka or krone på stikke) — Formal Specification
Norway · gambling-stake · Confidence: low
← Back to game pageImplementation Suitability
Suitability: engine-readyFidelity: probabilistic-proxyGap-fills: 8
# Kaste på stikke — AI-optimized specification
## General idea
Kaste på stikke ("Pitching to the Peg") is a traditional precision coin-pitching and wagering game. Players ante coins into a common round pool, then pitch individual coins toward a fixed wooden target peg. Proximity to the peg determines pitching rank, which grants priority rights in a subsequent coin-flipping (*romling*) stage to claim coins from the pool.
## Formal model
- **Abstraction taken:** The continuous trajectory and ground roll of the coin are abstracted into a 2D landing coordinate sampled from a bivariate normal distribution centered at target parameterized by throwing distance and agent pitching skill. The *romling* settlement is modeled as sequential Bernoulli trials ( per coin).
- **Player/agent count:** 2 to 6 agents (standard model: 4 agents).
- **Turn model:** Two-stage round structure: sequential pitching phase, followed by ranked settlement flips.
## Formal rules
R1. **Initial Setup:** Each player starts with an initial coin purse coins [gap-fill]. The target peg is located at at distance meters from the throw line.
R2. **Ante Phase:** At the start of each round, every solvent player () contributes 1 coin to the pool .
R3. **Pitching Phase:** Each active player pitches one coin. The landing position is sampled from , where [gap-fill].
- Distance to peg is computed: .
- If ( meters [gap-fill: direct physical touch]), player scores a `DIRECT_HIT` and claims the entire pool immediately, bypassing R4.
R4. **Settlement Phase (Romling):**
- Players are sorted by distance in ascending order: rank order where .
- Player tosses all coins. For each coin, outcome is `HEADS` () or `TAILS` (). Player collects all `HEADS` coins into their purse .
- If any `TAILS` coins remain, player tosses the remaining coins, collecting all `HEADS`. This repeats down the rank order until all coins in are claimed.
- If coins remain after the lowest-ranked player, the cycle wraps back to [gap-fill].
R5. **Elimination:** Any player whose purse drops to is eliminated (`BANKRUPT`).
R6. **Victory Condition:** The game terminates when either:
- Exactly 1 player holds all coins (total monopoly).
- [gap-fill] A maximum limit of rounds is reached; the player with the largest purse wins.
## Edge cases
- **Tie at the end condition:** If two players have identical purse values at match timeout, the player with the lowest average throw distance across all rounds wins [gap-fill].
- **A player with no legal move:** Bankrupt players are skipped; solvent players must pitch on their turn.
- **Repetition / cycles:** Bankroll shifts on every round where coins are won. The global round cap prevents infinite zero-sum cycling [gap-fill].
- **Simultaneous claims on the same resource or position:** If two coins have identical distance , tie-breaker is determined by a single coin flip [gap-fill].
- **Player elimination mid-round; turn order repair:** Bankrupt players are removed from the active queue at the end of the round; the next round proceeds with remaining solvent players.
- **Resource exhaustion:** Total coins in the system is strictly conserved; the pool is strictly emptied every round.
- **Termination guarantee:** Guaranteed by finite coin purse conservation, absorption state of gambler's ruin, and the hard cap of rounds.
## State representation
- **Persistent state:** `bankrolls` (array of integers ), `round_number` , `active_players` (list of alive agent IDs).
- **Per-round state:** `pool_size` , `distances` (array of floats ), `rank_order` (sorted list of agent IDs), `unclaimed_coins` .
- **Observable information:** Perfect public information. All purse values, throw distances, and coin flip outcomes are fully visible.
## Action space
- `pitch_coin(aim_offset_x, aim_offset_y)`: Pitch action specifying aim point relative to peg.
- `shake_and_toss_pool()`: Action triggering the Bernoulli flip sequence during Romling.
## Playability notes
- **Single-player-vs-AI:** Highly suitable as a probabilistic bankroll management game against AI throwers with varying skill distributions.
- **Group games (n>4):** Scales naturally up to 6 players. 2-agent duel is a clean, classic heads-up gambler's ruin duel.
- **Physical-skill games:** Underhand spin, trajectory friction, and ground-bounce dynamics are abstracted into 2D Gaussian landing kernels.
- **What is lost in the adaptation:** Tactile clinking of copper coins in cupped hands, street-corner tension, and physical soil texture readings.
## Implementation hints
- Very fast, lightweight Monte Carlo simulation model.
- State is an integer array of balances plus a 2D coordinate sampler.
- Suggested first milestone: Implement the 2-player Gaussian toss distance and sequential Bernoulli coin-pool collection loop.
Download as JSON
Download the full game corpus entry for this game as structured JSON.
Download JSONCite this page
Pitching to the Peg (traditional Norwegian coin-pitching game; also known as kaste på stikka or krone på stikke) (Norway). (n.d.). Games of Countries. https://gamesofcountries.com/games/no-norway/kaste-pa-stikke
@misc{gamesofcountries-no-norway-kaste-pa-stikke,
title = {Pitching to the Peg (traditional Norwegian coin-pitching game; also known as kaste på stikka or krone på stikke) (Norway)},
year = {n.d.},
publisher = {Games of Countries},
url = {https://gamesofcountries.com/games/no-norway/kaste-pa-stikke}
}