Formal Specification
Table shuffleboard — Formal Specification
Belgium · dexterity-throw · Confidence: high
← Back to game pageImplementation Suitability
Suitability: engine-readyFidelity: probabilistic-proxyGap-fills: 3
# Sjoelen — AI-optimized specification
## General idea
A turn-based tabletop sliding game where each player attempts to score 30 wooden discs into four compartments (valued 2, 3, 4, and 1) across three successive sub-rounds. A set-collection multiplier awards 20 points (double the face sum of 10) for every quartet of discs containing one in each compartment. The core strategic challenge is real-time risk allocation: balancing attempts across under-filled compartments to complete 20-point quartets versus securing direct points in the high-value 4-slot.
## Formal model
- **Abstraction taken:** Physical sliding dynamics (hand velocity, trajectory angle, puck-to-puck rebounds, arch clearance) are abstracted into an aim-choice model where the agent selects a target arch . A stochastic resolver evaluates success (entering target arch, deflecting into adjacent arch, bouncing back, or blocking the entry), parameterized by agent accuracy.
- **Player/agent count:** 2 to individual players or 2 teams.
- **Turn model:** Sequential turns; each player completes a full 3-subround board before play passes to the next agent.
## Formal rules
R1. **Setup:** A board has 4 compartments with face values . Counts initialize to for all . The player begins with a pool of active discs.
R2. **Sub-round 1 (Beurt 1):** The player slides all 30 discs sequentially. Each disc is aimed at target arch .
- Outcome: disc enters compartment (), enters adjacent compartment , or fails to enter any arch (remains un-scored).
R3. **Sub-round 2 (Beurt 2):** Un-scored discs are collected and slid sequentially toward target arches. Scored discs are added to .
R4. **Sub-round 3 (Beurt 3):** Remaining un-scored discs are collected and slid in the final attempt.
R5. **Scoring calculation:** At the conclusion of Sub-round 3 (or when all 30 discs are scored):
- Compute full quartets: .
- Score for quartets: .
- Score for remainders: .
- Total board score: .
R6. **Match resolution:** Across boards, the agent with the highest aggregate score wins.
R7. [gap-fill] In team play, the match score is the sum of all individual board scores.
R8. [gap-fill] If a tie exists after boards, a 1-board sudden-death playoff is played between tied players until a winner emerges.
## Edge cases
- **Tie at the end condition:** Handled by R8; 1-board tiebreaker repeated until scores differ.
- **A player with no legal move:** Not applicable; sliding active pucks on one's turn is always legal and mandatory as long as active discs remain.
- **Repetition / cycles:** Bounded; each board strictly terminates after at most total disc slides.
- **Simultaneous claims on the same resource or position:** Not applicable; players take complete turns sequentially.
- **Player elimination mid-round; turn order repair:** [gap-fill] If a player forfeits mid-game, their score for unplayed boards is 0; remaining players complete scheduled boards in rotation.
- **Resource exhaustion (pieces, tokens, space, time):** The 30 discs are physically conserved and reset to pool at the start of each new board.
- **Termination guarantee:** Each sub-round consumes a strictly positive number of remaining discs; each board terminates in at most 3 sub-rounds, guaranteeing finite match completion.
## State representation
- **Persistent state:**
- `board_scores`: array of size .
- `current_player_idx`: active player index.
- `current_board_idx`: active board index ().
- **Per-turn ephemeral state:**
- `compartments`: vector .
- `active_discs`: integer count of pucks available in current sub-round.
- `subround`: current sub-round ().
- **Observable information:** Fully observable public state; all compartment counts and puck positions are visible to all agents.
## Action space
- `slide_disc(target_arch: int in {1, 2, 3, 4}, speed: float, angle_offset: float)`:
- Precondition: `active_discs > 0`, `current_player_idx` active.
- Effect: Resolves stochastic outcome, updates compartment tally or un-scored pool, decrements `active_discs`.
## Playability notes
- **Single-player-vs-AI:** Natural; human and AI compete on identical boards, optimizing target slot selection based on current compartment occupancy.
- **Group games ():** Seamlessly supported through sequential individual play.
- **Physical-skill games:** Outcome probability modeled as:
- **What is lost in adaptation:** The tactile auditory satisfaction of wooden puck collisions and nuanced ricochet shots off the board walls.
## Implementation hints
- **Complexity:** Very low state complexity; optimal play policy can be solved via dynamic programming over state vector .
- **Branching factor:** 4 discrete choices per disc throw.
- **Verification test:** Verify that state calculates to points exactly.
Download as JSON
Download the full game corpus entry for this game as structured JSON.
Download JSONCite this page
Table shuffleboard (Belgium). (n.d.). Games of Countries. https://gamesofcountries.com/games/be-belgium/sjoelen
@misc{gamesofcountries-be-belgium-sjoelen,
title = {Table shuffleboard (Belgium)},
year = {n.d.},
publisher = {Games of Countries},
url = {https://gamesofcountries.com/games/be-belgium/sjoelen}
}