Formal Specification
Traditional Prize Board Game — Formal Specification
Singapore · gambling-stake · Confidence: high
← Back to game pageImplementation Suitability
Suitability: engine-readyFidelity: directGap-fills: 1
# Tikam-tikam — AI-optimized specification
## General idea
Tikam-tikam is a discrete, stochastic lottery draw without replacement from a finite grid of sealed tabs. An agent manages a finite budget of tokens, deciding when to draw and when to stop based on Bayesian updates of remaining high-value prize probabilities as the board is progressively depleted.
## Formal model
- **Abstraction taken:** None; formal model of multi-armed bandit / lottery draw without replacement from an urn of size .
- **Player/agent count:** 1 to agents (or 1 agent vs. the Environment).
- **Turn model:** Sequential single-draw decisions.
## Formal rules
R1. Setup: A deck of tabs is generated with fixed prize distribution:
- Grand Prize ( points): count .
- Minor Prize ( points): count .
- Consolation Prize ( points): count .
- Blank ( points): count .
- Entry cost per pull token.
R2. Agent State: Each agent starts with a bankroll tokens and cumulative prize score .
R3. Turn Action: On an agent's turn, the agent chooses to `DRAW` (if and ) or `STOP`.
R4. Draw Resolution:
- If `DRAW`: .
- A tab is sampled uniformly without replacement from remaining deck .
- , and tab is permanently removed from ().
- Public board state is updated with the revealed value of tab .
R5. Win Condition: When an agent chooses `STOP` or , their final payoff is computed as . In multi-agent play, the agent with the highest net payoff after the board is depleted wins. [gap-fill] A maximum of 100 total draws ends the board.
## Edge cases
- Tie at the end condition: If multiple agents finish with equal net payoff, the agent who drew the highest single prize value wins.
- A player with no legal move: If an agent runs out of tokens (), the agent is forced to `STOP`.
- Repetition / cycles: Bounded strictly by the finite tab pool , strictly decreasing with every draw.
- Simultaneous claims on the same resource or position: Turns are strictly sequential; each pull draws exactly one tab from the remaining pool.
- Player elimination mid-round: When an agent exhausts tokens, they stop playing while remaining agents continue.
- Resource exhaustion: Monotonically decreasing pool ().
- Termination guarantee: Provably terminates in at most 100 steps due to finite drawing without replacement.
## State representation
- Persistent state:
- `remaining_tabs`: multiset of values in .
- `total_remaining`: integer .
- `agent_bankroll`: array of integers .
- `agent_scores`: array of integers .
- `active_player`: integer.
- Observable information:
- Hidden: Exact locations of remaining prize values on the physical board.
- Public: Initial distribution, count of revealed prizes by tier, remaining tab count , and all agent bankrolls/scores.
## Action space
- `act()` .
## Playability notes
- Single-player-vs-AI: Perfect mathematical decision benchmark; optimal agent solves an optimal stopping problem via dynamic programming.
- Group games (n>4): Extends seamlessly to multiple competitive agents drawing sequentially from the same board.
- Physical-skill games: Not applicable.
- What is lost: Tactile peeling of perforated paper and sticky glue texture.
## Implementation hints
Extremely simple and lightweight state machine. Can be executed millions of times per second for Monte Carlo policy evaluation.
Download as JSON
Download the full game corpus entry for this game as structured JSON.
Download JSONCite this page
Traditional Prize Board Game (Singapore). (n.d.). Games of Countries. https://gamesofcountries.com/games/sg-singapore/tikam-tikam
@misc{gamesofcountries-sg-singapore-tikam-tikam,
title = {Traditional Prize Board Game (Singapore)},
year = {n.d.},
publisher = {Games of Countries},
url = {https://gamesofcountries.com/games/sg-singapore/tikam-tikam}
}