GolfCardGame

Description

GolfCardGame is an ORS environment for evaluating agents on strategic decision-making in the Golf card game, where the goal is to achieve the lowest score. This environment wraps the Golf implementation from TextArena, a framework for text-based game environments.

Capabilities

• Testing probabilistic reasoning under partial information
• Evaluating risk assessment in card swapping decisions
• Assessing memory and pattern recognition for opponent behavior
• Testing strategic timing of when to draw, swap, or discard

Compute Requirements

Golf does not require a sandbox. It has minimal compute requirements.

License

MIT.

Tasks

There are two splits: train (300 tasks) and test (300 tasks). Each split contains 50 tasks across each of 6 variants:

• Golf-v0
• Golf-v0-train
• Golf-v0-raw
• Golf-v0-medium
• Golf-v0-medium-train
• Golf-v0-medium-raw

Each task is seeded for reproducibility.

Reward Structure

This is a sparse reward environment. Rewards are mapped from TextArena's native range of {-1, 0, 1} to {0.0, 0.5, 1.0} via (raw + 1) / 2.

We do not use LLM graders for this environment; reward is determined programmatically.

Data

Game state is generated procedurally by the TextArena engine using seeded randomness. No external data files are required.

Tools

Agents are given four tools:

• draw_from_deck(): Draw a card from the deck.
• take_from_discard(): Take the top card from the discard pile.
• swap_card(row, column): Swap the drawn card with a card in your hand at the given row and column (1-indexed).
• discard_card(): Discard the drawn card without swapping.

Time Horizon

Golf is a multi-turn environment.

Environment Difficulty

Medium. Golf requires understanding card values, probability, and strategic timing. Players must decide when to draw, what to swap, and when to end the round. Matching column pairs add tactical depth.

Other Environment Requirements

This environment requires an OpenAI API key (passed via secrets) to power the LLM opponent.

Safety

Agents in Golf interact only with a card game and have no access to external systems, the internet, or sensitive data. The environment does not present safety risks.

Citations

@software{textarena2024,
  author    = {Guertler, Leon and Banting, Wilfried and Pignatelli, Eduardo},
  title     = {TextArena},
  year      = {2024},
  publisher = {GitHub},
  url       = {https://github.com/LeonGuertler/TextArena}
}