PigDice

Description

PigDice is an environment for evaluating agents on the dice game where players balance risk and reward to reach a target score. This environment wraps the PigDice implementation from TextArena, a framework for text-based game environments.

Capabilities

• Risk assessment and probability-based decision-making
• Strategic stopping under uncertainty
• Balancing short-term gains against potential losses
• Two-player competitive gameplay against an LLM opponent

Compute Requirements

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

License

MIT.

Tasks

There are two splits: train (150 tasks) and test (150 tasks). Each split contains 50 tasks across each of 3 variants:

• PigDice-v0-short
• PigDice-v0
• PigDice-v0-long

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 two tools:

• roll_die(): Roll the die to accumulate points. If you roll a 1, you lose your turn's points.
• hold_score(): Bank your turn's points and end your turn.

Time Horizon

PigDice is a multi-turn environment.

Environment Difficulty

Easy to Moderate. PigDice involves straightforward probability calculations and risk-reward tradeoffs, but optimal stopping strategies require balancing current position against opponent's score.

Other Environment Requirements

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

Safety

Agents in PigDice interact only with a dice 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}
}