Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

STORY2GAME: Generating (Almost) Everything in an Interactive Fiction Game

Eric Zhou, Shreyas Basavatia, Moontashir Siam, Zexin Chen, Mark O. Riedl

TL;DR

STORY2GAME tackles the challenge of turning unconstrained LLM-generated stories into playable text-based games by coupling a story generator with world construction and engine synthesis. The approach uses a three-stage pipeline—story generation with preconditions/effects, world generation, and executable action code generation—plus a dynamic action module for just-in-time player actions. Evaluations show high per-action compilation rates and substantial grounding of dynamic actions, though semantic alignment can be imperfect for complex actions. The work demonstrates a scalable path toward largely self-generated game worlds that preserve narrative creativity while maintaining a grounded state for coherent gameplay.

Abstract

We introduce STORY2GAME, a novel approach to using Large Language Models to generate text-based interactive fiction games that starts by generating a story, populates the world, and builds the code for actions in a game engine that enables the story to play out interactively. Whereas a given set of hard-coded actions can artificially constrain story generation, the ability to generate actions means the story generation process can be more open-ended but still allow for experiences that are grounded in a game state. The key to successful action generation is to use LLM-generated preconditions and effects of actions in the stories as guides for what aspects of the game state must be tracked and changed by the game engine when a player performs an action. We also introduce a technique for dynamically generating new actions to accommodate the player's desire to perform actions that they think of that are not part of the story. Dynamic action generation may require on-the-fly updates to the game engine's state representation and revision of previously generated actions. We evaluate the success rate of action code generation with respect to whether a player can interactively play through the entire generated story.

STORY2GAME: Generating (Almost) Everything in an Interactive Fiction Game

TL;DR

STORY2GAME tackles the challenge of turning unconstrained LLM-generated stories into playable text-based games by coupling a story generator with world construction and engine synthesis. The approach uses a three-stage pipeline—story generation with preconditions/effects, world generation, and executable action code generation—plus a dynamic action module for just-in-time player actions. Evaluations show high per-action compilation rates and substantial grounding of dynamic actions, though semantic alignment can be imperfect for complex actions. The work demonstrates a scalable path toward largely self-generated game worlds that preserve narrative creativity while maintaining a grounded state for coherent gameplay.

Abstract

We introduce STORY2GAME, a novel approach to using Large Language Models to generate text-based interactive fiction games that starts by generating a story, populates the world, and builds the code for actions in a game engine that enables the story to play out interactively. Whereas a given set of hard-coded actions can artificially constrain story generation, the ability to generate actions means the story generation process can be more open-ended but still allow for experiences that are grounded in a game state. The key to successful action generation is to use LLM-generated preconditions and effects of actions in the stories as guides for what aspects of the game state must be tracked and changed by the game engine when a player performs an action. We also introduce a technique for dynamically generating new actions to accommodate the player's desire to perform actions that they think of that are not part of the story. Dynamic action generation may require on-the-fly updates to the game engine's state representation and revision of previously generated actions. We evaluate the success rate of action code generation with respect to whether a player can interactively play through the entire generated story.
Paper Structure (18 sections, 6 figures, 3 tables)

This paper contains 18 sections, 6 figures, 3 tables.

Table of Contents

  1. Introduction
  2. Related Work
  3. Story2Game
  4. Generating Stories with Preconditions and Effects
  5. World Generation
  6. Game Engine Generation
  7. Dynamic Player Action Generation
  8. Essential Object Preconditions
  9. Fundamental and Additional Preconditions
  10. Preceding Event Preconditions
  11. Attribute Effects
  12. Evaluations and Results
  13. Game World Initialization Experiments
  14. Dynamic Action Generation
  15. Conclusions
  16. ...and 3 more sections

Figures (6)

  • Figure 1: Game Generation Pipeline.
  • Figure 2: Dynamic Action Generation Pipeline.
  • Figure 3: Average node counts of different types based on story length
  • Figure 4: Percentage of times each category was chosen as a precondition. 30 total items and characters considered, each applied with 3 novel verbs.
  • Figure 5: Novel Action Generation Confirmation Percentiles
  • ...and 1 more figures