Analyzing Actions in Play-by-Forum RPG

摘要

Interactive Storytelling (IS) technologies are enabling richer experiences for electronic games. Current computational IS models are based on investigations about how games are planned by game designers and actually played by the audience. Unfortunately, most research efforts are limited to the structured data for obtaining insights about IS models. This paper presents a study aimed at determining which actions modeled by Role-Playing Games (RPG) are more important for actual gameplay and how these actions are related. For doing so, we first extracted textual data from existing adventures found in a play-by-forum RPG portal. Such gameplay data is written in Portuguese and reflect natural game-play without observer intervention. Our analyses from a natural language processing perspective provide valuable insights for IS models in reducing the gameplay chasm between electronic and tabletop RPG.