Glyph: Visualization Tool for Understanding Problem Solving Strategies in Puzzle Games

TL;DR

This paper introduces Glyph, a visualization tool designed to compare and analyze individual and collective player strategies in puzzle games, aiding researchers and designers in understanding behavior patterns and identifying anomalies.

Contribution

The paper presents a novel visualization technique that enables detailed comparison of player strategies while reducing visual cognitive load, addressing limitations of existing methods.

Findings

Core features helped users understand player strategies effectively

Prototype system showed potential despite usability improvements needed

Initial testing indicated usefulness for academic and industry analysis

Abstract

Understanding player strategies is a key question when analyzing player behavior both for academic researchers and industry practitioners. For game designers and game user researchers, it is important to gauge the distance between intended strategies and emergent strategies; this comparison allows identification of glitches or undesirable behaviors. For academic researchers using games for serious purposes such as education, the strategies adopted by players are indicative of their cognitive progress in relation to serious goals, such as learning process. Current techniques and systems created to address these needs present a few drawbacks. Qualitative methods are difficult to scale upwards to include large number of players and are prone to subjective biases. Other approaches such as visualization and analytical tools are either designed to provide an aggregated overview of the data, losing the nuances of individual player behaviors, or, in the attempt of accounting for individual behavior, are not specifically designed to reduce the visual cognitive load. In this work, we propose a novel visualization technique that specifically addresses the tasks of comparing behavior sequences in order to capture an overview of the strategies enacted by players and at the same time examine individual player behaviors to identify differences and outliers. This approach allows users to form hypotheses about player strategies and verify them. We demonstrate the effectiveness of the technique through a case study: utilizing a prototype system to investigate data collected from a commercial educational puzzle game. While the prototype's usability can be improved, initial testing results show that core features of the system proved useful to our potential users for understanding player strategies.