Modeling Individual and Team Behavior through Spatio-temporal Analysis
summary
A new methodology called Interactive Behavior Analytics (IBA) combines a custom labeler StratMapper and a visualization tool Glyph with Dynamic Time Warping and clustering to produce human-interpretable models of player and team behavior. Researchers label replays (24 initial labels reduced to 9 for paper), two professional DotA 2 games required about 10 hours of labeling, achieved Cohen’s kappa 0.72 (78% agreement), and Glyph exposes sequence differences (one-sided vs. back-and-forth games), support-player similarity, and transitions like gank→kill and push states.
AI-written summary · about
chapters (4)
related talks
08mVisualizing the Spatio-Temporal Evolution of Gameplay using Storyline Visualization: A Study with...
18mHow the Experts Do It: Assessing and Explaining Agent Behaviors in Real-Time Strategy Games
32mIdentifying playstyles in football through spatial networks
04mInteractive Sports Analytics
06mUnderstanding Player Patterns by Combining Knowledge-Based Data Abstraction with Interactive ...
35mDecoding bias and narrative in competitive video games
53mBalancing gameplay of aRPG using player behavioral patterns
17mWhat Moves Players? Visual Data Exploration of Twitter and Gameplay Data