Computer Vision for MOBA Analytics: A Dataset and Baseline for Visibility Analysis in Dota 2

2026-06-25

AI Summary

This study addresses the limitation of existing MOBA game analyses, which predominantly rely on structured data and fail to capture the actual in-game visibility available to teams during matches. To bridge this gap, the authors introduce Dota2-Vis, a novel video dataset, and propose the first visibility analysis framework based on dual-perspective gameplay footage and manually annotated minimap images. Leveraging the YOLOv11 model family, they process 288 full-HD match videos and 2,477 minimap images to infer the presence states of opposing players. Experimental results demonstrate that YOLOv11l achieves superior performance in handling dense and cluttered minimap scenes, generating highly reliable visibility curves. These curves effectively reveal behavioral patterns at the levels of individual players, heroes, roles, and entire teams, thereby complementing and extending traditional structured-data approaches.

Abstract

Introduction: Most Multiplayer Online Battle Arena (MOBA) analytics studies rely on structured data, which does not directly capture what each team could actually see during a match. Objective: This work introduces Dota2-Vis, a video-based dataset, and a baseline pipeline for visibility analysis in professional Dota 2 matches. Methodology: The dataset comprises all 144 matches from The International 2025, recorded from both team perspectives, totaling 288 Full HD videos, together with 2,477 manually annotated minimap images. We evaluate multiple variants of a modern object detector for player-icon detection and use the best-performing model to estimate opponent-visible player presence over time. Results: YOLO11l (large) achieved the best overall performance, reliably identifying player icons even in dense and visually cluttered minimap scenes. The resulting visibility curves reveal player, hero, role, and team-level patterns that complement conventional MOBA analytics, highlighting behavioral differences that are difficult to obtain from structured data alone. The dataset and code are publicly available at https://github.com/RicardoRCarvalho/dota2-vis/.

Problem

• MOBA analytics
• visibility analysis
• Dota 2
• computer vision
• structured data limitation

Innovation

• visibility analysis
• MOBA analytics
• video-based dataset
• object detection
• Dota 2