An Analysis of Target Tile Pathfinding in Classic Pac-Man Ghost AI

Background: This study analyzes the Target Tile pathfinding algorithm used by ghost characters in Pac-Man (1980), where simple mathematical rules create distinct behaviors. Methods: Using a custom probabilistic simulation, we evaluated the efficiency of ghost targeting systems. We measured capture efficiency (time-to-kill) and compared the Euclidean and Manhattan distance heuristics. Results: The simulation showed that the standard mixed-strategy configuration (combining pursuit, ambush, flanking, and retreat) achieved an average capture time of 155 ticks, significantly outperforming a uniform "all-chaser" configuration (240 ticks). The Euclidean distance provided a slightly higher path efficiency. Conclusion: These findings confirm that interacting simple deterministic rules produce complex emergent behavior, making the original algorithm an effective model for teaching pathfinding and designing lightweight game AI.