SHADOW TAG: DESIGN AND IMPLEMENTATION OF A 2D SURVIVAL ARCADE GAME WITH A DELAYED MIMICRY AI

Abstract

Artificial Intelligence (AI) is increasingly applied within digital game development to build intelligent mechanics that boost player engagement, adaptability, and strategic thinking. Standard 2D arcade games typically rely on scripted enemy behavior and fixed movement routines, which makes gameplay predictable and limits replay value. This paper presents Shadow Tag, an AI-driven 2D survival arcade game built around a Delayed Mimicry AI, in which the antagonist replays the player's own past movement trajectory after a set time delay. Rather than relying on conventional pursuit logic such as pathfinding or direct target tracking, the proposed AI turns the player's earlier actions into an evolving obstacle, demanding ongoing strategic planning and spatial awareness.
The game is built with Python and the Pygame framework using an object-oriented design that combines player control, a delayed shadow AI, procedurally generated obstacles, collectible objects, collision detection, animation handling, particle effects, persistent score tracking, and a state-driven interface. The delayed mimicry effect is produced by recording the player's position each frame and replaying those coordinates after a set delay, yielding an opponent whose behavior continually shifts in response to what the player has done. Supporting systems such as procedural level generation, real-time collision handling, high-score saving, and consistent 60 FPS execution round out an immersive, responsive play experience.
The working implementation shows that the proposed framework successfully blends intelligent AI behavior with real-time mechanics while keeping performance stable and gameplay interactive. The delayed mimicry model raises gameplay difficulty without demanding computationally heavy machine learning, giving an efficient and scalable AI approach for arcade-style titles. The architecture leaves room for future additions such as adaptive difficulty, reinforcement-learning-driven enemy behavior, multiplayer support, procedural level design, and more advanced AI decision-making. Shadow Tag thus offers a fresh way to bring behavior-driven artificial intelligence into modern 2D survival games while showing how AI concepts can be put to practical use in interactive entertainment and game development.

Keywords