AI-driven robotic soccer teams use coordinated patterns like offensive formations, defensive structures, and transition dynamics, resembling cooperative equilibria where agents adopt complementary roles. At the adversarial level, teams infer opponent intentions, creating recursive reasoning described as strategic depth. Recent advances combine reinforcement learning with game-theoretic reasoning, with research groups such as DeepMind showing agents can develop cooperative and competitive behaviors in simulated multi-agent environments. However, transferring these capabilities to embodied robots introduces challenges including sensor noise, communication latency, mechanical limitations, and unpredictable environmental disturbances, requiring robust real-time adaptation. Robotic soccer serves as an experimental platform for studying collective strategic intelligence, examining how distributed agents form shared strategies and how cooperation emerges under competitive pressure. Game theory acts as a conceptual bridge between individual robotic decision-making and team-level intelligence. Modern robotic systems increasingly rely on machine learning for decision-making in environments with uncertainty, high dimensionality, and rapid temporal dynamics, with real-time decision-making demanding continuous sensory processing and action selection within milliseconds.