Classic arcade fighting games employed sophisticated yet resource-efficient techniques to create dynamic enemy AI. Developers used finite state machines where enemies transitioned between predetermined behaviors like attacking, blocking, and retreating based on player proximity and health. Pattern systems created challenging opponents through carefully designed attack sequences that players could learn and counter. The AI frequently read player inputs to create responsive difficulty, with many games implementing dynamic difficulty adjustment that scaled enemy aggression based on player performance. Memory limitations led to clever programming tricks like probability tables determining attack likelihood rather than complex decision trees. Enemy behavior often followed simple but effective rules: maintain optimal fighting distance, exploit player weaknesses, and adapt attack frequency. These constrained systems created the illusion of intelligent opponents while working within strict hardware limitations of 1980s and 1990s arcade machines.