Player Performance Estimation Through Adaptive Puzzle Levels Generation in Game-Based Assessment

Game-based assessment frameworks can effectively measure student proficiency in basic integer arithmetic by integrating procedural content generation with dynamic difficulty adjustment. By utilizing the B-Block platform, this approach tailors puzzle complexity to the individual learner’s zone of proximal development. The system systematically calibrates challenges based on four weighted criteria: block count, operator complexity, operand magnitude, and operand range. This model-free architecture ensures that educational content remains responsive to real-time performance logs, maintaining an optimal balance between engagement and cognitive demand.

The efficacy of this methodology is supported by a strong positive correlation of 0.734 between gameplay performance and traditional paper-based test scores. Statistical validation, characterized by a low root mean square error of 0.031, confirms that the algorithm accurately estimates student ability while maintaining consistent assessment standards. By aligning technical puzzle parameters with pedagogical arithmetic requirements, the system provides a reliable, scalable alternative to conventional testing methods.

This research establishes a robust framework for automated educational evaluation, demonstrating that adaptive game design can serve as a valid proxy for formal academic assessment. By leveraging artificial intelligence to classify questions and generate content, the system offers a personalized learning experience that mirrors the rigor of standardized testing. These findings suggest that integrating adaptive procedural generation into educational software provides a viable path for enhancing the accuracy and scalability of performance monitoring in digital learning environments.