MagicStrategies investigates how individual learning styles of students engaged in web-based education can be accommodated by creating personalized learning content. We propose methods to match learning styles and content types. To that effect a software simulation was created, which can generate any number of learner profiles and attempts to match them with learning objects from a repository. The simulation demonstrates and tests a matching algorithm between learning styles and learning objects attributes. In addition, we show how the system can automatically present educational content in more appropriate media types if the students perform poorly with the current set. Thus we bridge Computer Science and Mathematics with Social Science and Instructional Design and use methods from all worlds cooperatively. Our methodologies apply to leaner-content, employee-task, employee-team, soldier-mission, and other similar relationships.