Background: A topical research question is how pre-service teachers develop professional competencies with regard to game-centered approaches such as 'Teaching Games for Understanding' (TGfU). In this context, research on teacher education assumes reflection on teaching examples -- including real-life classroom practices -- to have positive effects on teachers' professional development. The role of media used in such examples is unclear, however. While text-based teaching examples might be more favorable for novice learners due to a step-by-step presentation of information, video-based examples that capture the complexity and reality of teaching situations might better foster the learning of advanced pre-service teachers. Method: In a field experiment, we taught principles of TGfU to 136 undergraduate physical education (PE) pre-service teachers and randomly assigned them to three conditions: they either reflected by themselves on a video-based or on a text-based teaching example or completed an open-ended planning-task. For the video-based example, a PE teacher and his class staged different scenes of a teaching unit following a script that focused on 'badminton-like' games considering steps and pedagogical principles of TGfU. The video-based example was transcribed and converted into a text-based example. Before the intervention, we assessed the pre-service teachers' knowledge of gamecentered approaches. After the intervention, we measured their knowledge of TGfU and affective-motivational variables (interest, challenge, enjoyment, annoyance). Results: The participants had a (rather low) comparable knowledge before the intervention. After the intervention, we found that reflecting on the video-based example had increased the knowledge of TGfU more than reflecting on the text-based example or completing an open-ended planning-task (medium effect). We found no interaction effect of prior knowledge and condition on the knowledge of TGfU at posttest. Participants with a video-based example reported higher interest and enjoyment after the intervention (medium effect). All other affective-motivational variables did not show significant differences. Discussion: Other studies concentrating on teaching in classrooms found that text-based examples fostered pre-service teachers' competencies better than video-based examples if they had rather low prior knowledge. One explanation for the present finding could refer to the specific field of study: For novice PE pre-service teachers the supposed benefit of text-based teaching examples does not advance the learning process more than the supposed benefit of video-based examples (realistic display of constitutional aspects of PE). Since interest is closely related to (intrinsic) motivation, and the learners' motivation is fostered by realistic and authentic context, our finding regarding the condition effect on interest and enjoyment is consistent with assumptions of situated learning theory. Higher motivation, in turn, could lead to a more elaborate analysis of the video-based example, which may ultimately explain the higher level of knowledge, too. Conclusion: The current findings indicate that video-based examples are a very promising means for PE teacher education. This refers in particular to autonomous work in teacher education because in the present study the pre-service teachers reflected on the teaching examples by themselves.