Scientific competitions are held for more than a hundred years now and many paradigms have been experienced meantime depending on domain and pursued objectives. In Computer Science, most of the programming competitions are based on the time-to-complete paradigm for more than 30 years. Though this approach is undoubtedly successful for some contestants and some objectives, lately new paradigms have appeared that try to attract more students, to cover more areas of Computer Science, and provide for achieving more learning objectives. This paper presents an academic experience on an internal Computer Science competition, spanning on more than three decades, the focus being on its evolution and its continuous adaptation to the current learning needs, objectives, and contexts. Alongside with the motivation of changing the traditional paradigm, this paper describes in detail our new approach core ideas, our experience with it, while highlighting all the risen issues. After having had two editions already, we can provide some insights on its impact on team-based competitions, on its proved value for both promoting students' ideas and strengths and working on their weaknesses, whilst competing for both fun and success in a flexible way.