Sports Data Analytics is a growing market despite the fact that it operates in a domain that is highly fragmented, highly specialized and dynamic in its nature. Facing such challenges, the Sports Data Analytics community has developed methods that other domains can learn from. In this presentation we are looking at when models can be successfully learned by an AI and were it makes more sense that domain expertise is translated into analyzable data structures. The audience will learn how in Sport intelligent systems are built by combining Machine Learning, Analytics Translation and good old-fashioned AI.