Netting Expectations: The Football Ontology

This dissertation develops a formal ontology of association football grounded in Basic Formal Ontology (BFO) and the Common Core Ontologies (CCO). Its primary aim is to provide a principled account of the entities, processes, and information structures that constitute professional soccer, with particular attention to how contemporary performance data is generated, interpreted, and used. Despite the growing influence of data-driven analysis in sport, existing representations of soccer data often conflate distinct ontological categories, leading to ambiguity in interpretation and limitations in interoperability. The work proceeds in three stages. First, it analyzes key continuant entities in soccer, including players, roles, positions, and dispositions, clarifying their relationships within a realism-based ontological framework. Second, it models soccer occurrents—such as passes, shots, goals, and officiating decisions—as structured processes situated within temporal and institutional contexts. Third, it examines expected goals (xG) as an information entity, arguing that xG values are best understood as predictive and counterfactual measures rather than as properties of individual shots or players. The resulting ontology demonstrates how performance metrics can be rigorously connected to real-world events while preserving critical distinctions between actions, outcomes, and probabilistic expectations. This approach supports clearer interpretation of soccer data, improved semantic interoperability across analytic systems, and more responsible use of predictive metrics. While developed for soccer, the modeling strategies advanced here are applicable to other sports and to domains that rely on complex event data and probabilistic evaluation.