Most football analytics estimate the value of an action by its downstream effect on scoring, possession, or similar outcomes. This paper argues that such action-value approaches are only part of the picture. It proposes a complementary way to describe a pass: by how it relates to the opposition's spatial structure on the pitch.
The authors introduce a method for learning pass archetypes from spatio-temporal tracking data. These archetypes are meant to characterise the tactical impact of a pass, going beyond simple outcome-based valuation. The work appears on arXiv as a replace-cross announcement, though the abstract does not detail the specific modelling techniques or results.
Because the source is a single abstract, the article only reflects the stated motivation and intended contribution. No comparison with other methods or empirical findings is available in the provided text.