ONGOING · CURRENT
Hybrid 3D Expected Goals (xG) Model for Soccer
Building a 3D xG predictive model that captures shot geometry: distance, angle, body part, and defender pressure in three dimensions, beyond what traditional 2D features can express.
WIPcurrently building
PythonPyTorchSoccer Analytics3D Geometry
THE STORY
Traditional xG models flatten the world to (x, y) shot coordinates and a few categorical features. Real shots happen in three dimensions: the ball arcs, the keeper moves, the defender contests in space. I'm exploring whether adding the third dimension materially improves goal probability estimates over the standard 2D baselines.
CODE
REPO TBA — being cleaned up for public release