Getting started:

Getting Started
Using Statsbomb
Using Wyscout
Sign test
Using t-tests

1, Event Data:

Visualising football
Plotting shots
Plotting passes
Passing networks
Case study: decentralised football
Pass heat maps
Case study: attacking play
Case study: 5-second rule
Improving your visuals

2, Models:

Statistical models
Linear regression
Case study: Possession
Introducing expected goals
Fitting the xG model
Case study: the geometry of shooting
History and application of xG

3, Scouting:

Statistical scouting
Radar plots
Case study: spreadsheet scouting
Other approaches to scouting

4, Valuing actions:

Evaluating actions
Markov chains
Expected Threat - Position-based
Calculating xT (position-based)
Expected Threat - Action-based
Possesion Chains
Calculating xT (action-based)
Case study: value and context

5, Randomness and prediction:

Simulating a match
Unavoidable randomness
Simulating results
Predicting football matches
Calculating match outcomes
Statistical modelling
Team KPIs
Points Moving Average
Case study: Long term trend at Hammarby

6, Pitch control :

Controlling Space
Making Voronoi Diagrams
Case study: defence in attack
Pass probability

7, Positioning and Movement:

Valuing off-ball actions
Case study: off-ball runs
Off-ball movement metrics
Expected Goals including player positions

8, Physical Data:

Physical data
Calculate distance-based metrics for player fitness
Modelling the relationship between EPV added and distance covered

9, Clustering:

Clustering progressive passes
Clustering players

10, Making a Web App:

Installing Streamlit
Twelve Visualisations
Twelve Community

Soccermatics

»
Improving shot, pass and dribble models
Edit on GitHub

Improving shot, pass and dribble models

© Copyright 2022, David Sumpter (code and text), Aleksander Andrzejewski (code). Revision 10ccf92d.

Built with Sphinx using a theme provided by Read the Docs.

Read the Docs v: latest

Versions: latest; main

Downloads

On Read the Docs: Project Home; Builds