Scientific Validity

Linear, logistic, polynomial and regularised regression in sport: what the coefficients mean, how to read one honestly, and why the error term is the honest part.

Which sporting metrics actually move together — real coefficients for football, basketball and tennis — how correlation feeds a model, and why it never proves cause.

Frequencies, Bayes’ theorem, the distributions that fit each sport, correlation and regression — and the three ways statistical work in sport usually goes wrong.

Poisson, Elo, logistic regression, boosting and neural architectures — which model produces which kind of number, and how each one is validated.