Machine Learning & Analytics
Independent project · Target March 2027
Hockey Metric Reliability
How many games before a given player metric is signal rather than noise, and how large a within-player change has to be before it is detectable rather than drift. The same reliability machinery as SwingDNA, on public hockey data.
Overview
A reliability study of hockey player metrics: how many games it takes before a given metric is signal rather than noise, and how large a within-player change has to be before it counts as a real change rather than drift.
It reuses the same reliability and minimal-detectable-change machinery developed for SwingDNA, applied here to public hockey data. The question is the same one in a different domain — telling a real difference apart from the measurement noise around it.