Week 6 · Day 3 — Model Evaluation: Not All Mistakes Are Equal

Week 6: "The Scout" | Day 3 of 5 Theme: Confusion matrices, precision, recall, F1 — knowing when your model actually works

The Problem with "Accuracy"

Imagine your model predicts injuries. Your dataset has 950 healthy players and 50 injured ones.

A lazy model that always says "healthy" gets 95% accuracy. But it completely misses every single injury.

That model is useless. Accuracy lied to you.

Today you learn the metrics that actually matter — and when to use each one.