ml math and statistics that actually show up

ML math is useful when it helps you read a system without worshipping the output. You do not need to become a mathematician before you can ask better model questions.

This chapter keeps the math tied to engineering moves: read a vector score, compare a model to a baseline, notice spread in a sample, catch leakage, and write a sanity report before sharing results. The drills use plain Python so each calculation stays visible.

By the end, you should be able to explain what a score is made of, what baseline it must beat, what uncertainty or leakage could make it fragile, and what evidence should block a premature claim.