"Great article, but when you write
"Standardize X(n) so that its variance does not depend on n, that is, introduce Y(n) = X(n) / SQRT(Var[U(1)]). This step consists of re-scaling the Y-axis."
should it not read Y(n) = X(n) / SQRT(n) ?
"You won't need a maths degree to read "Elements of Statistical Learning"? Perhaps not if you just want to follow along, but I'd be surprised if anyone without degree-level maths knowledge can actually understand it and solve the…"