Possible typo  problem 2.23 // var of H1 on page 65
After finishing problem 2.23 I checked my numbers for bias and var vs those presented in the text on page 65. They agree, up to the fact that for the variance of H1 (linear functions), I find 1.676... which would round to 1.68, not 1.69. I evaluated the relevant integral(s) numerically using Mathematica 10, at varying precision/accuracy settings.
