I'm going to reveal myself as entirely too geeky here, but my primary complaint about this approach is that it relies on a linear scale for evaluating utility,

I'm going to reveal myself as entirely too geeky here, but my primary complaint about this approach is that it relies on a linear scale for evaluating utility, when much research suggests that utility curves are frequently logarithmic. (Example: Going from earning $20k to $100k per year is a huge difference with substantial implications for financial security, but $1m to $1.08m has a substantially lower impact.) One could argue this article looks at a restricted range where the log behaves more lineary, but if we're going to apply mathematical modeling to our life choices, ... :-)