I don’t want to conclude that lottery might be rational, but I don’t think it is self-evident that the right way for deciding between different probability distributions of utility is to compare the expectation value. We are not living a large number of times, we are living once (and, even if we did, bare summated value would neglect justice).
P.S. Welcome to Less Wrong! You may already have read these, but just so you know: the About page and FAQ are both very handy to new users, and the 2010 Welcome Thread is a good place to introduce yourself formally if you so desire.
I don’t want to conclude that lottery might be rational, but I don’t think it is self-evident that the right way for deciding between different probability distributions of utility is to compare the expectation value. We are not living a large number of times, we are living once (and, even if we did, bare summated value would neglect justice).
It’s not self-evident, no, but it can be shown under some reasonable assumptions. If you don’t play according to expected utility, then you take the risk of being convinced to do something silly, like buying one lottery ticket for a high price, swapping it for a second ticket, and then selling your new ticket for a lower price.
P.S. Welcome to Less Wrong! You may already have read these, but just so you know: the About page and FAQ are both very handy to new users, and the 2010 Welcome Thread is a good place to introduce yourself formally if you so desire.