Ben Kuhn comments on Expected value estimates you can take (somewhat) literally