Matt Goldenberg comments on When Observation Beats Experiment