Douglas_Knight comments on Contrarianism and reference class forecasting