timtyler comments on Contrarianism and reference class forecasting