Will_Sawin comments on Raising the forecasting waterline (part 1)