Charles Dillon comments on Data on forecasting accuracy across different time horizons and levels of forecaster experience