As in our example above, the total auxiliary series may contain a greater portion of private household consumers which are more weather dependent than industrial customers. The forecasts for the auxiliary series will then be too sensitive to weather changes than desired and if the setting “Max model size” is set to anything greater than 0, the program will estimate the proper reduction of the weather dependencies. In the previous step, the hourly averages for a number of weather parameters were saved along with the hourly day type patterns. We will now express the forecast series as a linear combination of the hourly day type quotients and the deviations from the day type weather.