Causal Forecasting

Hi,

I had a couple of questions on Causal forecasting.

1. If we are doing week level Causal forecasting and have day level promotions - how are the effects from multiple days in history applied in the future. For example, historically I had promotions running on Fri, Sat & Sun in a particular week and I decide to run the same promotion in future only on the Sunday.

Would the cumulative effect from all the 3 days (in history) be normalised and applied to the future which has only a 1 day promo planned ? Or is there some other logic ?

2. In the section related to Daily Causal Forecasting in the user guide, there is reference to the week level forecast being calculated and spread down to week.