Ah, I see - You could find the control store by comparing the trend and seasonality columns of two stores for the same time period with the Numeric Scorer node. Then select the control store candidate with the greatest R-squared values. What do you think?