Hmm. I see weights determined by past spending patterns. If they changed weights based 1-year old versus 10-year old data. Where is the manipulation in 1-year old data? I admit it, I don't get it. It seems the older the data, the more past bias.