I’m trying to talk myself into Madson. I’m on a phone, so didn’t do the addition or the table, but if you look at 3 year averages, he is significantly better when he is not used on back to back days. Most of the “better” is in 2016-2017 though. Last year, he was pretty bad across the board, though these are all really small sample sizes for any one year. So, it seems like bringing him in and making sure to limit back to back days might be something their analytics team could have mined for a cheap potential upgrade.