Good question! For this post, I only applied regression testing to predicted v. actual. I am away from my database currently to see the actual coefficient, but it was still around the .21 mark, slightly under. My skew this week happened mainly with inactives and injuries. All those dots on the x-axis really didn't help with my overall accuracy. I am going to work on squishing those out in weeks to come.

In backtesting, my highest tested r^2 was about 0.40, but this has only happened after I've had a half season or more of data to work through.