Is it possible that your projection have problem with the most variance dependant variable wich is TD reception/Rush TD? Except for those games of the decades (like Julio's).
I'm juste curious, i'm not that great at maths :P
You are viewing a single comment's thread from:
It is quite possible, though those are accounted for in my multiple linear model I use to help adjust 75th quantile metrics. I utilize a rolling average of targets, touchdowns, and red zone events to help make next week projections. I just think I have bad sample sizes, and super high variance so far. I am trying to keep an eye out for these events though, and am constantly looking for ways to continue to tweak!