My approach in week 2 is to take what we think we learned, and bet the opposite. So STL is a hot mess? Well they may will cover. Pitt looks unstoppable. I will probably take the points vs CInci.
There is some real value due to what "we think we know" after 1 week. This is not to say things can't be learned. But the value of being a contrarian, I bet, near its height in week 2 when people overreact to a small sample size.
So maybe I bet against SF, on SD and on Washington. (Or at least, I lean that way).