OK, Advanced NFL Stats data is in. Here’s the chart, using an average between ANS and ML win probabilities:

One major difference with ANS and Moneyline is that ANS has the Jets over the Patriots. Brian Burke, who created ANS, talks about this in his article this week. This could definitely be a contrarian pick candidate.
This week, there is this one rare no brainer pick… Carolina all in over St. Louis. ANS, Moneyline, and conventional wisdom suggest Carolina to win, but yet 22% of the crowd picked St. Louis. The Giants over Minnesota has something similar. Therefore, I will be bumping both teams all the way up in my pick.
Philly over Dallas (32% public pick, 65% average win probability), Pittsburgh over Baltimore (31%, 69%) and – to some extent, based solely off ANS – Jets over the Patriots (5%, 53%) are good contrarian picks here. I’m going to stick with the base here, because I think there’s enough contrarianess versus the public:

So for Dad, since he’s back in town, I asked him which of the green highlighted teams he would pick as an upset. He said he’d pick Dallas, Baltimore, and New England… which aren’t upsets, but interesting to see that he fits with what the general public picked. When I asked which team would he pick if he had to pick an upset, he picked Philly over his own Dallas Cowboys (he doesn’t believe in Jason Garrett, or any coach not named Tom Landry or Jimmy Johnson for that matter).
I ended up going with Washington all-in over Chicago (16% picked, 43% win prob) and keeping everything else base. It was hard trying to decided if I should hedge my picks… Carolina and the Giants had too high of win probabilities to go opposite on.