This week the lines-makers seem to have gone a little too far for almost every NFL game. My model picks the underdog in each matchup (against the spread, obviously), except the Titans-Jags game. Surprisingly, it picks the favorite, the Jaguars, in that game (and they covered).
According to oddsshark.com, underdogs are 110-111-4 against the spread this season. So it seems unlikely that the line-makers are consistently off. Perhaps my model is over-regulated and holding predicted scores too close to the mean score. I will do more analysis, but know that this is a model that would win almost 53% of its bets over the past 30 years if it bet on every game.