While in my Statistics II class, I developed a spreadsheet to generate outcomes of NFL games using specific statistics from teams (each weighted accordingly). I then matchup the teams and compare the score difference I generate to the current Vegas spread.
I am not expecting this to predict 100% of outcomes (or even close to that); nor should you. It will take trial and error to see which statistics matter the most but I have referenced the previous five years to get my current information. However, the NFL is constantly cycling through which teams are good/bad, players move from one team to another, and of course there are teams, like the Chiefs last year, that just surprise everyone.
First Pick Of The Year:
Seattle -5 (Projected Score Difference Of 11.488 with a STDEV of 3.937)