First-time poster, bet a lot of football and football.
It's my opinion that all of the numbers discussed here are kind of beside the point, because the odds already reflect every aspect of the game in question. The person who set the line, and the market that has moved the line since, has done so to reflect reputation, the market's expectation, and each of the fundamentals named here, from yards per pass and offensive efficiency, from turnovers to running attack to running defense. That one number can serve as a proxy for all that other stuff.
*If* you think yards per pass is the most important thing, you can give extra weight to it when you cap a game, but do so in the knowledge that the ultimate result of the game is a coin flip. There are just too many variables involved when you've got 53 players on each side, a dozen coaches on each side, and seven refs on the field and more in the booth.
Since odds' (and point spreads') fundamental purpose is to split the market down the middle, thus allowing the bookmaker to win no matter what happens in the game (and creating essentially a coin-flip situation), the thing you can do to better your chances of guessing right is to take plus-money wherever possible against any 50-50 line.
This is far easier to do in world football, in a three-way market (home win, draw, away win). If you know that particular teams or particular leagues show an extreme tendency toward underdog wins or underdog draws, you can tailor your strategy accordingly.
In the NFL it is best to fade the public, because the public is stupid. The public doesn't understand recency bias, for example. Regression to the mean is real.