"What the analysis does is look for relationships between data and outcome," Alamar explained.
Alamar says football is a game of situations and so he takes raw statistics and puts them in context -- for example first and 10.
"How good is a team at passing at first and 10, how good is a team at running at first and 10, how often do they pass the ball at first and 10 and the same kind of stat for defense they are facing," Alamar said.
You don't have to understand the analysis to get the result. Alamar's bottom line for Sunday is New York Giants 28, New England Patriots 27.
Alamar created his predictive model for the Wall Street Journal in 2007. Since then, his Super Bowl picks have been right four out of five games.
Vegas odds have the Patriots by three and the professor's prediction isn't scoring any points in class.
"Tom Brady is not going to lose twice to the same team; it's just not going to happen," Menlo College student Joe Prea said.
Alamar has worked for the NFL. He knows the game is more than numbers, but he's sticking by his formula.
"If it was just a quarterback against quarterback, Tom Brady would win, but quarterbacks never play against each other they just play against other defense and this week Eli Manning has the easier job," Alamar said.
The model's five year track record is 68 percent. One misstep -- it had the 49er's beating the Giants.
"At the end of the day the model gives us likely outcomes, not certainty," Alamar said.
