Since everyone is coming out with their conference predictions this time of year, I thought I'd try to do the same, with a different methodology. Instead of making an educated guess, I used trends from the past 3 years to make my predictions, a very rudimentary predictive analytics. For example, Michigan won 3 games in 2008, 5 in 2009, and 7 again in 2010, for average improvement of 1.33 games a season. Since you can't win only a quarter or 2 quarters of a game (unless you ask Missouri fans in 2009), I rounded down to 1, for a projection of 8 wins this season.
While this method is obviously flawed, as teams who lost a lot of games have more room to move up, it still is pretty close to what most predictions are for the season. It looks like it will be Michigan St. versus Ohio St. for the inaugural Big Ten title game if past trends continue into this season.