This offseason I developed a win-probability model for College Football. The Model is built off of play-by-play data from 2009-12 and considers the Down, Distance, Spot, Lead, and Time Remaining of each play, as well as pre-game factors such as who is home and FBS vs. FCS teams. You can read more about the development and accuracy here.
Feel free to leave a request in the comments if I didn't include one you would like to see.
Alabama at LSU
My model doesn't handle over time right now, and unfortunately the bulk of the exciting parts happen with less than two minutes left so they are tough to see. I'm surprised LSU's win probability wasn't higher when they recovered the fumble and had 1st and goal, but then again they did end up losing the game.
Texas A&M at Auburn
Auburn's two fumbles at the end of the game are the near vertical orange lines. They probably could have overcome one of the fumbles, but not both.
Ohio State at Michigan State
Michigan State had a chance to put this away at the end of the 1st half but two huge plays by Ohio State turned the tide.
Notre Dame at Arizona State
I haven't had time to calculate individual player Win Probability Added for a game, but I would guess Everett Golson had one of the worst games imaginable in terms of WP lost for his team.
Kansas State at TCU
Oregon at Utah
The Touchdown-that-wasn't occurred at the start of the 2nd quarter, so at exactly 15 minutes of game time elapsed. The magnitude of that play isn't really captured by this, but over the last 5 years home teams with a 14 point lead near the start of the 2nd quarter win 96% of the time. When they are tied they only win 61% of the time. That blunder and subsequent touchdown was a huge swing for Oregon.
Baylor at Oklahoma
Virginia at Florida State
Georgia Tech at NC State