Glicko rating system for games

I've been impressed with the Glicko system as I've seen it used at freechess.org and I'm working on a java game rating package that will hopefully include Glicko, Glicko2 and Elo. I'd like to use the system as a predicitve rating system for college football. It could be used to rate more than just games played with football as it has been said that each play is a game in miniature. I would like to rate each player using each play as a competition. That way I'd have more data and would have a better confidence rating. I could also rank offenses and defenses seperately. That way I could not only matchup offense vs. defense, but defensive tackles vs. offensive linemen, corners vs. receivers, etc. The big problem with this is where do I find stats that show who was in on every play. Initially I'm using a predictive period of 4 years meaning that after 4 years if a team did not play I would assume the maximum rating deviation. I'm hoping I could get 4 years of data to do this somewhere, but the data entry problems are pretty large. If I just left it at competitions that's 119 teams * 4 years * 12.5 games / team (about considering conference championships and bowl games). If you consider each play a competition (which I would like to do) for each player thats another 22 * 75 per game. So approximately 9,817,500 data entry items. (Not something I'd like to do myself, even if I had all the data.) If I had all the data in some format already, I'm sure I could automate a lot of the process. I'd also have to come up with some criteria for what is a "win", what's a "loss" and what's a "draw". Obviously on offense a first down is a win. Otherwise, I'm thinking plays of 2 yards or less are a loss as well as 3 down stops and turnovers on downs regardless of the yards gained. Should a 3 yard gain be a draw on a first or second down? Obviously the opposite would be true on defense. For field goal attempts a made field goal or successful fake is a win, and the opposite is a loss.
Punts stopped inside the 20 would be a win regardless of field position. Kickoff returns past the 20 would be wins for the return team. Punts not stopped inside the 20 would have to have some other criteria, say > 35 yards gained is a win, 25-35 yards gained is a draw and < 25 yards gained is a loss? Once I see some data there could be tweaks to any of these criteria.
Anyway, it would be nice, though I may be just dreaming.