I have been using this system since 2009 and it wasn't until this year that I decided to put my own spin on it and there have been some interesting results. I usually don't start using this system until about week 7 because it entails plugging in the points for and points against for the first 6 games of the season for each team playing that week. After that you calculate the average points for and average points against for the teams that are facing each other that particular week. You also remove the highest points for and lowest points for and highest points against and lowest points against for the teams facing each other. You would then add the average points for, for the visiting team to the average points against for the home team and subtract 20.5 from that number. This gives you the projected points that the visiting team is going to score. You do the same thing with the home team (add the average points for, for the H team to the average points against for the V team and subtract 20.5). This will give you the projected final score of the game.
Since there hasn't been 6 games played yet this year I have been skipping the part where you eliminate the highest and lowest points for and highest and lowest points against for each team facing each other and have just used all points for and points against. The system has not worked well at all so far this year with a record since week 2 of 30-44 but it has predicted some pretty good upsets. The problem is the inconsistancy and not knowing when it will actually work for the team that may or may not come through with one of these upsets.
I have to restart the computer I'm on right now so I'll be back to tell you the rest of this system story in a moment.