If I Ran the College Football Play-offs

First of all, it would be eight teams, not four.  And, the criteria used to pick those teams would be wins and losses (W/L), strength of schedule (SOS), and conference championship participation (CCP).  

The W/L of a team are easy to determine.  The SOS would be determined by wins and losses of opponents and the opponents of a team's opponents.  After an exhaustive (i.e. five or six minute) search on the internet, the current SOS ranking that is the closest to the type I would use is predictionmachine.com. 

So, I’d take a team’s SOS ranking, multiple it by 0.1, add it to their losses, and then subtract it from their wins.  Then, the conference runner-up gets an additional 0.5, and the conference champion gets an additional 1.0, in every conference that has a championship game.  That final number would be their play-off rating (PR).  For example, Alabama’s SOS ranking is 4; multiple that by 0.1 = 0.4.  1 loss + 0.4 subtracted from 12 wins = 10.6.  Add another 1.0 for CCP  = 11.6 PR.  

Four Team Play-off

#1 Alabama (11.6 PR) vs. #4 Ohio State (6.8 PR)

#2 FSU (10.3 PR) vs. #3 Oregon (8.7 PR)

Same four as chosen by the committee.  However, when you go to eight teams, it changes quite a bit.

Eight Team Play-off

#1 Alabama (11.6 PR) vs. #8 Ole Miss (5.2 PR)

#2 FSU (10.3 PR) vs. #7 TCU (5.8)

#3 Oregon (8.7 PR) vs. #6 Miss. State (5.9 PR)

#4 Ohio State (6.8 PR) vs. #5 MIzzou (6.2 PR)

#9 Arizona (4.7 PR), #10 UGA (4.6 PR), and #11 Baylor (4.4 PR) would be left on the outside, looking in.

            What about the chance of one of the “lessor” conference champions sneaking into the play-offs, you ask?  Under my system, Conference USA champ, Marshall would have a 0.1 PR this year.  Mid-American champ, N. Illinois would have a -2.1 PR, and Mountain West champ, Boise State, would have a 2.7 PR.