A couple of mathematically oriented comments...
Let us presume that the quality of college football teams lie along a standard bell shaped curve. As you move away from either of the extremes, the number of teams at whatever level increases until you get to the middle of the curve. What that means is that as you allow more teams into the playoffs, there will be more teams at almost the same level of quality as the last team let in. Thus, while it might be hard to decide which two teams most deserve to play for the championship, it will be even harder to decide which should be the fourth team that deserves to be in the playoff (so this is why the selection committee will surely get criticized). I might add, that the amount of controversy over which team should be the last team allowed in would only intensify if the number of teams in a playoff increases to eight.
Let us suppose that one team is decidedly better than the others in the top four. They would beat any of the other top teams four out of five times they were to play each other. (I doubt that any team is good enough that they would beat another top four team every time they played. Look at the SEC.) With a single game playoff, the probability that the dominant team would win the national championship would be .80 (eighty percent). If, however, they have to go through a two game playoff, the probability that this better team would win the championship drops to .64 (sixty-four percent). A college playoff system might be good for generating revenue (and distracting "student athletes" from their academics - most players on even the best team will never play a game in the NFL), it is not a great system for determining which team deserves to be declared the national champion.
(Okay, I know that these comments will not be popular, but they are what mathematics tells us.)