As Sgt. Wolverine noted (as have others), a playoff only crowns a playoff champion, not a true National Champion. That's why, instead of trying to shoehorn some logic into the BCS or crying out for a 6, 8, or 12, or 16-team playoff, college football might as well just go back to its earlier pre-1998 system and just accept the "Mythical" before the NC and give the power back to the human polls.
Now hear me out. I look at sports at the elite level the same way I view entering an elite university, the gaggle of beautiful people trying to enter a trendy club, or obtaining a great job. At this level, when everyone/everything has virtually the same resume and skills, people are left trying to distill out the slimmest of stratification to justify why one team is better than another, and by default people believe that head-to-head somehow "proves" who is better. Yet, as we've seen numerous times, there is a myriad of variables that can greatly affect even the outcomes of these head-to-head match ups, let alone a full playoff. First off, when you play an opponent can have at least as much of an effect as who you play. We all remember 2006, when OSU and UM were two of the best three teams in America, played each other in a solid game, and then had to wait 50+ days to take on Florida and USC. Rust, overconfidence, over-scheming, etc. can all set in over that time, and what you get are two blowouts that I severely doubt would have happened if the games were played a week after the Big 10 finale. Conversely, look at Texas this year - they played 4 ranked teams on consecutive weekends, and then finally lost on a last-second touchdown grab. Teams get hot, teams get tired, players get hurt, etc., and when the talent is approximately equal, all a playoff shows is who was able to get hot, avoid injuries, and find a favorable draw for 3-4 weeks. So not only does a playoff only crown a playoff champion, but it now also crowns a playoff champion for the past month.
I still remember the kerfuffle that Colorado made a few years ago when they got red-hot at the end of the season, destroyed Nebraska (the fighting Eric Crouches) and beat Texas for the Big-12 title, and complained they were shut out of a chance at the NC even though they were playing the best football in the country at the time (of course, they then were destroyed by Mr. JJ Harrington and Oregon, yet more proof that "when" is a big deal). Of course, this same team had already lost to Fresno State and rather convincingly to Texas earlier in the season. This team wasn't the best team in America that season, but match them against Miami that year, one week after beating Texas, and who knows how that game plays out.
And the benefit of a favorable draw also plays into the inherit weakness of a playoff in CFB because it all falls on who you play. Look at the participants/scores in last year's BCS games:
Rose - USC/Ill (49-17)
Sugar - Georgia/Hawaii (41-10)
Fiesta - WVU/OU (48-28)
Orange - Kansas/VT (24-21)
NC - LSU/OSU (38-24)
I see a slew of blowouts plus one game (KU/VT) that neither team really seemed to want to win. So does that mean USC, Georgia, WVU, KU, and LSU were the best 5 teams last year, and LSU was the clear champion? Of course not. In fact, I would hazard to guess that if you flipped LSU with Georgia, that NC game turns out differently. Or flip WVU and KU, and maybe OU doesn't get obliterated and the world is deprived of the Bill Stewart Face. And yet, because these lineups were defined by a combination of tie-ins and expected ratings, people somehow believe that they provide an "objective" metric for how good each team was. That just doesn't make sense to me. These match ups never aligned the best teams against each other no more than the traditional bowl tie-ins created the best match ups for determining a true NC. All you have are a bunch of somewhat-arbitrary match ups that may place a superior team in the crosshares of some insanely-hot team, or a team that just lost its best player for some reason (remember Cinci back when Kenyon Martin played - that team was a legitimate NC contender in basketball until Martin injured himself right before the tourney).
Well, the same thing would happen with a playoff. How do you decide who is "better" when the BCS and the polls have yet to show an ounce of consistency in determining who the best 2 or 3 teams are a given week? Sure, SOS, conference strength, head-to-head match ups, etc. provide a component, but ultimately you are just left with more distractions and after-the-fact justifications, not a true barometer of what teams truly are the "best".
For example, take this season. Provided Alabama loses this weekend to Florida, why should Florida be a #1 seed and Oklahoma a #2, or vice-versa? How about Texas, Utah, Alabama, and PSU battling over the #3-6 spots? Depending on which slot each of these teams fall into, their chances of winning or losing vary wildly. Presuming Boise St. and Ball St. round out this 8-team playoff, Texas would play a weak Ball St. team while Alabama and PSU would hammer away on each other for the opportunity to then play Oklahoma or Florida in the next round. One week earlier, Alabama is the #1 team in America, and now you have them fighting for their life against another team that was one missed field goal away from a perfect season. That's the second inconsistency of a playoff - you may only be as good as the team you match up with that weekend, yet everyone somehow presumes that all the other teams that lost either contemporaneously or earlier are somehow worse. So not only do you only crown a playoff champion for the past month, but you now only crown a playoff champion who played that particular configuration of opponents.
And yes, a round-robin works in theory, but there is no way you can expect college football teams to play into February, even with a shortened regular season. The games are too nuanced and too physical - the reason it works in March Madness is because players and coaches generally can gameplan and recover for another opponent in a few days, while in football the players would be walking corpses after late December.
Finally, the third factor that people tend to forget about a playoff series is that certain teams enjoy the benefit of home field advantage even when it was not necessarily intended. Since I presume the only way a playoff will work is with the consent of the bowl alliance, most of these playoff games would need to be played on "neutral" fields sponsored by the various bowls. True, you should reward teams with "better" records with home field, but you'll inevitably have the problem we've seen over the past few years with certain warm-weather schools (looking at you, LSU and USC) playing meaningful games virtually at home because of these relationships. Heck, USC virtually never has to leave Southern California during bowl season, and I still believe that LSU playing in the Superdome last year helped them immensely against OSU. Sure, the traditional bowl games were not anything better (USC still has the Rose Bowl, LSU the Superdome, Florida/FSU/Miami the Orange Bowl, etc.), but at least there you could argue bowl tie-ins were at work and every team kind of understood the worst-case scenario at the beginning of the season. Admittedly this is the weakest of my arguments, and that is why I left it until the end. That said, now you are left crowing not just a playoff champion who played a particular configuration of opponents over a shortened time frame, but also a playoff champion who played those games at particular fields.
If you have made it this far, then congrats, and you are almost done. I used to be a huge fan of playoffs in football, believing that it somehow crowned a truer champion than the USA Today/Coaches poll or the BCS title game. Yet, the more I thought about it, I don't see it doing anything more than slapping a different coat of paint on the same broken-down car. Implementing a playoff would, ultimately, appease some fans for a short period of time until its flaws would begin to show, probably when a team like BYU runs the table or a 3-loss SEC team gets insanely hot and wins the NC. Personally, the only thing I dislike more than the uncertainty the polls created is the faux-certainty and legitimacy that the BCS gave to each season's MNC. At least with the polls, early-season struggles were at least partially remembered, and getting hot at the right time wasn't necessarily rewarded with any more or less outrageousness than in a playoff. And for those concerned that the polls would overlook worthy teams like Utah or Boise St., I'm all for a rule that would allow a mid-major ranked in the top-10 to knock out a lower-ranked opponent with a tie-in from one of the "big" bowls. Sure, that might cause the occasional outrage and isn't perfect, but then at least you give these teams a chance to showcase their abilities against a top-flight opponent. Best-case scenario, a top-flight team like Utah goes undefeated and miraculously bypasses all of the 1-loss teams and is named the MNC; worst-case scenario is they go undefeated and are still passed over by a 1-loss power team, which is the exact same result we've seen since the beginning of the BCS/time. I am all for abolishing the BCS; I just don't think we need to replace it with some convoluted playoff system.