Choosing Playoff Participants: Poll, Committee, or Autobid?

Submitted by oakapple on May 16th, 2012 at 3:15 PM

All of the key players (conference commissioners, bowls, TV) seem now to agree that some kind of four-team playoff is coming to college football. The challenge now turns to exactly how the four participants will be selected.

Three main options have emerged: 1) Polls; 2) A Selection Committee; or 3) Autobids for the four best conference champions. I'd like to explore the pros and cons of these options and suggest the likely outcome.

There are also hybrid options, which I'll get into below.

Polls would be the most straightforward extension of the system we have now: instead of the top two qualifying, the top four would qualify. This is not to say that the BCS standings would be computed as they are today, only that some combination of computer and human polls would determine who gets in.

A Selection Committee would be a system similar to basketball, where a small group of experts would weigh the candidates and choose the best four.

Autobids would take the decision almost entirely out of human hands: the four best conference champions would qualify for the playoff. Of course, you'd still need polls to decide the four best, but the influence of the polls would be greatly minimized. For instance, last season the rankings of the six BCS league champions were 1, 3, 5, 10, 20, and unranked. I doubt that any rational observer would have had a serious argument with that order.

A few Hybrid proposals have been widely circulated. For instance, Jim Delany has suggested autobids for conference champions, but only if they're ranked at least sixth. This system would enhance the value of winning your conference, but would eliminate weak champions from playoff consideration. (Last year, UCLA had a shot at winning the Pac-12 with a 6-6 regular-season record.)

The pure "Autobid" option seems to me highly unlikely to be selected, because it would allow weak conference champions into the playoff. One could argue the relative merits of Oklahoma State and Alabama last year. But in a system where four teams qualify, could anyone but a Wisconsin fan really claim that the Badgers, despite winning the Big Ten in a down year, had a better season than the Crimson Tide? Seriously?

A playoff that fails to select the four best teams is not credible. Conference champions aren't necessarily the best, except perhaps in their own league (and sometimes not even there). And in any event, the SEC would never agree to that—and no proposal will be adopted that the SEC doesn't support. That's the political reality.

Another political reality is that you need a system that accommodates Notre Dame. Irish-haters may say, "Let 'em join a league." But no one with any actual authority in the matter is trying to freeze the Irish out: ND Athletic Director Jack Swarbrick is a full voting member in the BCS negotiations. And the various interested parties (bowls, television) won't accept a system in which the Irish can't qualify. Whatever you think of them, the Irish are a big draw. So there needs to be a system that accommodates highly-ranked teams that didn't win a conference.

The pure "Polls" option is the easiest to understand: reach the top four, and you're in. But it can lead to some anomalies. Last year, for instance, Stanford was #4 in the BCS standings, but the Pac-10 champion Oregon Ducks, who beat Stanford in the regular season, were #5. The Delany proposal, which gives priority to conference champs, as long as they're in the top six, would correct for that. So would a rule that allows conference runners-up only if the conference champion also qualifies.

Wisconsin Athletic Director Barry Alvarez favors a selection comittee, which he feels is more transparent: if you don't like the result, you'd know exactly who was responsible. This differs from the current BCS standings, where most people don't even know what algorithms the computers are relying on.

But where would you find a committee of manageable size, whose loyalties wouldn't be in question. Alvarez is apparently unaware of the irony, when he suggests Kirk Herbstreit as an ideal member of the committee, saying the former Buckeye "is neutral, is on top of it, [and] talks to coaches around the country."

Can you imagine the uproar if there's a close call between #4 and #5, and Ohio State is one of the two teams? Or if it's Michigan? No one would believe that Herbstreit was neutral in that transaction. And of course, as ESPN's #1 booth analyst, he'd be helping to decide the participants in the very game he was going to broadcast. It's hard to imagine a more obvious conflict of interest.

This is not to single out Herbstreit. I can't imagine who you could put on that committee, who would be accepted as truly neutral. There is hardly anyone in college football who doesn't have some allegiance (real or perceived) to a particular school or conference.

I do realize that the NCAA has a selection committee for basketball, but it's a completely different situation. The lowest seed ever to win the tournament was #8 Villanova in 1985. The lowest seed ever to reach a Final Four was #11 (which has happened twice). So none of the teams that get left out, the so-called bubble teams, have any realistic shot at winning it. Narrowly missing the tourney is no doubt frustrating to the fans of those particular schools, but the rest of the country doesn't care.

In a four-team football playoff, the difference between #4 and #5 is immense. You might not like the BCS standings (Barry Alvarez clearly doesn't), but there are so many inputs to them that there is no one person you can blame if you don't like where your team is ranked. On any committee, there invariably would be a backlash in close cases.

That doesn't mean the BCS formula is perfect. For instance, the coaches' poll ought to be dropped: it's a clear conflict of interest. And the computer rankings that participate today are not the best ones. But the basic idea of a ranking based on some combination of computers and human polling is the right way to go.

There's a lot of talk in college football about "honoring" the regular season. It's a bunch of hooey. Under any of the suggested formats, your team will have to have a very good regular season to qualify for the playoff. Either you'll have to win your conference, or be ranked in the top four to six, or some combination of the two.

In short, what I think we'll wind up with is some kind of hybrid qualification model, which will favor conference champions, but with a provision for wild cards, and/or a requirement that conference champs be rated at least sixth. And I think the rating mechanism is more likely to be a combination of polls and computers, not a committee.



May 16th, 2012 at 3:38 PM ^

I think we should poll coaches and ADs which candidates they would like to form a committee that would then determine which conferences get autobids.

Edward Khil

May 16th, 2012 at 3:49 PM ^

...for the first 3 of 4 teams. The next team(s, if there are fewer than 3 in the top 6) should be the highest BCS-ranked team(s) outside of those already selected.

And the Coaches' poll doesn't need to be abolished. However, it should only be conducted after the season is over.

A selection committee is way too problematic.


May 16th, 2012 at 3:52 PM ^

....about requiring polls that participate in the BCS formula to refrain from conducting pre-season polls. I certainly hope that the USA Today/Coaches and AP polls adopt the approach used by the Harris Poll.

The Delaney hybrid proposal seems the most sane of the bunch. Clearly sanity will have little to do with what approach is selected, but one can hope.


May 16th, 2012 at 8:35 PM ^

I, too, would love to see the elimination of preseason polls. But the problem here will be, even if preseason polls are de jure eliminated, what's to prevent sportswriters and other Harris Poll members from saying what their "unofficial" ranking choices would be? Even if they only announce their "unofficial" top 5, it's the top 5 that actually matter here, so it's still a problem. You can't really keep the voters from thinking about how they would rank teams before all the information is available (i.e. before a few meaningful games have been played). And once they've started thinking about it, the biasing has begun, whether or not they've published an officially sanctioned set of votes. The only place it might work is for the coaches' poll, because they don't have the time to spend thinking about polls in the first place.


May 17th, 2012 at 2:44 PM ^

But if you make the votes by each voter public, they would have to justify their rankings. If this is the scenario, you would hope there would be uproar If they had an SEC team ranked 5th despite playing only non-autobid teams and Boise State out of the top ten even if they were also 4-0 and beat MSU, for example.   By waiting 4 games into the season, you will at least have some basis to go on. 

It would also prevent the defenders from stressing their team's early season victories over overrated teams by saying they beat FSU who was ranked #10 at the time (despite finishing the season with 6 losses)


May 16th, 2012 at 4:07 PM ^

I choose a system were all FBS teams throw their names into a hat, which is shuffled and drawn lottery style. The top three winners then proceed to a three-way rock-paper-scissors match, where they play three rounds (it's go on shoot, we're not communists), and the top two, after best of three, play each other for a chance to go to the national championship NEXT year. Now this is where it gets tricky... the 4th team is voted on by a knowledgeable brain-trust committee of Norwegian ice-dancers, who also have supreme executive power to veto any decisions made during the RPS process. The games will be held only in foreign countries at a location no further than 3 miles from an active volcano.

At least it would be less confusing than the current system.


May 16th, 2012 at 4:44 PM ^

A hybrid approach with 8 teams is much more intuitive and just makes sense.

Take all 6 AQ conference winners and force them to be in something like the top 15 in the polls. Then you have 2 wild cards for the highest teams remaining (AQ or non AQ). If a conference winner is not in the top 15 in the polls, you simply add the next highest wild card. Notre Dame would be guaranteed in as a wildcard if they reach the top 8 or something - a fair poll threshold due to the pros/cons of not being in a conference.

And no, taking 8 out of 120 FBS schools instead of 4 does not significantly "cheapen" the regular season any more than 4 teams would. If anything, it avoids the very possible problem scenario of >4 undefeated teams.


May 16th, 2012 at 4:59 PM ^

However, I was limiting the analysis to things that actually have a chance of happening. It's overwhelmingly clear that we aren't getting an 8-team playoff this time. I do believe we'll have one eventually.

And once you're at 8 teams, I think you've solved the problem, as long as you don't corrupt it with too many autobids. I can't remember a #9 team that had a credible argument that they were the best in the country. So with an 8 team event, you'll surely be getting everyone that belongs.

But if you do that, you have to accept the possibility that a #8 team will get hot at the end of the season, and win it all.


May 16th, 2012 at 5:18 PM ^

And you're right that it seems to be 4 teams or nothing at this point. I think it's just a matter of taking baby steps to the optimal 8 team playoff model. I wonder, do most of these decision making bigshots "know" that the 8 team model is the best and are just slow-playing getting there so as to not create too big a stir too quickly, or do they genuinely think 4 teams is best? Hmmmm.

As far as the #8 team getting hot, heck, I think that would be great. It's what makes playoffs the playoffs. I don't see folks complaining about it in the NFL, NBA, etc.

Good diary and good discussion.


May 16th, 2012 at 6:37 PM ^

Committee is the ONLY way to go.  I'm thinking of last year's situation where 11-1 Stanford was in the polls ahead of 11-2 Oregon.  Oregon CRUSHED Stanford 50-something to 30-something.  Oregon had a close loss to USC, whom Stanford barely beat in 3 OT's.  So on a head-to-head, clearly Oregon was the better team than Stanford.  Oregon's lower poll position was due to the fact they had the kajones to schedule LSU ... and get whupped ... whereas Stanford played 4 creampies OOC.

To me, only a Committee can correctly sort out that Oregon, not Stanford, was more deserving of the #4 bid.  And only a Committee might then say, no, neither deserves it -- Oregon deserves it over Stanford, but Oregon had their chance against LSU and got annihilated.  So only a Committee could say, we'd put Boise State or Wisconsin in the #4 spot, based on all those factors.  (To me, with Okie State and Bama in the #2/#3 slots, there was no other obvious #4 given the LSU/Oregon result early in the year; maybe I'm a homer but perhaps Wisconsin sort of proved themselves, although BSU's win over UGa definitely made BSU look good in that slot too.)

The polls failed in getting this right.  And no computer can account for all of those subjective decisions -- you'd never be able to program in every possible scenario.

Every other NCAA championship uses a committee, IIRC.  So there is no reason this should be any different.


May 17th, 2012 at 7:58 AM ^

But it wouldn't have allowed for his alternate possibilities of BSU or Wisky getting the bid as more deserving.

I agree that a selection committee is the way to go.  It works.  The eveidence comes from all the other NCAA tournaments.  Sure, it will cause controversy, but every system will generate controversy.


May 17th, 2012 at 9:09 AM ^

I agree on a committee, but for different reasons. The major success of the basketball tournament selection committee has been to encourage teams to schedule games against good teams in the non-conference portion of their schedules. Currently, poll voters have almost totally excluded schedule strength from their considerations (see Stanford/Oregon as an example). By clearly stating that SOS will be a consideration on par with win-loss record, the committee could shape non-conference scheduling for the better (I'm looking at you, Wisconsin).


May 16th, 2012 at 7:57 PM ^

that deemphasizes the current polls works for me.  If they want to blow up the current polling system, I'd be fine with poll driven selections, but until that happens, using them is pointless.  It doesn't matter HOW you use them, they are crap so the result from using them will be crap.


May 17th, 2012 at 2:52 PM ^

They need a committee to do the rankings every week, not just end of the season.  Start with the 8th game of the year or something and then you can have enough data points to compare the teams.  Basically re-do the BCS rankings but with a committee, they can feel free to use whatever data points necessary (Harris polls, RPI, computers, etc.) but the committee has to come up with the rankings.  That way you have true transparency.  Do they value Alabama's loss to #1 LSU as a worthy of a higher ranking than OSU's much better wins but much worse loss? 


Who did they play, where did they play them, results.  Strength of schedule has to be included but can't be such a huge factor that you get a team with 3 losses but a ridiculous strength of schedule a high ranking.  I think it needs to be a fine line, something like a 50% bonus points for a strong win, only a 1/4 of the normal decrease for a one strong loss (everything from the 2nd loss on is treated as full losses).