Yesterday I said:
The example of Miami and Florida State (in 2000) shows that college football can’t rank every single team based on their record. Despite the complaints about how unfair it is that non-BCS teams have no shot at the national championship, no person in their right mind that’s not a Mountain West or WAC homer would put up with Utah-Boise State as the 2009 National Championship game. Members of BCS conferences would complain that they’re being punished for being in good conferences and the tendency to schedule cupcakes would get even worse. Ranking every team based on record, without regard to schedule, benefits the non-BCS conferences but it rarely actually selects the best teams that managed to escape good conferences. The system is biased against the non-BCS teams for a reason, people. A playoff is the only approach fair to both the BCS and non-BCS conferences.
But if we can’t seed teams based on their record, how do we seed teams? Under the framework Ed Gunther uses to frame the argument, we can’t use a BCS-like ranking system; it’s too subjective for our objective playoff. So what can we use? Gunther proposes the following strawman:
The anti-playoff side likes to frame the issue another way. If we created a playoff, like the pro-playoff fans want, but didn’t have the rankings, we’d need a way to choose which teams get to participate in the playoff. One of the most reliable ways would be to take the conference champions and a few wildcard teams, just like the NFL does. So here’s the NFL playoff laid overtop of college football: first off, all of the NFL’s divisions have the same access to the playoff and title game. So all eleven conferences (the college version of divisions) would all have to be equal and have the same access to the playoff. So the SunBelt is on equal footing with the SEC, the MAC with the Big10, etc. Sound good? Let’s keep going. To automatically get into the playoff, all you have to do is win your division/conference. So the champion of the SunBelt is in, while the second place team in the SEC might not be, depending on the wild card. Winning the MAC holds the same weight as winning the Big10 or Big12. Do we really need to go on? No.
Yes, we do, because the notion of selecting all 11 conference champions isn’t the insane strawman you seem to think it is. Winning the MAC might hold the same weight as the Big 10 or Big 12, but no one in their right mind thinks they’re making it to the championship game, negotiating their way through more than six BCS teams also littering the bracket, unless they have some mettle. What’s more, if the second-place team in the SEC doesn’t necessarily get in (which isn’t really terribly different from what exists now and what people want, 2008 Texas and 2006 Michigan notwithstanding), that means all the teams in the SEC have to give their all to get that one guaranteed bid to the playoff. (Psst! Importance of the regular season!)
A more appropriate comparison would be with the NCAA basketball tournament, which selects 31 conference champions and 34 at-large teams. In a sixteen-team playoff, selecting 11 conference champions would leave room for five at-large teams. Those at-large teams would likely all be BCS conference teams in any practical system, giving the BCS conferences 11 spots. (In simulated 11/5 systems based on the BCS standings last year, TCU picked up the last at-large bid. However, in my simulated system that used a committee-of-me, I plucked Georgia Tech ahead of TCU, Oklahoma State, and a dark horse bid by Pittsburgh.) Maybe the Sun Belt champion isn’t, strictly speaking, one of the top 16 teams in the country, and maybe they don’t strictly “deserve” to go to the playoff – maybe they’re significantly worse than any of the BCS conference teams in the playoff. So they’ll probably end up stashed at the bottom of the ladder, with the 15 or 16 seed, providing a relative cupcake for the 1 or 2 seed. The MWC and WAC champions are often very good teams, but the MAC, C-USA, and Sun Belt champions aren’t, so the 13 seed would be a relatively mediocre conference champion but one from a weak BCS conference or one of the better mid-majors. The top three seeds would play relative cupcakes in the first round, and once you got to four or lower, the intensity of the games ratchets up considerably. Teams at the top of the ladder would covet one of those top three seeds and an easy first-round game. (Psst! Importance of the regular season! No tanking down the stretch!)
Gunther’s point is that it’s unfair (or widely seen as such) to the best conferences to treat the BCS conferences and the non-BCS conferences completely equally, and thus some sort of subjective ranking system is needed to balance pro-BCS conference bias and anti-BCS conference bias (the latter of which is aka a faux golden mean). Thus even a playoff would need a BCS-type system to determine what teams were best over the course of the whole season. But here Gunther seems to have a rather broad definition of “ranking”. It’s true that the NCAA basketball selection committee creates a seed list of the 65 teams in its tournament to guide the seeding process, but it’s not a hard and fast rule. And that’s precisely what I use when creating my simulated 11/5 system at the end of the year, selecting the five at-larges and seeding the 16 teams in the tournament myself. The BCS is designed to select two teams, not 16 (especially when three of those sixteen are outside the top 25), based on polls (which are rooted in nothing but subjective opinions) and computer rankings (which are convoluted, often designed for gambling and not picking a champion, distrusted, and no one knows how they work anyway). The NCAA basketball selection process is carried out by a group of people who are given simple and reasonable computer numbers, such as the RPI and a simple strength of schedule formula, and other relevant facts. As long as the playoff selectors weren’t motivated to say “let’s include Notre Dame even though they don’t deserve to be in because they draw eyeballs and give them a higher seed than they deserve so they go deeper and pop more ratings”, the latter approach would be far superior for college football, possibly even if we were to stick with the BCS.
You’re not getting rid of controversy. There’ll just be controversy as to who gets in from the at-large pool. At least we won’t have any more undefeated teams with no chance of playing for a national championship. A five-team at-large pool is big enough that it should include any team with any legitimate claim to being the best team in college football, and by the point we get to the edge we’re talking about two, three, or even four-loss teams that probably don’t have a real shot at winning the whole thing anyway. (Which is why I’m not making it any bigger and including less worthy teams.) Does anyone really think that the teams on the bubble of the NCAA basketball tournament ever have any real shot at winning the national championship, George Mason notwithstanding? As we’ll see later, a 16-team playoff does a good job of including every team that, in past years, loudly proclaimed they were worthy of a shot at the national championship – even in the chaotic year of 2007.
I think Gunther’s fatal flaw is both implicit throughout his examination of people who oppose a playoff and made explicit in his introduction. Gunther seems to think (or at least consider a reasonable strawman) that if the season were objective all the way through, it would crown the team that was the absolute best: “Team A beats Team C, and Team B beats Team D, then Team A beats Team B = Team A is the best.” In other words, if the NFL had no regular season, if it just started right in with a 32-team tournament, the team it crowned would always be the best. Gunther doesn’t seem to consider that upsets can occur anywhere. What if the best team lost in the first round? What if a team got an upset, had lucky things happen to eliminate tough opponents before they got to them, and made it at least to the Super Bowl as a mediocre at best team? If Roger Federer loses in the first round, does that change the fact he’s the best player, or does it just mean that someone managed to get to him on that day? At least with the Florida Gators ranked #1 in the preseason we’ll know that if they fall out of the title picture it’s because they’re not as good a team as we thought. Ditto the Pittsburgh Steelers or New England Patriots and the NFL playoff picture. But if the team that everyone thought was best loses in the first round, how do we know that isn’t just because they had an off day? And don’t past rounds of the tournament become as meaningless as a regular season would be with each successive round? Isn’t it logical that if a regular season without a playoff (the BCS) doesn’t produce a clear-cut champion, a playoff without a regular season doesn’t produce any reading of the best team either?
(A quick irrelevant digression, that I couldn’t find any better place to put: there are, apparently, some opponents that would moan about a playoff producing rematches between teams that met in the regular season, rendering the original game between the teams irrelevant because this game is the one that counts. Didn’t the team that won the first matchup already prove they’re better? Gunther’s proponents would counter that that game didn’t really “prove” anything, since it’s one fallible game, but why should this game “prove” anything any more? They’re both individual fallible games. Most sane playoff proposals should be set up to avoid rematches at least in the early rounds.)
Devil’s advocate time: you can make a case that, because college football barely even gives you a hint as to what the absolute best team might be, the burden of that would have to fall on a playoff that isn’t well suited for that purpose, and Gunther’s strawman in the last paragraph would have more relevance than in other sports. Precisely because we mix up two definitions of who’s “best”, the winner of the playoff would be considered, indisputably, the best team in the country. Was Texas really better than USC in 2005, or just on the day they took the field? We’ll never know, but we take it for granted that they were, and with a playoff a team doesn’t even need to be ranked in the top two to end up being considered the undisputed best team in the country, regardless of whether they actually were. That’s why the NHL awards one trophy for having the best record in the regular season and another for winning the Stanley Cup Playoffs (although best-of-seven series make it less likely that a team will just get lucky). That brings me to what Gunther sees as the core of the argument.
A playoff won’t give us the best team at the end of the season, only the hottest or the one best able to avoid – or pull off – upsets. Under Gunther’s framework, the counterpoint of this is simple and seemingly self-evident: that the BCS doesn’t give us a single, clear-cut champion, just who a bunch of pollsters think should be the champion. But it’s not necessarily the case that a clear-cut champion is the sole province of a playoff, just as it isn’t necessarily the case that the best team can only be crowned by the BCS. Each system can agree with the other sometimes. Again, no one disputed that Texas was a deserving champion after the 2006 Rose Bowl; it might as well have come after the end of a long playoff. That’s because there wasn’t a “split title” where another team claimed they should be champions. Similarly, sometimes a team is so dominant that the championship game is just a coronation – consider the 1972 Miami Dolphins, or the 1985 Chicago Bears, or the Bulls teams of the 1990s, or even the 2009 UConn women’s basketball team. No one doubts that the best team won, because they were dominant in the regular season in a way no one else was. A matchup between two teams everyone thinks is the best two in the country isn’t any worse because it came at the end of a playoff – in fact the playoff itself may suffer for it because it just seems like a prelude to the main course. Ultimately, part of the reason no one ever follows through on their threats to leave the sport because of the injustices of the BCS is because if you have a matchup between two titans, it doesn’t matter how you got there.
So, to what extent does a playoff give us the best team, and to what extent does the BCS give us a single champion? Arguing under Gunther’s framework, proponents of a playoff would argue that the playoff consists of a lot of very good teams, and the winner of the playoff is obviously the best of them. Opponents would argue that the BCS standings reflect a consensus on what the best two teams are, which means it’s more clear-cut than it often receives credit for (again ignoring the mid-major bugaboo). Both of these arguments are patently false – a team that barely snuck into the playoffs could go all the way and win it all, and that wouldn’t mean they were the best team over the course of the season, and the latter argument is even more absurd, as anyone who thinks the BCS reflects a “consensus” hasn’t looked very closely at the numerous years of controversy, the corruption of the polls, and the fact computers can go against the consensus of humans. What’s more, I never hear the latter argument, since Gunther misconstrues where opponents are coming from, and the closest I hear to the former argument is in another context.
But still, at least the former argument hasn’t been completely demolished (yet). The 2007 Giants or 2008 Cardinals may not have been the best teams in the country, but they clearly must have been better than they got credit for (or that their record suggested) if they won three games against supposedly better teams, four in the case of the Giants. Where the argument goes wrong is that while it’s one thing for one team to pull off a string of upsets, it’s quite another if they’re the beneficiary of another team’s upsets. If the 5 seed in the NCAA Tournament makes the Final Four because the 1, 2, 3, and 4 seeds all fell before facing them, they haven’t really proven anything, other than that they can beat the 6, 7, 8, or even 9 seed. George Mason faced freakin’ Wichita State in the Sweet 16 the year they made their famous run.
Fortunately, Gunther proposes two more realistic – and common – arguments. Building on the brief success of the last argument, proponents note that usually, a playoff should crown the best team in the country, certainly more often than the BCS produces a champion everyone’s happy with. Gunther’s opponents counter that at least the BCS will never give the title to the team that in reality is eighth best (or even worse). Or does it? Think back to 2007, when there was a complete clusterbleep regarding who would face Ohio State in the national championship game. Any team in the top nine in the BCS standings could have conceivably been plucked to play the Buckeyes, and the only reason 12-0 Hawaii at #10 didn’t have a shot was because they were a non-BCS team. USC, which ended the season seventh, won the Rose Bowl, but crushed an Illinois team a lot of people didn’t think deserved to be in the BCS at all. USC had a mediocre resume with losses to Stanford and Oregon, but they could have easily gone to the BCS Championship Game if enough pollsters took pity on them, gotten lucky, and won. How do we know LSU, who won the title game that year, was really better than any of the eight teams below them, except maybe the ones they played? (Don’t try to muddle the issue by claiming it wouldn’t have happened under the old bowl system. What if USC had beaten Ohio State in the Rose Bowl?)
Technically it’s more accurate to claim, at least for Gunther’s vision of those that oppose a playoff, that the BCS will still give the title to a team that can claim to be the best team – but that’s in fact a minor concession on their part, since an eight-team playoff in 2007 would have achieved much the same goal, no matter how many upsets occurred – even if the eight seed won the title, they could conceivably claim to be best in the regular season as well. This brings us back to college football’s small sample size. What if I told you that in 2005, the best team in college football wasn’t Texas or USC, but West Virginia? The Mountaineers only had one loss on the entire season – they were just unlucky that day, just as Florida was unlucky last year when they lost to Ole Miss. And they did win their bowl game, topping the champions of the mighty SEC in Georgia. But they finished eleventh in the bowl-determining BCS standings, a seven seed in an eight-team tournament that granted BCS champion auto bids, and based on the BCS standings, a ten seed in a sixteen-team tournament that granted auto bids to all 11 conference champions.
This brings us back to the devil’s advocate position at the end of the last argument – college football’s regular season is so insufficient that almost any winner of a playoff could be considered the best in the regular season as well. At no point in the BCS era would an eight-team playoff based on the BCS standings have selected a non-conference champion with more than two losses. On the conference champion side, what if I told you that the best team in 2008 was Virginia Tech (who did win their bowl game, albeit against Cincinnati, an almost-as-weak Big East champion) and the best conference the ACC? You’d laugh until you looked at the clusterbleep of the ACC standings and saw that nearly every team had a shot to go to the ACC Championship Game. The SEC likes to claim they should get the benefit of the doubt for sometimes-weak records because every team in the SEC is so great that there’s so much parity that teams beat each other up; doesn’t that go double for the ACC? (The Hokies finished the regular season nineteenth in the BCS standings; under the same playoff formats as before, V-Tech would have been dead last in an eight-team playoff and 13th in a 16-team playoff.) You can’t claim the Giants were the best team in the NFL in 2007, and you can’t even claim the Cardinals were the best team in the NFC in 2008 despite the fact they won a division (well, you can, but it’s difficult); the NFL schedules are too balanced. You can make a case for any of thirteen (well, twelve) teams being the best in the country every single year, admittedly of varying levels of plausibility. (You hear that, Stewart Mandel?) A 16-team playoff would still select a smaller proportion of teams in FBS than any other playoff existing today. It may not be the ideal scenario for people who oppose a playoff for Gunther’s reasons, but that’s the way college football is and shall be.
(That Mandel link leads me to bring up an argument none of Gunther’s analysis brings up, which is the difficulty level to make a Cinderella run. The Cardinals would not happen in college football because they would not get the benefit of the doubt just for winning the division; winning a weak conference would not guarantee home field advantage in any round, as it did for the Cardinals, and locking up your conference early would not necessarily be an excuse to tank. The Giants had to win three tough road games against very good teams, though, just to make the Super Bowl against a fourth, and even in a sixteen-team format one of those games would probably be significantly easier; to beat 1, 2, and 3 seeds in the last three rounds of a sixteen-team format you would start out beating a 6 or 7 seed. A George Mason run might be of comparable difficulty in a sixteen-team format, and harder in a smaller format but with better teams to pull it off.)
The response Gunther’s opponents would have to the argument that a playoff should usually produce the best team is twofold, and in some sense, we covered them both earlier. In fact, the second response is precisely that it devalues the regular season. The first one:
A playoff breaks their definition of a best champion because first, teams will play different amounts of games in the season. With an 8-team playoff in college football, some teams would play 12 games and some would play 16 – that’s 33% more games, which is too big of a competitive gap to equally compare teams and their achievements.
But that’s pretty much immaterial, since ideally, the teams in the championship game have already established themselves as being on another level than the teams that aren’t in the playoffs, during the regular season. That’s why we need to make the playoff big enough to accommodate every team with a claim to be the best in college football. And teams within the playoff would play a varying number of games, with the teams in the championship game playing only one more game than the teams they beat in the semifinals. Given the way Gunther phrases this argument, and given the way Gunther’s opponents would presumably be okay with the bowl system where winning teams play one more game than losing ones, it seems to imply opponents would be okay with a gap that small. Wider gaps, like the one between the participants in the championship game and the quarterfinal losers, are more problematic; fortunately, my solution to the sanctity of the bowls solvdevalues es that problem, or at least widens it by one round. Stay tuned.
In fact, this argument itself suffers from two problems. First, it’s effectively saying the college football season is too small for a playoff. It’s too small for the regular season alone to give us sufficient data either; deal with it. The second one Gunther acknowledges, but not as a problem:
Basically, the anti-playoff side knows that their subjective champion is debatable, and the way they choose to make that debate fair is to make sure every team has the same amount of information (aka, number & type of games) available for voters to look at. If a few of the teams have more performances on a bigger stage, it makes the situation unfair even before voters begin the debate, the big no-no of the subjective side.
“But the whole point of a playoff is that it removes the subjectivity of a poll”, you say. But this is where we get into the problem Gunther’s opponents have with the notion of a playoff determining a claimant to the title of the best team with the same or similar veracity as the BCS, and the reason why I proposed that notion as a devil’s advocate argument, proposed by Gunther’s opponents rather than to them. It seems to me that Gunther’s opponents would accept the games in a playoff, but not necessarily the winner of the playoff as the automatic national champion. It’s as though the NHL’s Presidents Trophy were awarded to the team with the most combined points in the regular season and postseason. After all, the AP doesn’t always accept the winner of the BCS national championship game as its champion, because of the body of work their champion produced over the course of the season, and they still do a poll after the Final Four and don’t have to select the winner of the national championship if they don’t want to. Gunther’s opponents don’t want to separate the regular season and the playoffs, because they’re all still games played by the teams in question. Most people would call this a false position, a strawman inflicted on themselves, since the vast majority of people have no problem separating the regular season and the playoffs, and this is one reason I’m doubtful Gunther is properly reading the motives of opponents for anyone but himself.
But it’s harder to shake, and less of a strawman, when you consider that in college football, the playoff needs to play the role of helping determine the best team over the course of the whole season. The regular season is insufficient for that purpose, so any postseason, in the eyes of Gunther’s opponents, needs to be both a regular season and a playoff. We’ll see elements of attempts to resolve this contradiction later, when we take a look at some proposed playoff formats, including Gunther’s own suggestion. But this problem is rarely made explicit – the one attempt to resolve this contradiction other than Gunther himself is more concerned about the overall sanctity of the regular season – and so I don’t think it’s on top of mind for most opponents of a playoff. Besides which, the only real solution is to make the regular season itself longer – or institute a playoff so teams have more motivation to schedule tough to improve playoff seeding and chances of making the playoff. Whichever way you slice it, we’re not going to get better at determining the best team in college football without a playoff, and ultimately, people with the same ultimate motivations for opposing a playoff Gunther attributes to them might actually be better off with one.
And even if the team that would have been the best gets upended in an upset, well, we love upsets in March, don’t we?