Fixing The College Football Postseason
It is commonly said that college football has the best regular season followed by the worst post-season. These are perhaps not unrelated phenomenon, as the lack of a robust playoff structure (bad post-season) at the end of the season make the regular season more intense than anything any other sport has to offer (best regular season). For the longest time, it was the only major sport at any level without any sort of playoff system, wherein the champions were decided by news reporters and polling coaches. Eventually there was another round added to create a four-team playoff, which people hoped would improve the post-season by reducing controversy over who the national championship, while leaving the rest intact with an important regular season and preserving the bowl system. It succeeded only in the reducing controversy, went out of its way to fail at maintaining the importance of the regular season, and in combination with other factors completely failed at the third part1.
The good news is that year in and year out, we have a national champion and comparatively little controversy over it. The devaluation of the regular season, though, and the collapse of bowl relevance, essentially leave a system with one happy team (and sometimes a conference of teams that dramatically overestimate the value of association) and the rest having failed at the only thing that matters. It has produced some boring games and also lead to stagnant television ratings and fan interest in the championship game, the viewership of which has not moved since the four-team tournament was implemented.
In my view bowl system is actually underrated, preferable not only the current playoff but fun in its own right, but that ship has sailed. As soon as ESPN got the contract for the playoff, they geared all of their promotions towards it. “Who’s in?” they asked. Age-old rivalries were presented in the context of playoff even when neither team had a chance of making it in because of what effect their rivalry game would have on strength of scheduling of the teams in championship contention that actually mattered). Meanwhile, the NCAA changed a bunch of rules that pushed players and teams to make decisions more quickly, including the hiring and firing of coaches, players announcing their intent to transfer or enter the draft, and so on. These changes originally occurred before the bowl games, and are increasingly taking place during the regular season itself (among teams outside of championship contention, of course). None of these changes can be undone, and some were inevitable with or without a playoff.
Even though I was against the creation of a playoff, I believe at this point the only way forward is to lean into it. The main argument against doing so – apart from bowls – is the importance of the regular season. As it happens, though, the four-team playoff destroyed that as well. With the myopic focus of “Who’s in?” meant that the playoff criteria would be the most important thing, and the playoff criteria attached little importance to a lot of regular season achievements. Losing games didn’t matter if the playoff committee decided it didn’t reflect on the quality of your team. Conference championships are neither necessary nor sufficient to make the tournament. The result is that most years only three conference champions are included, and even among the Power 5 conferences the Pac-12 rarely makes the playoffs and the Big 12 and ACC only do so inconsistently2.
Last year, a proposal was put forth by an exploratory committee for a 12-team playoff. This turned out to be a surprisingly good number that allow the system to incorporate conference championship, allow a G5 team in, provide enough at-larges to keep the strongest conferences happy, without overrunning the tournament with mediocre teams on their fifth mulligan. Eight meant tough decisions regarding conference champions and at-larges, while sixteen lets in an awful lot of teams with weak claims to a shot at the title given that they were not especially competitive for even their conference championships. We also have experience with 12 teams because that’s how many teams make the top tiers of bowls, using roughly the same criterial proposed last summer. Unfortunately, discussions broke down recently over what amounts to relatively small details over which teams3 are included and a few other bits. One example of the latter is whether games are played on campus stadiums and what place bowl games have in the system, most specifically the Rose Bowl and its traditional New Years Day scheduling and Pac-12 vs Big Ten matchup.
On the one hand, it wouldn’t be that hard to incorporate bowl games. Just put most of the games at a neutral site, slap the bowl name on them, and go! On the other hand, bowls are among the things that have made college sports unique, and the history is pretty intense. To be honest, an Orange Bowl between Texas A&M and Wake Forest is really just kind of sad and a perfectly seeded playoff system with home field advantage mirroring the NFL kind of sterile and inappropriate for a league that is different in many important ways. I believe, however, that we can thread the needle of tradition and progress here.
The idea originally came from a proposal on Twitter that I liked a lot, but believe needs a few tweaks.
11-Team CFP proposal. Final Revision. P5, G4, Bowls, NYD tradition, expanded access, expanded $. Everyone wins. Please share! pic.twitter.com/ItdXQGDmkH
— Joel Cost (@drtoothsleuth) November 5, 2021
The basic gist is that the bowl games are incorporated as semi-finals with teams assigned to them on the basis of historical associations. So, the Big Ten and Pac-12 go to the Rose Bowl, the Big 12 goes to the Cotton Bowl, ACC to the Orange, and SEC to the Sugar. With the exception of the Rose Bowl, the remaining slots are filled by survivors of the at-large/G5 round. This proposal only has 11 teams, and the top five conference champions get an automatic bye (assumed in the chart to be the P5, but a P5 champ is replaced by G5 team if they are higher ranked).
I like the concept, but would make some tweaks with the execution. What I like about this plan is that it gives each conference a destination. I was raised in Southwest Conference territory and making the Cotton Bowl was a big deal in that conference. From there my alma mater (University of Houston) went to Conference USA which had a similar arrangement with the Liberty Bowl. Right now, as members of the American Athletic Conference aim to get into a major bowl, but which bowl changes every year. In his plan the goal would most likely be the Cotton Bowl, if you make it past round one, and that’s an improvement over the status quo.
Nonetheless, I would make a few changes. The first is that I would move the Rose Bowl and Sugar Bowl4 to the semi-finals. Unlike in the Gross plan, that would mean that it is not always the Pac-12 vs Big Ten, but each year it would have the opportunity to be if those conferences advance. Most years they would at least have one or the other, I think. For the Rose Bowl, it’s a good mix I believe the Rose Bowl would probably prefer that to be a quarter-final between those two. It also prevents the playoff from being pushed too far into January.
I’m not too hung up on specifics of conferences and bowls (see previous footnote). The important thing is that each conference gets its own quarterfinal bowl and if they are one of the top four conference champions, their bowl gets to host the semi-final. So, if you’re the Pac-12, you’re aiming for the Fiesta Bowl. But if your champion is ranked #5 or worse among conference champions, you’ll be going to someone else’s quarterfinal bowl. Tentatively, I gave the Big 12 to the Cotton Bowl, the Big Ten the Orange Bowl, the SEC the Peach Bowl, and the ACC the “Emerald Bowl”5, and the American Athletic Conference the Liberty Bowl6.
For the most part, the teams would be sorted by seeding with the following preferences and alterations:
- Teams that played one another during the regular season should be scheduled away from one another in the first two rounds of the playoffs where possible without moving teams more than one place in the seeding or flipping home field advantage.
- The Rose Bowl bracket always includes the Big Ten and Pac-12 champions, in opposing sub-brackets.
- The Sugar Bowl bracket always includes the SEC champion.
I’m not sure whether you do this formally through computers or have a selection committee with the understood preferences. Either works, though I think the latter is better7. If you have two committees they can adapt so specific circumstances and decide the tradeoffs. For example, if the best teams in the country far and away are the Big Ten and Pac-12, they might decide to shift one of them over to the Sugar Bowl bracket. Also, it might become apparent that you do want rematches in the first round rather taking a chance on one sucking up the semi-finals. Or they might decide not to worry about rematches at all until they have an idea of how it plays out viewer-wise. The important thing is that the higher ranked teams get the bye and home field advantage.
It may seem wrong to you to fiddle with the seeding, but even in the best of circumstances seeding is more art than science. With the CFP, in the last eight seasons the 4-seed beat the 1-seed 22% of the time, and the 3-seed beat the 2-seed 44% of the time. It’s simply not like the NFL where you have the sort of balanced scheduling where you can look at records and seed accordingly. In addition to tradition, this is also why I like neutral site games after the first round (even if there is a “home bowl” aspect. Seeding already matters a lot by giving the four teams a bye. Giving them a bye plus home field advantage (even if you split ticket availability) is unnecessary.
I ran the simulation in my mind for the last four complete seasons (excluding 2020) and I was able to meet all of the above-mentioned sorting preferences without moving a team more than one place form their natural committee seed, and not changing byes or home field advantages. There were a lot of potential matchups that would have been exciting. College football is great.
This runs a slight risk of being “too complex” but I think even if people don’t understand it they will be more patient when it comes to seeding and matchups than inclusion/exclusion, which was the problem with the complexity of the BCS. What it loses in intuitive transparency, it gains in giving people better matchups. It’s possible this year will be different, but one of the worst-rated championship games in the modern era was the LSU-Bama rematch and my guess is that interest will be limited in playoff rematches compared to new matchups. The committee probably would have changed the seeding this year to avoid a Bama-Georgia rematch in the first round, they just would have done so quietly. I would just be transparent with it. The people want good matchups! That’s usually the purpose of seeding to begin with8, it’s just in college football you have more variables.
The last change I would make may be one of the more controversial, but since I am putting my plan out there I feel obliged. In my view, every team in the FBS should have a path to the championship that does not rely on others losing. While some make demands that every conference champion should be included (as is the case at every other level of college football) but to be honest I’m not interested in making room for a 9-4 Northern Illinois to get a shot at the tournament. So, my rule would be that any conference champion that wins at least eight games in their conference and 12 games against FBS teams should be included in the playoffs. That means they get one loss against an FBS team or one FCS opponent, but not both. The only time you’re likely to see this make a difference is when you have two G5 conference champions meet the criteria, which happens but is rare. You also may be worried about the edge it gives teams in the bottom conferences, but going near undefeated is hard and rarely with a team in one of the lowest two or three conferences that does it9 I would probably have to take the “L” on this one, though, until or unless they expand to 14 or 16.
That just leaves the bowl games, and I have a lot of thoughts on the bowl season that I will save for another time. But as it pertains to the playoffs, whichever major bowl gets left out because their conference champion was not in the top four, can get the premier bowl game outside of the playoffs with the top two teams that haven’t already played one another. So, in 2019 the Fiesta Bowl was left out of the playoffs because of the Pac-12 champion being ranked fifth among conference champions, and the eligible teams were Michigan and Auburn. In 2021, since there were two P5 conferences that didn’t host playoff bowls (Pac-12 and ACC) you would have the top two and the next two after that. It would make for some good games and maybe there would be motivation in the form of two teams playing another both making a case that they should have been included.
I’m the first to admit that this is mostly fan-boy stuff and is unlikely to come into fruition. Even so, this is a plan that would weave together the tradition of bowls and have a more robust playoff system that everyone can play towards.
- There were of course other goals for having a playoff, such as producing more good games and making lots of money. It succeeded at the latter, but not at the former. The ones I list here are for the four-team playoff in particular, as opposed to eight, twelve, or sixteen.
- The Big 12 has only ever sent Oklahoma, who is leaving the conference for the SEC in the next couple of years. Apart from them, only TCU has come close. Future Big 12 member Cincinnati made it this year. The ACC sent Florida State one year but after that only Clemson has gone though they had a streak over multiple years. The Pac-12 hasn’t sent anyone since 2016.
- The competing proposals are virtually identical in terms of which teams would be invited, but differ slightly in how they are determined. It’s mostly a symbolic debate over the distinction between the power conferences and the others, but one which could have more concrete ramifications down the line in terms of TV contracts and NCAA structure.
- I decided that of the remaining bowls, the Sugar Bowl was the most prestigious, so I gave it the other semi-final. It could be the Orange or Peach or whatever. Throughout this piece I will be making some specific decisions because I need to in order to explain my plan, but I am attached to very few of them. So, if you think it’s obvious that the counterpart to the Rose Bowl should be a Peach Bowl seeded for a SEC vs ACC showdown, I’m not arguing with you on that provided that the Big 12 and G5 conferences have an in.
- One rule I would have is that all of these bowls need real names. The same way they made the Peach Bowl become the Peach Bowl again, I’d make the Duke Mayo Bowl – or whatever bowl they choose – take a real name. I chose Emerald because I am eyeing the former Belk Bowl in Charlotte and emeralds are North Carolina’s state mineral. There was a bowl in San Francisco by that name, but I don’t think that’s an issue here as bowl names have been recycled before.
- I had to figure one out for the AAC because Cincinnati would have hosted the semi-final round. Presumably all of the G5 conferences will get their own bowl but I don’t need the others even tentatively.
- If you want a nitty-gritty idea of how it might work structurally, there would be a Rose Bowl Bracket Committee comprised of representatives of half of the conferences including the Big Ten and Pac-12, and a Sugar Bowl Bracket Committee with representatives from the SEC and four other conferences. Each Committee has a maximum of two conferences they can insist be in their bracket – for the Rose Bowl it would obviously be Pac-12 and Big Ten and for the Sugar Bowl it would be the SEC possibly with another – but after that they pick the teams they want. Each Bracket then picks which of the top four conference champions they want from those still available, then they pick among the four teams designated to have home field advantage in round one, and then they pick from the four teams that travel in week one. The two groups can consult. As I mention elsewhere, “just do it by seeding” assumes too much of the accuracy of seeds. So, if this seems arbitrary to you, I have bad news for you about the seeding and indeed the entire selection process of systems past, present, and future. Most of the time, the brackets are likely to pick the next highest seeded team. Bowls and brackets get equal financial shares so there is not an incentive to choose by brand name
- Well, in particular the purpose of seeding is to save the best matches for the end. So, you put the barely-ins against the top seeds. Since seeding in this case is really hard to do, I think it matters less.
- Every G5 team that would qualify has been from one of the two strongest G5 conference with only two exceptions, from the Mid-American Conference (2013 Northern Illinois, 2017 Western Michigan). In both of those cases there was no other dominant G5 team so they would have taken the sixth slot. There was a third potential case in 2020 with Coastal Carolina if not for Covid, but even that one wouldn’t have created a seventh spot because the Pac-12 champion was ranked below both (it was a weird year and not a good year to design playoffs for).