Skip to main content
Menu
The Week In College Football: Conferences Are Weird, So Our Predictions Are Too

Conferences are weird. And not just because some of their names don’t make any sense. (The Big Ten has 14 members?! The Big 12 has 10?!) Although most college football conferences hold championship games, others are prohibited from holding them (and get screwed because of it). They feature rapacious money grubbing and encourage teams to destroy rivalry games. No wonder independent Notre Dame disdains conference membership altogether.

More importantly, for our selfish interests, conferences cloud interpretations of FiveThirtyEight’s playoff predictions — which you can find here. The weird structures and dynamics of conferences sow lots of confusion. Since we launched our College Football Playoff model last week, we’ve received lots of reader questions, and many of them boil down to one of two “conference conundrums.” Elsewhere on FiveThirtyEight, Nate Silver has explained some tweaks to our model that should make the numbers more sensible. But in case you’re still scratching your head, let’s run through the two big reasons conferences can befuddle:

1. The Two-Team Conundrum: How can [Ohio State/Alabama/LSU/TCU] have higher odds of making the playoff than winning its conference?

Because some conferences have enough good teams that they could send two squads into the playoff. Alternatively, other considerations in the selection process might outweigh a good team’s loss in a conference championship.

For example: Our model gives No. 21 Alabama a 43 percent chance of being selected into the playoff but a 35 percent chance of winning the SEC. In a simulation that my colleague Jay Boice ran, a two-loss Alabama team that doesn’t get to play in the SEC championship game still makes the playoff 25.3 percent of the time (probably alongside a one-loss SEC champion). Similarly, No. 3 Ohio State has a 56 percent chance of making the postseason but a 44 percent chance of winning the Big Ten. If both Ohio State and Iowa remain undefeated heading into the Big Ten championship, there’s a good chance that both will make it in, regardless of the outcome.

Other cases are more complicated. What if an undefeated Ohio State team fell to a one-loss Iowa team in the Big Ten championship?2 It’s feasible to imagine almost any outcome in that case: The committee could go with Iowa, Ohio State, both teams or neither team.

In the conference previews below, I explain more about how those scenarios might play out. But know that there isn’t necessarily a one-to-one correspondence between conference championships and slots in the playoff; things could get messy. It’s the job of the model to sort all of that out — though, admittedly, we don’t know much (or anything, really) about how the committee weights conference championships, as it didn’t have to deal with any upsets in those title games during its inaugural season last year.

2. The Division Conundrum: For two teams in the same conference, how can one team have a better chance of winning the conference but another team a better shot at making the playoff?

We received lots of these questions, and the culprit is arbitrary conference divisions. Take Florida and Alabama, for example. The SEC has two divisions: East and West. The Gators have a 38 percent chance of winning the SEC, according to our model. That’s higher than Alabama’s odds! But Florida is given only a 17 percent chance of making the playoff, to the Tide’s 43 percent. That’s because the Gators play in (and have clinched) the SEC East, and ’Bama faces a tougher task in the other division, the SEC West (where LSU, Ole Miss and Mississippi State play). Florida, which has one loss, has an 80 percent chance of making the playoff if it wins out. But if Florida takes a second loss (say, at South Carolina on Saturday) but still wins the SEC championship, it might not. (That’s a nightmare scenario for the SEC, in which it might be denied representation in the playoff entirely.)

What-ifs of the week

Our College Football Playoff predictions have been updated with the rankings released Tuesday night, and we project Clemson, Ohio State and Alabama to make it, in agreement with the committee’s latest rankings. But we expect Baylor to have a better path than Notre Dame and to be the fourth team included.

But that’s what’s current. We’re already thinking about what’s next. Take a look at our “what-if table” below, which shows how our projected playoff odds would likely change if a team wins or loses its upcoming game. Also included is how likely we think it is that a team will win all its remaining games, and its chance to make the playoff if that happens.

CHANCE A TEAM WILL …
SCHOOL MAKE PLAYOFF MAKE PLAYOFF GIVEN WEEK 11 WIN MAKE PLAYOFF GIVEN WEEK 11 LOSS WIN OUT MAKE PLAYOFF GIVEN WINNING OUT
Clemson 67% 70% 37% 49% >99%
Ohio State 56 62 32 27 >99
Alabama 43 61 21 25 >99
Baylor 31 45 11 16 99
Notre Dame 30 31 2 38 74
Stanford 28 34 10 19 96
Oklahoma St. 23 26 7 16 98
Iowa 22 25 5 13 96
Oklahoma 17 38 3 18 87
Florida 17 23 3 18 80
LSU 12 16 <1 21 55
Utah 11 17 1 14 73
TCU 10 11 <1 21 49
Michigan St. 10 11 1 10 87
Michigan 7 9 <1 17 42
North Carolina 4 6 <1 9 46
Mississippi St. 3 6 <1 11 25
UCLA 3 3 <1 6 45
Houston 2 2 <1 22 7
USC 1 2 <1 31 4
Wisconsin <1 50 <1
Florida State <1 1 <1 28 2
Navy <1 <1 <1 11 2
Temple <1 <1 <1 18 1
Mississippi <1 21 <1
Memphis <1 <1 <1 12 1
Oregon <1 <1 <1 8 1
Northwestern <1 <1 <1 12 <1

Let’s use Iowa as our guinea pig. We currently give the Hawkeyes a 22 percent chance of making the playoff. But if they beat Minnesota on Saturday, those odds tick up to 25 percent; if they lose, they shrink to 5 percent. Although the rest of the Hawkeyes’ regular-season schedule isn’t that challenging, if they win out, they’ll face a tough matchup in the Big Ten title game. We give them a 13 percent chance to run the table. But if they do, they’re almost certainly in the playoff (96 percent).

What to watch for this week

Big 12
Game of the week: Baylor vs. Oklahoma

College football statheads: This is your game! While Oklahoma is No. 12 and Baylor is No. 6 according to the latest committee rankings, they are the No.1 and No. 2 squads according to ESPN’s Football Power Index (FPI), a computer-generated measure of team strength. Oklahoma is the best team in the country according to FPI, despite losing to lowly Texas. Baylor, on the other hand, is unsurprisingly high-rated by FPI because of its high-powered offense, which easily leads the nation at 57 points per game. Despite being No. 2 in FPI, the Bears are a 58 percent favorite to beat the Sooners because the game is at home in Waco, Texas. The game has big implications: If Oklahoma wins, its odds of making the playoff will rise to 38 percent from 17 percent; if Baylor does, the Bears will rocket from 31 percent to 45 percent likely to make the playoff.

ACC
Game of the week: Clemson vs. Syracuse

The ACC story hasn’t changed: Undefeated No. 1 Clemson is in the playoff if they win out, but it looks bleak for all other ACC teams (and for Clemson, should they lose). After last week’s victory over Florida State — in what was likely their last truly tough game — Clemson saw its playoff odds rise to the highest of any team (67 percent). But should the Tigers stumble, the ACC’s hopes fall off a cliff. Although one-loss North Carolina has almost wrapped up the ACC Coastal division, and has a 30 percent chance of winning the conference, its chance of making the playoff is a measly 4 percent — but if UNC wins out (and beats Clemson in the ACC title game), its chances rise to nearly 50/50.

Big Ten
Game of the week: Ohio State vs. Illinois

Even if the Buckeyes stumble, they won’t necessarily be out. Their odds are strongly affected by the conference conundrums I outlined earlier. A one-loss Ohio State team might not even get to play for the Big Ten title. This is why the undefeated Buckeyes have a 56 percent chance of making the playoff despite only a 44 percent chance of winning the Big Ten. A one-loss reigning national champion excluded for its conference championship game may still rate highly according to the committee.

SEC
Game of the week: Alabama vs. Mississippi State

Alabama crushed my beloved LSU Tigers, ending their hopes for an undefeated season. And, as a result, the Tide are now the No. 2 team in the latest College Football Rankings. ’Bama is not totally out should they lose again — either to Mississippi State on Saturday, or in the Iron Bowl against Auburn. In that scenario, a one-loss SEC champion could get into the playoff alongside the Tide, or the Tide could win the conference with two losses and still get in. But it’s not likely: In Jay Boice’s simulations, a two-loss Alabama team excluded from the SEC title game would be expected to make the playoff 25.3 percent of the time, and a two-loss Alabama that wins the championship game would make the playoff 34.3 percent of the time.

Pac-12
Game of the week: Stanford vs. Oregon

One-loss Stanford is the best Pac-12 bet to make the playoff, at 28 percent. The Cardinal are almost a sure thing if they win out (96 percent likely, according to the what-if table). But the threat to Stanford is that they’re the odd man out in a scenario with a one-loss SEC champion, an undefeated Clemson and an undefeated Big Ten or Big 12 champion.

Beyond The Power Five
Game of the week: Memphis vs. Houston

Notre Dame is looking good. On the heels of LSU’s loss to Alabama, the Irish have moved up to the No. 4 spot in the committee’s playoff rankings. But our model gives Baylor an ever-so-slight advantage over Notre Dame to make the playoff (31 percent vs. 30 percent).

But the real game to watch is Memphis vs. Houston. With Memphis’s crushing loss to Navy last week, its dream to be the mid-major that crashes the playoff party has ended. Houston, however, is still undefeated, and the two squads meet in the premier conference game of the season among non-Power Fives. To have any shot at the playoff, Houston will have to win out. But even if they do, their odds of making it are only 7 percent, by our estimation. So the Cougars need to keep praying for carnage among the elites.

CORRECTION (Nov. 11, 4:40 p.m.): An earlier version of the table in this article listed incorrect numbers for Wisconsin’s and Mississippi’s chances of winning out. Those odds are 50 percent and 21 percent, respectively. The table has been updated.

Footnotes

  1. All the rankings I’m using in this article are the committee’s.

  2. Iowa can afford to take a loss, maybe even two, since the Hawkeyes are a game ahead of Wisconsin and also have the tiebreaker over the Badgers.

Andrew Flowers writes about economics and sports for FiveThirtyEight.

Comments