2019 Death to the BCS Playoffs: Official Bracket

All right, the conference championships are in the books. Let's build a bracket.

I looked over the title games last week, and tried to picture how I anticipated the at large selection process would end up going. And so after a couple hours of deliberation, we're all set for the 2019 Death to the BCS Playoffs! But how do we take the results of so many college football games and pare this down to 16 teams? And why 16?

I ran with the idea that the authors of the book pictured here proposed almost a decade ago as the ultimate counterargument against the corrupt system that was the BCS. The current College Football Playoff is better, but not perfect in my eyes. There's still a level of corruption by the refusal to acknowledge all of the teams at the FBS level. This playoff fixes that by guaranteeing every conference champion a shot at the national title as every other level of NCAA football does. That's ten spots, and I fill out the field with six at large teams to make the numbers even. The best part? The first three rounds are all played at the home stadiums of the higher seeds; no more three-quarter empty bowl games in random places that rob athletic departments blind.

So how do I decide on the six at large teams who make it, and how do I seed the field? I try to be as impartial as I can, and I do that with a number of metrics outlined below and charted here:

• Playoff Points: This is an idea that I originally stole from the IHSA. They use a system of adding up the number of wins by opponents on a team's schedule and use that to try and determine a strength of schedule. This metric comes in three forms.
• First Degree Playoff Points (PP1): The closest to the original idea. For every opponent a team beats, they receive a number of points equal to that opponent's total wins on the season. Those numbers are added for every defeated opponent over the course of a season. As an example, Rutgers went 2-10 on the season with wins over Massachusetts (1-11) and Liberty (7-5), giving them eight PP1 for the year. Obviously, more wins means more chances for points, and wins over better teams means more points from each game. The caveat is that wins over FCS opponents do not award any Playoff Points.
• Second Degree Playoff Points (PP2): An extension of the original idea. PP1 answers the question to an extent of "How good were your opponents?" PP2 extends on that idea, by taking the mean of the PP1 of defeated opponents. Using the Rutgers example again, they beat Massachusetts, who finished with no PP1, and Liberty, who finished with 14. That gives Rutgers a PP2 of 7.00. More wins will generally mean more points, which I why I factor this metric as a mean rather than a sum total. This will result in some quirks; for example this year Miami (Florida) finished with a PP2 of 28.50 despite finishing 6-6; wins over teams like Virginia and Louisville helped bolster this number for them. Again, wins over FCS teams result in a score of zero for that win, which will drag the average down.
• Adjusted Playoff Points (aPP): Similar to PP1, except that for every loss by a team, I subtract the losses by their opponent from the total. This is my way to factor in losses as a metric, something not covered by either of the first two metrics. Obviously, losses to really good teams won't hurt, but losses to mediocre or bad teams will hurt worse. Going once again to the Rutgers example, their eight PP1 points still count, but are offset by each of their ten losses. The loss to undefeated Ohio State didn't cost them any points, but a Week 6 loss to 3-9 Maryland cost them nine points just from that game. All told, Rutgers finished with -33 aPP for the 2019 season. Again, wins over FCS teams don't award any points, but losses to FCS teams apply double the loss penalty. This year, there were three losses by FBS teams to visiting FCS opponents. As an example, Central Arkansas went 9-3 in their regular season (I don't count playoff games for this metric), including a win over Western Kentucky. Western Kentucky lost six aPP due to this loss.
• Computer Rankings: To try and make sure that I'm avoiding any bias, I work other metrics into the fray in the form of computer rankings. While this was a flaw of the BCS, part of the flaw was that margin of victory wasn't factored into theirs. I've found three sets of computer rankings that do, and I average them out to find one average ranking. This ranking is taken from the mean of Jeff Sagarin's rankings, the late David Rothman's rankings, and rankings from All My Sports Teams Suck.
• Results versus playoff teams: This is a fairly new metric kind of formalizing something I already look at to some degree. It's a metric I look at in stages; initially I start only with the ten conference champions as the playoff teams, and look at records for all the candidates at that time. As I add at large teams, I update the records accordingly before picking the next one. This finalizes when the 16 team field is set and I use it to help with seeding. As a general rule, losses aren't killer, and if you play playoff teams you're likely in better shape than if you don't, even if you lose. Wins over other playoff teams though will certainly help your resume.

I then try to organically combine all these factors when I build out the field. You will see that field seeded 1-16 after the jump, followed by my at large selection process and a rationale for how I seeded the field.

1. Ohio State (13-0, Big Ten Champion)- 84 PP1, 32.67 PP2, 84 aPP; 1.00 AVG CPU; 5-0 vs playoff teams (6th appearance, 1st since 2017)
2. LSU (13-0, SEC Champion)- 72 PP1, 27.08 PP2, 72 aPP; 2.33 AVG CPU; 2-0 vs playoff teams (2nd appearance, 2nd consecutive)
3. Clemson (13-0, ACC Champion)- 63 PP1, 22.50 PP2, 63 aPP; 2.67 AVG CPU; 0-0 vs playoff teams (5th appearance, 5th consecutive)
4. Georgia (11-2, At Large)- 69 PP1, 29.36 PP2, 61 aPP; 5.00 AVG CPU; 2-1 vs playoff teams (4th appearance, 3rd consecutive)
5. Oklahoma (12-1, Big 12 Champion)- 58 PP1, 23.09 PP2, 54 aPP; 5.67 AVG CPU; 0-0 vs playoff teams (7th appearance, 5th consecutive)
6. Wisconsin (10-3, At Large)- 64 PP1, 27.80 PP2, 58 aPP; 7.33 AVG CPU; 1-2 vs playoff teams (4th appearance, 1st since 2017)
7. Auburn (9-3, At Large)- 52 PP1, 22.56 PP2, 48 aPP; 9.00 AVG CPU; 1-2 vs playoff teams (3rd appearance, 1st since 2017)
8. Oregon (11-2, Pac 12 Champion)- 53 PP1, 24.00 PP2, 45 aPP; 11.33 AVG CPU; 1-1 vs playoff teams (3rd appearance, 1st since 2014)
9. Michigan (9-3, At Large)- 53 PP1, 24.78 PP2, 48 aPP; 11.00 AVG CPU; 1-2 vs playoff teams (3rd appearance, 2nd consecutive)
10. Notre Dame (10-2, At Large)- 61 PP1, 27.80 PP2, 56 aPP; 12.67 AVG CPU; 0-2 vs playoff teams (5th appearance, 3rd consecutive)
11. Memphis (12-1, American Athletic Champion)- 58 PP1, 22.55 PP2, 54 aPP; 13.33 AVG CPU; 0-0 vs playoff teams (2nd appearance, 1st since 2014)
12. Utah (11-2, At Large)- 57 PP1, 24.82 PP2, 53 aPP; 15.33 AVG CPU; 0-1 vs playoff teams (1st appearance)
13. Boise State (12-1, Mountain West Champion)- 62 PP1, 23.45 PP2, 57 aPP; 20.67 AVG CPU; 0-0 vs playoff teams (5th appearance, 3rd consecutive)
14. Appalachian State (12-1, Sun Belt Champion)- 54 PP1, 17.91 PP2, 53 aPP; 22.67 AVG CPU; 0-0 vs playoff teams (2nd appearance, 2nd consecutive)
15. Florida Atlantic (10-3, Conference USA Champion)- 42 PP1, 16.33 PP2, 35 aPP; 43.33 AVG CPU; 0-1 vs playoff teams (2nd appearance, 1st since 2017)
16. Miami (Ohio) (8-5, MAC Champion)- 27 PP1, 14.00 PP2, 9 aPP; 82.33 AVG CPU; 0-1 vs playoff teams (1st appearance)

To view the official bracket, click here.

Before I get into the seeding, here's how I picked the six at large teams competing this season.

1. Cincinnati, Virginia, Wisconsin, Oklahoma State, UAB, Notre Dame, Western Michigan, Air Force, Utah, GEORGIA, Louisiana
   Georgia is clearly the best team at the table; they have the most PP1 of anyone alive, the second-most PP2 in the country, and despite that bad South Carolina loss before falling to LSU, have the best aPP at the table as well. The computers back this up. Easy first pick. Auburn steps to the table out of the SEC.
2. Cincinnati, Virginia, WISCONSIN, Oklahoma State, UAB, Notre Dame, Western Michigan, Air Force, Utah, Auburn, Louisiana
   Another fairly easy choice for best team at the table. Wisconsin has the best Playoff Point numbers of anyone in the discussion, and the second-best computer rankings behind Alabama of all remaining teams. In a crowded conference with Michigan, Iowa, and Minnesota all in consideration, despite my projection last week, Michigan steps to the table out of the Big Ten.
3. Cincinnati, Virginia, Michigan, Oklahoma State, UAB, Notre Dame, Western Michigan, Air Force, Utah, AUBURN, Louisiana
   The decisions are getting a little tougher in terms of who to take here. I considered Notre Dame in this spot, given that they have the best Playoff Points numbers of the teams at the table, but they're behind Auburn in the computer rankings, and the Tigers have a win over a playoff team in Oregon. Two of Auburn's three losses are to playoff teams, and the third is to a Florida team that would be in consideration if they weren't suspended for this postseason due to playing two FCS teams. The Tigers get the nod, which is a great sign for Alabama, who steps to the table out of the SEC.
4. Cincinnati, Virginia, MICHIGAN, Oklahoma State, UAB, Notre Dame, Western Michigan, Air Force, Utah, Alabama, Louisiana
   For a long shot a week ago, Michigan getting in relatively comfortably has to be something of an upset. The Wolverines' Playoff Point numbers were among the higher ones at the table, their computer rankings were the best out of everyone in the discussion except Alabama (but their Playoff Point numbers are really low), two of Michigan's three losses are to playoff teams, and their best win is over a Notre Dame team still at the table in convincing fashion. It's just enough for me. Iowa steps to the table next out of the Big Ten.
5. Cincinnati, Virginia, Iowa, Oklahoma State, UAB, NOTRE DAME, Western Michigan, Air Force, Utah, Alabama, Louisiana
   At this point the Irish have fallen far enough; they're the best team left on the board with two losses, both to at large teams picked ahead of them. They have the most Playoff Points of any team left at the table and the best computer rankings other than Alabama. But Bama's way behind in Playoff Points, and Notre Dame's best win (by 32 over a Navy team not at the table but was certainly on the list to get to the table if not blocked by Cincinnati) is way better than Bama's best win (by 24 over a South Carolina team that beat Georgia? By 19 over Texas A&M? The pickings are slim). With Notre Dame in, BYU steps to the table next as an independent.
6. Cincinnati, Virginia, Iowa, Oklahoma State, UAB, BYU, Western Michigan, Air Force, UTAH, Alabama, Louisiana
   Anybody with four or more losses is out, which eliminates Oklahoma State, Virginia, Western Michigan, BYU, and UAB. Louisiana easily has the weakest resume of the six remaining, so they're out. Air Force is the next one out, even though their Playoff Point numbers are better than Alabama; the Tide are better according to the computers, and both of their losses were to playoff teams; the Falcons have one, plus a second to a Navy team out of the mix at this point. Iowa's win over Minnesota now means nothing, and at 9-3, they're gone. At this point, I'm left with three options: Alabama, with the best computer rankings but relatively terrible Playoff Point numbers; Utah, with the most Playoff Points on the board and middle computer rankings of the remaining teams, but no real signature wins; and Cincinnati, who has the only playoff win on the board (albeit against the 16 seed), all three losses were to playoff teams (including Memphis twice), and their "best" win is probably over a UCF team that's no longer in the conversation. At the end of the day, I opted for the two loss conference runner-up over a three loss Cincy team and an Alabama team that also didn't beat anybody. My first team out and toughest cut? Alabama.

Photo by AJ Mast (AP)

So in terms of seeding, we had just three undefeated teams this year, all of whom got top three seeds. However, unlike the College Football Playoff, I gave Ohio State the nod for the top overall seed. Every metric pointed in their direction: they had a record 84 PP1 and aPP, were the only team with more than 30 PP2, were the unanimous number one in the computer rankings, and had five of their 13 wins come against playoff teams. This wasn't close, so LSU will have to settle for the #2 seed. They were miles ahead of Clemson on the Playoff Point front and a hair in front in terms of computer rankings, while having some results against playoff teams. The lack of those, combined with the weakest resume of the unbeatens, results in Clemson having to settle for the #3 seed and, if chalk holds, having to go on the road for the semis. I went against the CFP committee in another place as well, giving Georgia the nod over Oklahoma for the #4 seed and the other guaranteed second home game. Oklahoma had an extra win, as well as arguably the better loss, but Georgia's metrics were better across the board.

Photo by Matthew Emmons (USA TODAY Sports)

The Sooners will have to settle for just the first round home game barring an upset as the best remaining team in the field. With a number of auto bid teams still on the table, I opted to start going with at large opponents from here, as Wisconsin's high Playoff Point totals and strong computer rankings, plus a win over a playoff opponent, was enough to get them the #6 seed. Despite three losses all in the regular season, Auburn came in next at #7 with two top ten computer rankings and an identical playoff record to the Badgers. I slotted my last Power Five champion in Oregon into the last first round home game slot as one of the few remaining two-loss teams in the field.

Photo by David Guralnick (Detroit News)

Michigan will travel out west to play them in the opening round as my last team with a playoff win on their resume. That head to head win over Notre Dame got them the nod for the #9 seed, while the Irish slot into #10 with the second-best Playoff Point numbers left on the board. With three one-loss conference champions left, I had to put one of them here next, and opted for Memphis; despite having fewer Playoff Points than Boise State, they were better according to the computers and had the "better" loss in a tougher conference, so they slot into #11. My last at large comes next in Utah; the computers had them higher than any of the remaining conference champs, and that, combined with a playoff result, was enough to give Utah the #12 seed.

Photo by Steve Conner (AP)

As per usual, the bottom four are Group of Five conference champs, but a pair of 12-1 teams get really tough first round games here. Boise State, despite having more PP1 than a majority of teams in the field, with lower computer rankings and the lack of a playoff result slot into #13. Their numbers are a little better than Appalachian State, who slot into #14. Florida Atlantic is then the clear #15 seed, and even though it means a first round rematch, 8-5 is awful, so Miami (Ohio) has no choice but to play the Buckeyes again.

I feel like this was one of the tougher brackets to build out in the years that I've done this, but I think I did a fair and balanced job of it. Next week I'm going to follow up on the yet-to-be-simulated 2018 Death to the BCS Playoffs, and hopefully the simulator will be ready to go for the 2019 Playoffs when the first round is slated to start on the 21st! Good luck to all the teams in the field!