The Allstate Playoff Predictor is running scenarios for the final time in the 2021 college football season.
With just conference championships ahead, there aren’t that many permutations of game outcomes remaining. But the College Football Playoff selection? That can still go a number of different ways.
We’ll use our trusty model — which is based on the committee’s past behavior and factors in FPI, strength of record, number of losses, conference championships (or independent status) and the current CFP rankings — to forecast the selection given a particular outcome. As always, remember that the Allstate Playoff Predictor deals in probability, not absolutes: After all, it’s predicting what humans in a conference room in Texas will decide.
Lastly, note that final predictions could come down to how these conference championship games are played. Alabama losing by a field goal and Alabama losing by three touchdowns could mean very different things. So our projections for a particular scenario could shift between now and Sunday morning depending on how these games play out.
Georgia, Michigan, Oklahoma State and Cincinnati win
Likelihood of occurring: 27%
Georgia and Michigan are locks to get in, and the model has Cincinnati not far behind at 88%. The real question is: Who gets the last spot between Oklahoma State, Notre Dame and Alabama? The Predictor leans reasonably strongly toward Oklahoma State but gives all three teams a shot.
It puts the Cowboys’, Fighting Irish’s and Crimson Tide’s chances at 64%, 29% and 19%, respectively (note that this adds up to over 100% because it still gives a non-zero shot that the committee punts Cincinnati out of the top four).
Why Notre Dame over Alabama? Two reasons. First, the literal number in the loss column has previously had an outsized effect on the committee above and beyond strength of record — and Alabama projects to finish fifth in strength of record anyway in this scenario. Second, while neither team is a conference champion, Notre Dame, as an independent, had no opportunity to be one. That, previously, has given the Fighting Irish an edge over non-champions in a conference. All that being said: It’s close, because Alabama is clearly one of the four best teams (the Crimson Tide rank third in FPI with a big gap between them and the No. 4 team, Michigan) in college football.
Georgia, Michigan, Oklahoma State and Houston win
Likelihood of occurring: 6%
In the previous scenario we more or less were looking at three teams vying for one spot. With Cincinnati out of the picture, it becomes those same three teams vying for two spots. The Predictor leans toward Oklahoma State and Notre Dame over the two-loss non-champ in Alabama for aforementioned reasons, but the Crimson Tide would still be almost a coin flip to get in. The model would give Oklahoma State a 79% chance at a berth, Notre Dame a 70% chance and Alabama a 50% chance.
Note that the model does not know about Brian Kelly’s departure, which Gary Barta said is something the committee will consider (which is absurd, but another story altogether).
Georgia, Michigan, Baylor and Cincinnati win
Likelihood of occurring: 15%
It comes down to Notre Dame vs. Alabama here. And for the same reasons mentioned above, the Predictor gives Notre Dame a slight edge but thinks the committee really could go either way: 61% for Notre Dame, 42% for Alabama. Would be one of the most interesting head-to-head comparisons we’d have seen on Selection Day.
Georgia, Iowa, Oklahoma State and Cincinnati win
Likelihood of occurring: 8%
Uncertainty across the board here, but the model would rank these teams in terms of chance to reach the playoff as: Georgia, Cincinnati, Oklahoma State, Notre Dame, Alabama. Any combination of Georgia plus three is possible, though.
Alabama, Michigan, Oklahoma State and Cincinnati win
Likelihood of occurring: 14%
There’s three locks here in Alabama, Georgia and Michigan. It all comes down to the final spot, and the Predictor leans toward Cincinnati landing it, in no small part due to the fact that the committee already has the Bearcats in the top four. Oklahoma State would have to leapfrog the Bearcats despite both winning conference championships.
Georgia, Michigan, Baylor and Houston win
Likelihood of occurring: 3%
This sets up a snoozefest of a selection as Georgia, Michigan, Alabama and Notre Dame walk into the playoff.
Georgia, Iowa, Oklahoma State and Houston win
Likelihood of occurring: 2%
Good chance it’s Georgia, Alabama, Oklahoma State and Notre Dame — and that’s what the Predictor says is the most likely foursome — though the model still gives a shot to Michigan even after a loss because the Wolverines are currently at No. 2.
Georgia, Iowa, Baylor and Cincinnati win
Likelihood of occurring: 5%
Same as above, just replace Oklahoma State with Cincinnati.
Alabama, Michigan, Oklahoma State and Houston win
Likelihood of occurring: 3%
Ooooh, this is razor-thin. Alabama, Georgia and Michigan are all in. But we’ve got Oklahoma State barely edging out Notre Dame here, 53% to 49%.
Alabama, Michigan, Baylor and Cincinnati win
Likelihood of occurring: 8%
This sets up what should be a pretty straightforward selection of Alabama, Michigan, Georgia and Cincinnati. The Predictor won’t totally rule out the committee finding a way to put Notre Dame in over the Bearcats, but it’d be a stretch logically, even for the committee, considering those two schools’ current ranks.
Alabama, Iowa, Oklahoma State and Cincinnati win
Likelihood of occurring: 4%
Georgia, Alabama and Cincinnati are all very likely in or better, so this one comes down to Oklahoma State vs. Notre Dame. The model favors Oklahoma State 57% to 39% in this case. Why the larger spread than the previous Cowboys vs. Fighting Irish scenario? I think that’s just simulation variance, which happens with projections.
Alabama, Iowa, Baylor and Cincinnati win
Likelihood of occurring: 2%
Alabama, Georgia and Cincinnati are in. Notre Dame would likely grab the last spot, but the Predictor isn’t willing to totally throw in the towel on Michigan. There’s a chance the Wolverines could potentially hold off the Fighting Irish for that last spot, because Michigan could drop two spots and still be in the playoff.
Alabama, Michigan, Baylor and Houston win
Likelihood of occurring: 2%
Alabama, Georgia, Michigan, Notre Dame.
Alabama, Iowa, Oklahoma State and Houston win
Likelihood of occurring: 1%
It likely would be Georgia, Alabama, Oklahoma State and Notre Dame. Though, again, the Predictor won’t rule out Michigan. It’s also too unlikely for the model to produce specific projections.
Georgia, Iowa, Baylor and Houston win
Likelihood of occurring: 1%
This, too, is so unlikely that the model won’t output individual team numbers, but Georgia, Alabama and Notre Dame should all be in. The question is: Who is the fourth team? The Predictor thinks it’s Michigan, because the Wolverines are currently ranked at No. 2 and might be able to hold on. But the door is really wide open if the committee wants to get crazy.
Could this be how Baylor slips in? Oregon? Ohio State? Iowa? The Predictor doesn’t think so (if it had to pick among those teams it would guess the Buckeyes, who are clearly the best of the group in terms of FPI ranking).
Alabama, Iowa, Baylor and Houston win
Likelihood of occurring: 1%
Almost the exact same as the scenario above, because once all the other favorites fall it likely won’t matter who wins the SEC championship game.