Now it’s political. Early in the season, playoff forecasts depend almost entirely on the quality of the teams and the difficulty of their schedules. But as the field is narrowed to just a handful of serious contenders, the wisdom and whims of the playoff committee loom ever larger. Will conference titles outweigh an early season loss? Will runaway wins, often forsaken as bad for the game, tilt the seedings? And is there any consideration to spreading the love around rather than lavishing it disproportionately on a single conference?
To understand how the playoff committee will answer these questions we have mined its history. Although only four years old, the committee produces a Top 25 for six weeks every year, so there is more data there than it seems. We’ve investigated a wide range of variables to see what best makes sense of the committee’s behavior. They are human, and the committee’s membership changes, so it’s not a perfect science. But we get good traction using inputs such as number of losses, strength of schedule, strength of record and overall team quality.
This year’s first rankings come out Tuesday night. We don’t expect a lot of surprises at the top, but outside the top few teams, there’s quite a bit of uncertainty. Below is our committee NOW-cast — our empirically based forecast of the committee’s rankings this week. The top teams break into four distinct tiers, with a lot of uncertainty within each tier.
Clemson (NOW-cast No. 1) and Alabama (No. 2) are the clear leaders of the pack. While neither has played as tough a schedule as Notre Dame, both have distinguished themselves in their wins. While Alabama is the better team, Clemson has played the tougher schedule, and has a better strength of record — the likelihood that an average Top 25 team playing their schedule would have at least as many wins.
We think the gap between these two — both in the eyes of the committee now, and in our predictive ratings — is much smaller than most. In fact, after Clemson’s 59-10 thrashing of Florida State (our No. 1 performance of Week 9), Clemson has closed the gap with Alabama to less than two points in our predictive ratings.
Notre Dame (No. 3) will get the respect of the committee by virtue of its 8-0 record, but dive deeper and the numbers don’t look kindly on the Irish. Our in-season only numbers have Notre Dame as only the No. 25 team in the country — and we have the Irish as only No. 10 on our predictive ratings. They’ve looked vulnerable in close wins over Vanderbilt, Ball State and Pittsburgh, and were not as impressive as the final score indicated in a 22-point win over a very weak Navy squad last week. We’ve long doubted their ability to win out and still give them only a 36 percent chance of pulling it off. It might not be pretty if they do find a way into the playoff — right now we’d make them a 19-point underdog to Alabama.
LSU (No. 4) and Kentucky (No. 7) have been two of the delightful surprises of the 2018 season. Both face their toughest tests this weekend and are considerable underdogs despite shiny win-loss records and home-field advantage. Unfortunately our numbers suggest that while both are solid teams, they are considerably outmatched against Alabama and Georgia. They are the two teams we expect to fall the most between the committee’s rankings this week and the end of the season.
Michigan (No. 5) and Oklahoma (No. 8) both rate in our Top 5, and are the top one-loss teams in the country. Michigan and Oklahoma are within half a point of each other, have one “good” loss, and are favored to win their conference. This combines to leave them with almost identical 50/50 chances to make the playoff. They both have multiple tough games ahead of them, but are a cut above the rest despite being doubted after early-season defeats. It is very likely one of these teams finds its way in.
Like Michigan and Oklahoma, Georgia (No. 6) has been a little overlooked following an early loss. The difference, in this case, is the seemingly insurmountable barrier Alabama poses to winning the SEC. While that is not the only way Georgia could make the playoff, it is the only one it controls, leaving it with only a 30 percent chance of making the final four.
Washington State (No. 9) and Utah (No. 13) are the Pac-12’s only hopes . . . but their playoff chances are each only 2 percent.
Ohio State (No. 10) is a big question mark. Our committee model is not high on it, due to a weak schedule and a lopsided loss. But the Buckeyes still control their own destiny — we show a one-loss Ohio State in the playoff 86 percent of the time. The question is whether the Buckeyes can bounce back from the embarrassment in West Lafayette and answer the many questions surrounding their program. We consider this, along with Notre Dame’s pillow-fight gauntlet, one of the key uncertainties that will shape the playoff.
West Virginia (No. 12) does not have a marquee win like Texas (No. 14) does, but is still in the conversation by virtue of having only one loss. It’s a mirage. Of contenders, only Washington State has faced an easier schedule. Their schedule going forward is the toughest of all contenders.
Florida (No. 11) has played well, but, with no chance left of making the SEC championship, has no path to the playoff.
With five weeks to go the playoff field may yet be settled on the field. There are seasons when the playoff four are relatively obvious, taking the politics (mostly) out of it. This season could still go either way. For most of us, the chaos is appealing. Give us too many one-loss teams, or too few — the messier the better. If so, you like this whole committee business better than you may want to admit. And for you, the best time of the year is upon us.
Cade Massey (a Wharton professor) and Rufus Peabody (a professional sports bettor) are co-founders of Massey-Peabody Analytics, which produces proprietary quantitative college football ratings. Each week, they simulate the rest of the college football season and the playoff committee’s selection process 20,000 times and explore some of the most interesting outcomes.
Read more on college football: