Cade Massey, a practice professor at the Wharton School, and Rufus Peabody, a Washington-based sports analyst, developed this ranking system for projecting future performance. Ratings represent a team’s predicted point differential against an average team on a neutral field. Current season statistics are adjusted for home field, opponent and game situation, blended with preseason expectations and weighted by their predictive ability.
When the College Football Playoff committee unveils its first rankings Tuesday night, we’ll find out what teams committee members believe are the best so far. But the rankings tend to blur “best” and “most deserving,” which makes forecasting the process tricky.
With three years of hindsight and continuing algorithmic tweaks, we’re getting closer to understanding what the committee is looking for.
Based on that analysis, we expect the top four teams in the first rankings to be, in order, Georgia, Alabama, Clemson and Notre Dame (see table below). More interestingly, perhaps, we give three of those teams a better than 50-percent chance of remaining in the top four when the regular season is over. If that happened, it would be the first time since the new system was put in place in 2014 that the committee picked more than two playoff qualifiers right out of the gate.
The Committee’s criteria for choosing the best teams versus the most deserving ones are complex, and our understanding of them imperfect, but analyzing three years of rankings give us an objective basis for forecasting an inherently subjective process.
Of the two factors, “best” is easiest to fathom because it’s what we do every week: evaluate how good each team is in terms of how well they’ll do in the future. Meantime, a convenient way to think about “most deserving” is what ESPN calls Strength of Record (SOR). It’s related to strength-of-schedule, but considers a team’s record in addition to whom they have played.
Technically speaking, SOR is the probability an average top-25 team would have at least as many wins, given their schedule, as a team has so far achieved. For example, in our analysis, an average top-25 team would have just a nine-percent chance of winning as many games as Georgia has – all of them! – if it played the Bulldogs’ schedule. That’s easily the lowest probability of any team, thus Georgia’s top SOR ranking.
Even after tweaking how our model accounts for SOR and other factors, we know we’ll miss some of the particulars in the committee’s first rankings, but we’re confident in the broader insights revealed by them. Here are a few emerging themes we believe will color the rest of the college football season:
>> Georgia and Alabama are in a tier by themselves: We strongly expect the top two teams to be the SEC leaders. They are widely considered two of the best teams (we have them No. 1 and No. 3) and are undefeated despite challenging schedules, so they also have the two best SORs (No. 1 and No. 2). They are very close – the order easily could be flipped – and for the first time we believe they will both make the playoff.
>> It is an unusually crowded field: Few sequences of teams Nos. 3-11 would surprise us. With so many carrying similar resumes, the committee faces a thankless task. Moreover, we expect the mess to persist – at this point we forecast more than seven zero- and one-loss Power Five teams when the committee has to make its final choices. Historically, conference champs have received a bump, so amid all the bedlam, the path to the playoff is still clear: win a Power Five conference with fewer than two losses, and you are almost certain (89 percent) to get in.
>> Notre Dame complicates committee calculations: We now give the Irish a 32-percent chance of winning out, and a 16-percent chance of making the playoff. But they are especially tough to forecast. Because the Irish are an independent, the committee base a selection on a conference championship (something it potentially can do for schools such as Alabama, Clemson, Ohio State, Oklahoma and Washington). You might think reasonably that would hurt Notre Dame, but the Irish play the nation’s third-toughest schedule and have obvious cachet. Could the committee put their finger on the scale for the closest thing there is to a national team? We suspect they’ll be given some conference championship “credit” despite their independent status. Doing so ups their playoff probability from what would have been 13 percent.
>> Miami and Washington are two sides of the same coin: Miami has great SOR but looks lousy in our numbers (No. 17), so we expect them to start high and fall, especially because they have to go through Clemson in the ACC (we’d make the Tigers 10-point favorites). Miami’s seeding this week should be a good barometer for how much weight the committee puts on wins and losses. Washington is the opposite, a team with a dreadful SOR (No. 19) but still high in our power rankings (No. 5). We expect them to start low and rise over time, though the Huskies likely have too much ground to make up given the strength of the field.
>> The Pac 12 and Big 12 are on the outside looking in: Georgia’s elbowing their way into the playoff bracket comes at entire conference’s expense. The Big 12 is hardly dead, but TCU’s unexpected loss at Iowa State reduces the chance the conference champ will emerge with only one loss to 54 percent, and makes the Big 12 winner just a coin flip to produce a playoff team. Washington’s still the favorite out of the Pac-12, and the last team standing, but the weakness of their record is a major problem in a season with so many strong contenders
We’ll learn how the committee views this year’s contenders Tuesday. But there’s a fair bit of football left, and how things play out on the field is a greater source of uncertainty than the committee’s preferences. Regardless, we’re buying our No. 2 Ohio State and selling our No. 17 Miami.
— Bob Tedeschi contributed
Read more college football coverage:
Expect Georgia and Alabama to be 1-2 in first College Football Playoff rankings – Washington Post