If you're a frequent reader of RBR, you've probably already seen our earlier piece on assessing strength of schedule. If you haven't, be sure to check it out before reading on.
As promised, we are going to apply the same methodology described in that piece to assess team schedules for the upcoming season. Instead of FEI ratings, we are using an average of the preseason 1-120 rankings supplied by Sporting News and Rivals.
Our methodology makes a big assumption that clearly will not hold: that every opponent will end up being precisely as strong as the preseason rankings indicate. Obviously, this will not be the case, as some teams will end up being much more difficult opponents than preseason rankings indicate (see last year's top two teams, Auburn and Stanford, for example). Meanwhile, others will end up being much weaker opponents than anticipated. Each opponent's strength will certainly need to be re-assessed when we have actual data from the season, but for now, preseason rankings are the best we can do.
Remember, the win/loss probabilities are considered for an average team in the top 12--in other words, how likely is it for an average top 12 team to win against an opponent of a given ranking played at a given location? Also, for teams who must compete in an additional 13th game in a conference championship format, that 13th game is included in the probability totals, with the highest ranked team in the opposing division as the presumptive opponent.
The table below shows all 18 teams that appear in the top 20 of both preseason ranking systems mentioned above. These national contenders are ranked according to the difficulty of their schedule in terms of going undefeated or finishing with 1 loss or fewer.
| Rank | Team Schedule |
Odds of top team going undefeated
|
Odds of top team losing 1 game or fewer
|
| 1 | LSU | 1.2% | 8.3% |
| 2 | Arkansas | 1.5% | 11.0% |
| 3 | Michigan State | 2.3% | 14.7% |
| 4 | Alabama | 2.5% | 14.6% |
| 5 | South Carolina | 3.4% | 19.3% |
| 6 | Ohio State | 4.1% | 22.4% |
| 7 | Nebraska | 4.5% | 23.6% |
| 8 | Oregon | 4.7% | 24.7% |
| 9 | Oklahoma | 5.4% | 29.4% |
| 10 | Wisconsin | 6.3% | 29.0% |
| 11 | Texas A&M | 7.0% | 33.7% |
| 12 | Notre Dame | 7.5% | 33.4% |
| 13 | Florida State | 9.0% | 36.7% |
| 14 | Oklahoma State | 11.4% | 45.8% |
| 15 | Stanford | 11.9% | 43.0% |
| 16 | TCU | 21.7% | 70.2% |
| 17 | Virginia Tech | 25.8% | 69.9% |
| 18 | Boise State | 41.9% | 82.5% |
Although it's pretty widely known just how varied schedule strength can be in college football, this table really drives home the point. For the most extreme illustration, note that a top team is nearly 35 times more likely to go undefeated against Boise State's schedule than the same team would be to go undefeated against LSU's. Even with less extreme examples from BCS conferences, consider that a top team is more than 8 times more likely to finish with one loss or fewer against Virginia Tech's schedule as opposed to LSU's, and more than 20 times more likely to finish undefeated.
Even for teams that are not on either extreme end of this list, note that a top team is 3 times more likely to go undefeated against Notre Dame's schedule as opposed to Alabama's. Just think: even that is a huge difference--one team has triple the chances of another of finishing unscathed. Now consider the differences in LSU or Arkansas's schedule as compared to the teams at the bottom of the list like Stanford, TCU, Virginia Tech and Boise State. It makes even a huge gulf like Alabama/Notre Dame look insignificant by comparison.
We've got plenty to say about each team's schedule and why it is that they appear where they do on this list, but we'll save those individual breakdowns for another day. In the meantime, just consider how tough it is to judge teams by their win/loss totals with schedule strength discrepancies this big. And yes, many of these probabilities will look much different once we get a chance to see how good teams really are, but for now they do tell a story at least about how difficult respective schedules are expected to be.