Texas's Loss To Georgia Doesn't Matter - Kinda.
If Texas takes care of business for the rest of the season, simulations suggest their chances of winning the CFP remain largely unchanged.
Texas’s 30-15 loss to Georgia was assuredly a demoralizing moment for their locker room. After their best start to a season in a decade, they came out on their home turf on national television to get whooped on the line of scrimmage and turn the ball over 4 times. In no world is that a positive development. However, the Longhorns can take solace that they can move forward without having their chances to win the College Football Playoff seriously changed thanks to quirks in the bracket’s structure and their immense talent.
The 12-team CFP does not seed their tournament simply by ranking the 12 best teams in the nation in order. The process is as follows, according to the CFP website:
The 12 participating teams will be the five conference champions ranked highest by the CFP selection committee, plus the next seven highest-ranked teams. […]
The four highest-ranked conference champions will be seeded one through four and will receive a first-round bye. The fifth conference champion will be seeded where it was ranked or at No. 12 if it is outside the top 12 rankings.
This year, as is expected to happen most years, the 5th-highest ranked conference champion is likely to be ranked outside of the top 12. This is most likely to be Boise State who is, at time of writing, ranked as the highest non-power 4 team in the country at #17 and has a Heisman frontrunner at running back. That means we suspect there to be a significant drop off in talent between the team seeded at #12 and all other teams in the tournament.
Texas benefits in particular from this arrangement since I believe they are now the odds-on favorite to end up as the #5 seed in the CFP. In that event, they would lose out on the bye they’d get as the #1 seed but their first two games would end up against the two lowest rated conference champions. We’d expect them to be heavy favorites in both of those games considering the immense amount of talent on their roster. But heavy favorites still lose on occasion, and teams on byes do not! So it’s not obvious whether this is as good as, worse than, or perhaps even better than securing bye as their much more likely #1 seed if they were to beat Georgia.
We can evaluate the relative advantage of each seeding by simulating different hypothetical versions of the CFP and see how often Texas ends up as the national champions. First, we’ll establish a couple different reasonable expectations for what the rest of Longhorns’ season could look like. Second, we’ll find some way to rate every team in the nation. And third, we’ll use those ratings to create win probabilities to power our simulation.
First things first, here are some reasonable possibilities for the rest of Texas’s season, assuming they win out the rest of the regular season like they’re supposed to:
They end up beating Georgia in the SEC championship game and the winner of the Big 10 ends up with 1 loss. → #1 seed
They end up beating Georgia in the SEC championship game but the winner of the Big 10 ends up undefeated. → #2 seed
They lose to Georgia in the SEC championship game and the Big 10 runner up also has 2 losses. → #5 seed
They lose to Georgia in the SEC championship game and the Big 10 runner up has 1 loss. → #6 seed
This covers a wide range of potential outcomes for Texas and thus will let us get a good idea for how their potential seeding impacts their future. For what its worth, I think the most likely outcome is #3.




So how will we accomplish the simple task of evaluating every team in the country by a uniform statistical measure that will easily allow us get the win probability of any team in an arbitrary matchup? For the purposes of this post, we’ll be using CFB Data’s ELO ratings, which has a natural equation to translate into win probabilities. Note that I do add a small home field advantage for the first round, but that doesn’t majorly impact Texas. We have our brackets to test, and our method of generating win probabilities, so what do the simulation results look like?
There is shockingly little difference between any of the seeds for Texas! Don’t get too caught up on the specific probabilities here - the exact probabilities are extremely dependent on all the other teams’ seeds in each hypothetical tournament. For example, right now BYU is slated as the #11 seed in the hypotheticals. If a team like Alabama or Notre Dame instead becomes the #11 seed, all of Texas’s probabilities of winning as a 2 or 6 seed drop by a couple percentage points.
It is safe to say that Texas is not suffering from having a much more difficult road to the #1 seed. One could make a reasonable argument that on paper #5 may be ever so slightly more desirable since Texas has the talent to sleepwalk through the best both the highest ranked G5 champ and the Big 12 champion. I would disagree because I believe this system of just predicting game outcomes doesn’t properly account for the potential impacts of injuries, which we’d expect to be evenly likely across games. But the raw data is there.
The core idea here is that the ratings suggest that Texas has reached a point where there are only a handful of teams in America which can threaten them. Because of this, the main benefit they get from a bye seems to be the marginal gain of avoiding injuries. So while it is a gut punch to lose a physical game at home on national television, they can rest assured their championship chances are still intact as long as they take care of the rest of their season.