Wednesday, November 26

The Numbers Behind the Big 12 South

I posted earlier about what should happen in the Big 12 South - with as much objectivity as is possible, Texas or Oklahoma should represent the division in the conference championship game and one of these two teams should represent the conference in the BCS Championship game if at least one of them finishes with just one loss. Which of those two teams is more deserving is largely a matter of personal opinion.

Now the question is what will happen?

To avoid too many possibly outcomes, I am going to assume that Texas and Texas Tech win their games this week, in which both schools are favored by more than 4 TDs.

If Oklahoma beats Oklahoma State
* The 3-way tie is between Oklahoma, Texas, and Texas Tech. The tiebreaker is BCS ranking, which means it is a contest of whether Texas' lead in the computers can outweigh Oklahoma's lead on the ballots. We don't typically see huge movements at the top when all teams win unless something really wild happens (say, UF/FSU going to overtime while OU wins by 70), so the current ballot margins will likely remain about the same. Oklahoma is +.076 in the Harris over Texas and +0.275 in the ESPN/USA. This means Texas would need to be +0.351 in the computer average - ie, ahead by a cumulative 4 spots after the high and low are thrown out. What this means is that the Texas magic number is one - if Texas is on average one full ranking ahead of OU in the computer average, they go to the conference championship. Currently, Texas leads Oklahoma in four of the six computer polls.

Anderson: Texas #3, Oklahoma #4
*** Defeating #2 Texas Tech (.815) moved Oklahoma from a .768 rating to a .786. Oklahoma State is currently #14 at .706. Texas is currently rated .803. Utah is sitting at .805 and their season is over, so Texas would likely pass the Utes. I am going to predict that the Texas lead holds in Anderson, Texas +2.
Billingsley: Oklahoma #2, Texas #3.
*** Texas won't fall below #3 in one week, thanks to USC not playing UCLA just yet. Could Oklahoma hop to #1 over Alabama? It's possible, but I'd say unlikely. Oklahoma +1, maybe +2.
Colley: Texas #1, Oklahoma #5
*** Colley's allows us to plug in results. Using Texas > A&M, OU> OSU, Alabama > Auburn, Florida > FSU, we get: Texas #1, Oklahoma #3. Texas +2.
Massey: Oklahoma #1, Texas #2
*** This seems likely to hold, as Alabama is a bit far below Texas to make the jump based on beating Auburn. Oklahoma +1.
Sagarin: Texas #1, Oklahoma #4
*** Texas has a wide margin here, but #2-4 are very close. Expect Texas #1, Oklahoma #2. Texas +1.
Wolfe: Texas #2, Oklahoma #4.
*** Alabama and Texas are well ahead of the pack. Texas Tech and Oklahoma are practically tied, so expect Texas #2, Oklahoma #3. Texas +1.

This is so close. I'm going to assume Oklahoma finishes #2 in Billingsley's, above #3 Texas. Then the scores are:
Oklahoma: 4, 2, 3, 1, 2, 3 = 2.5 truncated average.
Texas: 2, 3, 1, 2, 1, 2 = 1.5 truncated average.

This gives Texas exactly enough to stay ahead of the Sooners. Note that if the gap in Anderson is Texas +1, the gap in Billingsley is Oklahoma +2, or if Sag or Wolfe flip, that would give the Sooners the edge. While it seems like Texas needs everything to go right in order to avoid this outcome, the hidden factor is that any boost to Oklahoma's rating is going to slightly boost Texas' rating as a result of their head-to-head. (Billingsley might not do this, amazingly, but Billingsley is also not a poll where that is likely to matter as the real question is how he ranks Oklahoma relative to Alabama.)

Obviously, it still comes down to voting. Oklahoma needs to increase their lead over Texas in the ballot polls by about 14% in order to overcome this predicted computer edge for the Longhorns. This will largely be determined by how both teams perform this weekend, as well as how Florida performs against Florida State since they are the interloper.


If Oklahoma State beats Oklahoma
* The Big South tiebreaker is Texas Tech's head-to-head victory over Texas. The Red Raiders are division champions.
* If Missouri defeats Texas Tech, then it's pretty clear that Texas would be the conference's highest-ranked team. Sitting at 11-1, the only teams with arguments to be included above the Longhorns would be Alabama if they lost the SEC Championship Game (12-1), USC (11-1), and Penn State (11-1). A two-loss Missouri is not getting in over Texas, mostly because of record, but also because the head-to-head result negates the conference champion argument. Realistically, this outcome sends Texas to the national championship.

If Texas Tech beats Missouri, things get more complicated.
* Currently Texas is +0.17 in the harris average (just over 4 spots, mathematically) and +0.19 in the ESPN/USA (just under 5 spots). The teams between them are USC, Penn State, and Utah. Utah and PSU cannot lose, and USC is unlikely to lose. Oklahoma would likely fall below both teams, which helps Texas Tech more than it helps Texas. The same happens after the SEC Championship (for sure if Alabama wins; if Florida wins, that might help Texas more than TTU). In any case, Texas should be a full 3 spots ahead of TTU before we start analyzing the impact that a Red Raider victory over Missouri. I think a best-case scenerio for the Red Raiders is to move up to being more-or-less tied with Texas following a win in the B12CG, leaving it all up to the computers.
* Texas is currently +2 spots over TTU in the computer average. Notably, the Longhorns are ahead by 1-5 spots in every computer poll. Playing Missouri would help the Red Raiders. Colley's Matrix allows us to plug in up to five hypothetical games. Using UT > A&M, OSU > OU, TTU > Baylor, TTU > Missouri, and either result of the SECCG, Texas ends up either 2 or 3 spots ahead of Tech depending on the SEC outcome (closer if Alabama wins). Given this result, I'm going to project that Texas Tech cannot overtake Texas in the computer polls.
* If Texas defeats Texas A&M, the Red Raiders' national title hopes are essentially over, barring some massive SEC cannibalism that results in an all-Big 12 championship game. Keep in mind that if Texas loses to A&M, Oklahoma would go to the conference championship via tiebreaker if they win over OSU. This deprives Tech of the quality opponent Missouri to boost their computer averages. Therefore, Texas Tech can only advance to the national championship game if both Texas and Oklahoma lose in the next two weeks. Even then, they still trail USC in the BCs standings. Practically speaking, their national title hopes are over.


In conclusion:
* Texas has two routes to the national championship game if they defeat A&M: getting voted into the B12CG over Oklahoma and defeating Missouri, or OU losing either this week or in the B12CG.
* Oklahoma has one route into the national championship game: winning on Saturday, getting voted into the B12CG, and defeating Missouri.

Of these, Oklahoma's path is probably the most likely to unfold should the Sooners win out. I personally think the voters will shift by enough to put OU in over Texas, largely because the voters have screwed or tried to screw Texas at every chance they get, while continuing to give the Sooners the benefit of the doubt despite multiple tankings in BCS games. Okay, so it was objective up until that last point.