I missed last week's update, but the links are below for this week. The only good news about getting seriously ill before the Tech game is that I don't have much recollection of what happened during the game. So that's nice.
Ratings Odds and Ends
A few things to note about this year's teams. The first and most impressive, IMO, is that this year's Southern Cal defense is on pace to be rated the greatest defense of all time as measured by number of standard deviations from the mean. The other amazing part is that they will pass 2005 Ohio State by this measure. This is based on the D-1A only ratings which can be compared to any other season in major college football history. Some other items:
1.) Texas Tech, if the season ended today, would finish with the 18th highest rating of all time. Texas would be in 36th position despite the loss.
2.) Washington State's defense is on pace to be the worst of the BCS era by any current BCS conference school, and the only team close is the 2000 Connecticut Huskie squad. And they're not even that close. In fact, the only teams from BCS schools that have ever had worse defensive numbers were 1917 Mississippi, 1904 Florida, 1918 Arkansas, and 1901 Texas A&M. And those four schools' numbers are affected by a lack of data points in early seasons for southern teams. This can easily be argued to be the worst big school defense in the history of college football.
3.) Oklahoma's offense is scheduled to finish as the 10th greatest offense of all time. Their glaring weakness is that they have the worst defense of the Stoops era, and to date it is even worse than Blake's final defense in 1998.
4.) It's not surprising that this is Texas' second best offense of the Mack Brown era. What may surprise many, though, is that it is also the second best defense by this method, behind only 2005 and ahead of 2002. This method, of course, doesn't delineate between defensive and offensive points, so 2005's powerful offense certainly helped the defense. But it's important to note that Texas has played an amazingly difficult schedule in terms of offenses faced. And even overall, this is Texas' most difficult schedule under Brown, passing the slate faced by the 1998 squad.
Obviously the human polls count for 2/3 and so those will be extremely important. But those polls are obviously hard to predict because they are left to the whims of individual voters from around the nation, some of whom have proven themselves to be abject morons in the past and will continue to prove themselves such well into the future. So let's begin by analyzing the computer systems one at a time.
I'm not a big fan of this system, partially because I don't know how it works. But moreso because this sytem was devised specifically because the authors wanted to prove that Washington was better than Miami in 1991. They're probably correct, but starting out to prove a preconceived bias isn't a good way to devise a fair computer ratings set. Either way, let's look at their current standings as far as teams we're concerned about:
There's a lot of good news here, starting with the large gap between us and Florida. At this point it looks like even if Oklahoma were to beat both Tech and Oklahoma State, the scenario I'm looking at for this post because it's the only shot we've got at the Big 12 title game, they still wouldn't pass us. The bad news, though, is that for the purposes of determining the Championship Game representative from the South, a Florida win over Alabama wouldn't be included so Oklahoma will likely pass both Boise State and Florida in this set, lowering their computer poll deficit from 3 to 1 in this system. So going into the championship game, assuming OU beats Tech and Oklahoma State and we're not upset, call it 1 - Alabama, 2 - Utah, 3 - Texas, 4 - Oklahoma, 5 - Texas
Blech. This system is terrible, there's no rhyme or reason, and it's hard to predict. Let's take a look.
There is potentially terrible news in this one because of the razor-thin margin we have over the three teams behind us right now. After watching this system for a few years, though, you start to get a feel for how things will shake out. Right now I see this system in our scenario looking like this heading into the Championship Game: 1 - Alabama, 2 - Oklahoma, 3 - Florida, 4 - Texas, 5 - Texas Tech. This is truly a worst-case scenario. We go from a 3-team lead over Oklahoma to a 2-team deficit. Very bad news.
More bad news as we move through the ratings.
Colley's system is reproducible, so I put in most of the important games for the rest of the year. 1 - Texas, 2 - Florida, 3 - Oklahoma, 4 - Utah, 5 - Alabama, 6 - Texas Tech. This is not as certain as the others as there's a lot of volatility here, but I think the Top 3 is fairly solid, and that's what matters. We go from a 5-team lead to a 2-team lead.
Ah, finally we're into the three most mathematically and logically sound systems. However, these are also the most difficult to reproduce, so some guesswork
will be necessary here.
Massey's ratings take into account both game location as well as date of the game. More recent games are weighted more heavily. That means trouble for us. This is the one system where I am quite certain that Oklahoma would pass us in our scenario. However, the location factor should prevent Tech from being a wedge between us and the Sooners. In our scenario, I foresee a final ranking of 1 - Alabama, 2 - Oklahoma, 3 - Texas, 4 - Texas Tech, 5 - Utah. It's important, of course, that Tech not end up wedged between us and Oklahoma. Our 2-team lead turns into a 1-team deficit here.
Sagarin guards his formula very well, so this is all just guesswork.
There's a fairly sizeable gap between us and Oklahoma at this point, but I don't feel comfortable that the Utah-Oklahoma gap is large enough to hold on. A best guess here is a final ranking of 1 - Alabama, 2 - Texas, 3 - Oklahoma, 4 - Texas Tech, 5 - Utah. Sagarin takes into account game location but not recency of the game. At least he says he does in the schedule calculation, but I believe he takes it into account in the calculation as well.
Wolfe uses a pairwise comparison that I should in theory be able to reproduce but I haven't taken the time to attempt it yet. I know, shame on me.
Wolfe, like Massey and Sagarin, uses game location in his calculations. With this considered, we would have to guess that the order of the Big 12 South teams in our scenario would be Texas, Texas Tech, Oklahoma. That wedge between us and Oklahoma could be useful, the question becomes whether Texas Tech's weak schedule is too much to overcome by game location. But since we left Tech behind both of us in the Sagarin system, let's say the final rankings in this system look like this: 1 - Alabama, 2 - Texas, 3 - Texas Tech, 4 - Oklahoma, 5- Utah.
With these ratings, the final computer scores for Texas, Oklahoma, and Tech, would be Alabama - 1.00, Texas - 0.94, Oklahoma - 0.92, Tech - 0.86, Utah - 0.85, Florida - 0.83. Our current 0.1 lead over Oklahoma in the computers would shrink all the way to 0.02 points.
What does this mean? It means that if we maintain our current lead in the human polls over OU, our lead in the overall BCS standings would shrink from 0.0354 points to 0.0088 points. The key, then, is that we have to stay ahead of Oklahoma in the human polls. Even if OU wins out like we need them to do to get to the Championship Game, there's no guarantee we stay in front of them in the BCS. It honestly looks like a 50/50 shot at this point. It would be time to lobby voters, one of the many stupid effects of the system we have right now. And if any of my guesses above are wrong in Texas' favor at this point, we might need to be significantly ahead of OU in the polls. Time to hammer home the game locations and the Texas/OU head-to-head with the voters.