HookEm
11-27-2004, 09:34 PM
long, but good read from a Sooner board:
YCN special BCS analysis: Why Cal may well be Texas toast. Reply
--------------------------------------------------------------------------------
Just a warning first, this is an extremely lost post!
I have stated that I believe that Texas, with a good win over Texas A&M (a win by more than 7 points that isn't decided late in the game) will jump California in the BCS rankings this week and stay there. As a result, Cal will be left out of the BCS bowls unless either Southern Cal, OU or Auburn loses.
Meacham has asked that I post the analysis that made me conclude that Texas would jump California in the BCS rankings with a solid win over Texas A&M, and that California would not be able to get that top 4 ranking back in their final game against Southern Miss on 12/4. This would be somewhat of a disaster for Cal and the Rose Bowl. (so sorry)
Some of you may read this may have read some of the other BCS analysis I've done, so at least some of you are familiar with the way I work this thing. I'm going to do something a little different here, but I'm going to put my wizard cap on and dive into this analysis with all the reasoning powers that I can summon for this task.
I did the grunt work on this last night, in relatively hasty fashion, but I was entirely convinced that Texas was almost surely going BCS bowling with a good win over A&M. Today is where I get into serious detail about the why's and wherefore's of how this happens.
In order to do this analysis, it requires separate judgements based upon each of the computer rankings, and both of the polls. The polls are more of a common sense analysis rather than technical, as we all know the will of the pollsters sways in the wind.
I'll start with the computer rankings analysis.
For each individual computer rankings, only the teams ranked above and near the lowest ranked team of interest are shown. Those are the only teams that can have an impact on the relative rankings of the teams in question. Since we are dealing with evaluating individual computer rankings systems here, the rankings of the future opponents are from the respective computer rankings.
First I am going to list the current standings of teams relevant to this analysis in each of the six computer rankings that are used in the BCS ratings formula. After listing those standings, I am going to normalize all of the ratings systems used in order to have a standard "yardstick" by which to compare the relative ratings of each team as compared to the others. I viewed this as essential to getting the clearest possible picture of where each team really stands.
In order to normalize the teams, I am going to take the highest-rated team in each ranking and assign a value of 1.0000 to that team's rating in that poll. Every other team's rating as shown will be a number that represents the ratio of that team's rating to the rating of the top team in the poll.
As an example of how this works, let's say a rating system has the top team with a rating of .800 and another team at .720. I'll normalize the top team at 1.0000, and since the ratio of .720/.800 is .900, I'll assign a rating of .9000 to that team. I hope I've explained that clearly.
After I do that, I will get into a breakdown of each of the computer rankings to see where the teams should end up ranked after the remaining games. In those evaluations I will also list each team's future opponents, and their adjusted rankings in that computer rating system after all non-Division I-A teams have been removed from the ratings. (The Massey, Sagarin and Wolfe formulas include teams other than Division I-A.) And as previously stated, in these breakdowns I will use "standardized" ratings of each team for cross-comparison purposes.
Here are the six current computer rankings, with the rank listed first, then the rating (as used in that computer system), then the name of the team.
Anderson & Hester
1 .841 OU
2 .841 USC
3 .825 Auburn
4 .804 Utah
5 .787 California
6 .786 Texas
7 .784 Boise St
8 .783 Arizona St
Billingsley
1 334.347 OU
2 319.712 Auburn
3 319.166 USC
4 292.947 Texas
5 280.997 Boise St
6 278.160 Utah
7 274.438 Michigan
8 273.273 Georgia
9 271.557 California
10 271.475 LSU
Colley
1 .980034 OU
2 .960873 USC
3 .947769 Auburn
4 .893067 Texas
5 .890308 Arizona St
6 .888389 Utah
7 .865158 California
8 .863446 Boise St
Massey
1 5.460 USC
2 5.116 OU
3 4.961 Auburn
4 4.851 California
5 4.735 Utah
6 4.716 Arizona St
7 4.654 Boise St
8 4.651 Texas
Sagarin
1 100.50 OU
2 100.45 USC
3 94.12 Auburn
4 94.05 California
5 92.36 Texas
6 92.02 Utah
7 91.87 Boise St
8 91.04 Arizona St
Wolfe
1 8.984 OU
2 8.488 USC
3 7.918 Auburn
4 7.539 Texas
5 7.506 California
6 7.433 Utah
7 7.267 Boise St
8 7.208 Arizona St
So that is where all the computer rankings currently stand. For BCS calculation purposes, only the rankings for each computer rating system matter, but for this analysis the most important numbers are the ratings, which will be evaluated to see what kind of movement will occur in the rankings themselves. At issue are the following factors:
1. Where is each team currently ranked in this system, and how close is their rating to the other teams near them in the rankings?
2. What future opponents do these teams play, and where? It is important also to look at the current rankings of those future opponents individually for each computer ranking system, because they vary in each one.
3. Are any changes in the relative rankings of the teams likely to occur, and why?
And that is basically the analysis in a nutshell for the computer portion of this overall exercise. Now that I've covered that, here are the computer rankings using the adjusted numbers, and with future opponents and their rankings in those individual computers shown. After each individual rankings group is shown, I will analyze what I foresee for the final rankings in each one.
In any sort of analysis, in order to remove the white noise from the data, it is necessary to make some assumptions based upon what is MOST important and what isn't. In addition, it is necessary in any individual analysis to base your reasoning on a predetermined and limited set of outcomes. Otherwise the data becomes garbage, or the analysis starts to look like a Russian novel. Here are the assumptions that I have used, and go with them where you may.
1. I am assuming that all of these highly ranked teams will win all of their remaining games with teams that are ranked below them. That assumption may not make some people happy, so at the end of this analysis I will attempt to look at the amount of change and chaos any reasonably possible outcomes might create in the final BCS standings and bowl assignments.
2. I am assuming that with the large pool of opponent's opponents, that the actual impact on the individual computer ratings of each team will be relatively insignificant. I will only look at what effect playing future opponents is likely to have on the rankings of our teams of interest, in terms of adjusted strength of schedule.
3. I am going to assume the the world will not end prior to the bowl season being over. Keep your fingers crossed.
Here are the adjusted computer ratings:
Anderson & Hester (normalized ratings, not actual)
1 1.0000 OU - plays vs either 33 Iowa State or 27 Colorado in Big 12 Championship Game (12/4)
2 0.9999 USC - plays 28 Notre Dame (11/27), at 37 UCLA (12/4)
3 0.9812 Auburn - plays vs 12 Tennessee in SEC Championship Game (12/4)
4 0.9562 Utah - season ended
5 0.9360 California - plays at 63 Southern Miss (12/4)
6 0.9348 Texas - plays 13 Texas A&M (11/26)
7 0.9324 Boise St - plays at 97 Nevada (12/27)
8 0.9312 Arizona St - plays at 87 Arizona (12/26)
I'll go into greater detail in this first analysis, because then I can skip the boring stuff later.
Remember that the purpose here is to see where California and Texas wind up in the entire group of computer rankings relative to each other. The other teams are not in there for filler, but are necessary for this analysis to make good sense. It's clear from looking at the numbers that the top four teams are unassailable from below, given the large .0202 gap from #4 to #5, and the mediocre quality of Cal's final opponent. Boise State and Arizona State play very weak opponents, so they cannot move up. You wonder if maybe Texas couldn't rise the amount needed, given the fact they are playing a highly ranked opponent, and they also inherit that opponent's strength of schedule. But wondering about something isn't really fact-finding, so we will limit this to Texas and Cal. Texas trails Cal by a paltry .0012 here, and there's not a shadow of a doubt that the .0012 and considerably more is going to disappear this week with a Texas win. So much so, that with a weak Southern Miss team as Cal's remaining opponent, a Texas win is going to definitely mean UT #5, Cal #6.
Let me point out at this time that in the computer ratings systems, margin of victory has no bearing whatsoever on how teams get ranked. Texas gets the same benefit from beating A&M if they win by one in four overtimes or by 84 in a game shortened by the mercy rule. (just kidding)
Let me also state that the selection of teams shown only includes all the teams down to the last one were are interested in, and any other teams that could possibly change ranking relative to either Texas or Cal.
Billingsley (normalized ratings, not actual)
1 1.0000 OU - plays vs either 41 Iowa State or 29 Colorado in Big 12 Championship Game (Sat, 12/4)
2 0.9562 Auburn - plays vs 13 Tennessee in SEC Championship Game (Sat, 12/4)
3 0.9546 USC - plays 37 Notre Dame (Sat, 11/27), at 46 UCLA (Sat, 12/4)
4 0.8762 Texas - plays 25 Texas A&M (Fri, 11/26)
5 0.8404 Boise St - plays at 94 Nevada (12/27)
6 0.8320 Utah - season ended
7 0.8208 Michigan - season ended
8 0.8173 Georgia - plays 36 Georgia Tech (Sat, 11/27)
9 0.8122 California - plays at 56 Southern Miss (Sat, 12/4)
10 0.8120 LSU - plays vs 35 Arkansas (Fri, 11/26)
Texas is literally locked in at #4 in Billingsley's rankings. There is a giant gap between Texas and the top three, and a major margin down to Boise St, who finishes with a dud opponent. The next two teams have finished their seasons, and there isn't any way for any of the teams below that to make up ground. Cal is sitting way below at #9, with no prospect to climb above anyone with such an average closing opponent. But with LSU only .0002 behind Cal and playing a team ranked 21 places higher here, guess where Cal is going to go? Texas #4, Cal #10.
Colley (normalized ratings, not actual)
1 1.0000 OU - plays vs either 35 Iowa State or 29 Colorado in Big 12 Championship Game (Sat, 12/4)
2 0.9804 USC - plays 30 Notre Dame (Sat, 11/27), at 39 UCLA (Sat, 12/4)
3 0.9671 Auburn - plays vs 12 Tennessee in SEC Championship Game (Sat, 12/4)
4 0.9113 Texas - plays 14 Texas A&M (Fri, 11/26)
5 0.9084 Arizona St - plays at 89 Arizona (12/26)
6 0.9065 Utah - season ended
7 0.8828 California - plays at 58 Southern Miss (12/4)
8 0.8810 Boise St - plays at 96 Nevada (12/27)
Once again, Texas is locked in place, too far behind #3, an leading a #5 that plays a weaker opponent, #6 is done for the season, and Cal plays a much weaker opponent. Cal isn't going anywhere either, not with Utah ahead by a good margin and Boise St playing a sad team. Texas #4, Cal #7
Massey (normalized ratings, not actual)
1 1.0000 USC - plays 31 Notre Dame (Sat, 11/27), at 36 UCLA (Sat, 12/4)
2 0.9370 OU - plays vs either 32 Iowa State or 28 Colorado in Big 12 Championship Game (Sat, 12/4)
3 0.9086 Auburn - plays vs 14 Tennessee in SEC Championship Game (Sat, 12/4)
4 0.8885 California - plays at 68 Southern Miss (12/4)
5 0.8672 Utah - season ended
6 0.8637 Arizona St - plays at 85 Arizona (12/26)
7 0.8524 Boise St - plays at 103 Nevada (12/27)
8 0.8518 Texas - plays 12 Texas A&M (Fri, 11/26)
This time Cal is locked at #4, while Texas can move up. Without a doubt UT will pass Boise St, who they trail by only .0006, and it wouldn't be shocking to see them get all the way up to #5. But that's just what might happen, not what will happen. I think it is very safe to split the middle, because Arizona State playing #85 Arizona will definitely drop their rating, while Texas playing #12 A&M will definitely cause theirs to climb a significant amount. Texas #6, Cal #4.
Sagarin (normalized ratings, not actual)
1 1.0000 OU - plays vs either 31 Iowa State or 30 Colorado in Big 12 Championship Game (Sat, 12/4)
2 0.9995 USC - plays 35 Notre Dame (Sat, 11/27), at 38 UCLA (Sat, 12/4)
3 0.9365 Auburn - plays vs 17 Tennessee in SEC Championship Game (Sat, 12/4)
4 0.9358 California - plays at 55 Southern Miss (12/4)
5 0.9190 Texas - plays 10 Texas A&M (Fri, 11/26)
6 0.9156 Utah - season ended
7 0.9141 Boise St - plays at 104 Nevada (12/27)
8 0.9059 Arizona St - plays at 88 Arizona (12/26)
Here we finally have a side-by-side ranking of Cal and Texas, where an upward move by Texas would do double damage, in that Cal would have to move down. There is nothing in this group of teams that would indicate that UT and Cal would end up at #4 and #5 in either order. While the margin of .0168 between Texas and Cal isn't all that large, the question of whether or not Texas would pass Cal isn't a clear one to determine. So for the sake of erring on the side of caution, I'll just leave them where they are. Texas #5, Cal #4.
Wolfe (normalized ratings, not actual)
1 1.0000 OU - plays vs either 30 Iowa State or 32 Colorado in Big 12 Championship Game (Sat, 12/4)
2 0.9448 USC - plays 35 Notre Dame (Sat, 11/27), at 42 UCLA (Sat, 12/4)
3 0.8813 Auburn - plays vs 15 Tennessee in SEC Championship Game (Sat, 12/4)
4 0.8392 Texas - plays 12 Texas A&M (Fri, 11/26)
5 0.8355 California - plays at 48 Southern Miss (12/4)
6 0.8274 Utah - season ended
7 0.8089 Boise St - plays at 96 Nevada (12/27)
8 0.8023 Arizona St - plays at 97 Arizona (12/26)
Down to the last computer rankings, and again it is side-by-side for Texas and Cal. There's no upside or downside for either team, and UT won't lose ground to Cal. Texas #4, Cal #5.
Now that I've finished this part of the analysis, let's see where these teams figure to end up in each of the computer rankings all in one spot, what their average computer rankings are, and what that means in terms of that component of the BCS rating formula.
Texas 5 4 4 6 5 5
Cal 6 10 7 4 4 5
Dropping the best and worst rankings for each team, we have
Texas 5 + 4 + 5 + 5 = 19
Cal 6 + 7 + 4 + 5 = 22
Skipping to the result, we have
Texas .8500
Cal .8200
Texas will have a lead on Cal in the computer rating of .0300 at the very minimum. Currently Texas leads Cal by .0200 in this category.
I'm done with the computers, and the analysis says that Texas will increase their lead in this component by at least .0100, which reduces the amount that they need to gain in the polls. So let's look at the polls, and we won't use normalized numbers here, since they are raw vote totals. There is no particular reason to look at anyone but Texas and California here, because it is the difference in points between only those two teams that is in question.
First, though, let's get an idea of where the BCS ratings are and would be with the computer ratings change factored in
Currently the BCS ratings are:
4 California .8695 .8518 .8300 .8504
5 Texas .8142 .8262 .8500 .8301
The numbers above left-to-right are: AP poll points percentage, USA Today/ESPN poll points percentage, computer rankings percentage, current BCS rating.
So with the change in the computer rankings, the numbers look like this:
4 California .8695 .8518 .8200 .8471
5 Texas .8142 .8262 .8500 .8301
Before the polls are factored in, Cal leads Texas by .0170. That seems like such a cold, abstract number, .0170. What can be done with that? And what does it mean?
In a typical case it means exactly this. The BCS rating is an average of three parts - the two polls and the computer rating. The formula takes the three parts and divides by three to get the percentage, expressed as a 4-place decimal of the total possible computer and poll ranking points that a team can have. If a team is unanimously #1 in both polls and the computers, then that team gets the holy grail of BCS rankings, a perfect score of 1.0000. No poll votes and no top 25 computer rankings and you get exactly 0.0000, which means thanks for stopping by.
There is no need to average the three parts to find the relative ranking of two teams to each other, just add the three parts.
4 California .8695 + .8518 + .8200 = 2.5413
5 Texas .8142 + .8262 + .8500 = 2.4904
That is a total difference of .0509. That is the total points lead that Cal has on Texas when you factor in the change in the computer rankings for their future opponents. Ordinarily that would be some pretty serious points, but not in this case. The reason for that is that Cal and Tex are right next to each other in the polls as the top two teams in the rankings with one loss. And there is where Texas has a big glaring advantage on California.
Let's look at those poll rankings again for Texas and Cal, but this time let's look at the whole top 6.
AP Poll
1 1603 USC
2 1541 OU
3 1536 Auburn
4 1413 California
5 1340 Utah
6 1323 Texas
USA Today/ESPN Poll
1 1510 USC
2 1440 OU
3 1436 Auburn
4 1299 California
5 1260 Texas
6 1246 Utah
Let's see by groupings where these votes are concentrated.
In the AP poll, all of the first place votes are worth a combined 1,625 points, 2nd place 1,560, 3rd place 1,495. Add those together and you get 4,680 points. USC, OU and Auburn combine for 4,680 points. So none of Cal, Utah and Texas has a ranking on any ballot above #4.
All of the 4th-6th place votes are worth a total of 4,095 points, and Cal, Utah and Texas have 4,076 of them, leaving only 19 total points for teams ranked below #6 from the 4-6 voting positions
In the USA Today/ESPN poll, those same 1st-3rd place points total 4,392 points. USC, OU and Auburn have a total of 4,386 points. That leaves 6 points in any form available for the teams ranked below them.
Those six points plus the 3,843 points that the 4th-6th place votes are worth totals 3,849 points, while Cal, Texas and Utah combine for 3,805 points. That is all but 44 of the total points available for the top 6 teams.
What that exercise is telling us is that almost all the available points that can affect Cal or Texas positively are going to come from the three team group of Utah, Texas and Cal. So what are the real implications of that? Here is what they are:
1. Utah's regular season is done; a bowl game is next.
2. Cal plays at mediocre Southern Miss on the final day of the season.
3. Texas plays against a ranked Texas A&M team on the day after Thanksgiving.
Quick question: Which team has by far the greatest upside here?
This isn't rocket science; it's poll science. A little bit voodoo, but mostly common sense. And we are certainly to the debatable portion of this issue. And if I am at fault in my analysis, this is where I will get attacked.
Texas has a long-standing reputation as a team that cannot get over the two-loss hump with Mack Brown. This is a widely held belief among the pollsters for sure; or else Texas would not have been slapped much more severely for losing to #2 Oklahoma than Cal was for losing to #1 USC. What has been worse for Texas is that they've been the team that can't beat the good teams, as their record in recent years against top 25 teams hasn't been good at all.
Things are different for Texas this year. Texas fought all the way to the end against OU, and although once again they lost, they fought to the last and held Oklahoma to 16 fewer points than any other opponent this year. They might have struggled in some other games, but so has Cal, and Utah has had a dramatically weaker schedule this year. If Texas was to beat A&M in solid fashion, how many points do you think they would steal away from Utah, and especially from Cal? And if they beat A&M by 21 points, the same amount that Utah beat A&M by, or by more points than that, how many points do you think they would take away from Utah and Cal?
Here is where my reasoning comes to fruition. Texas doesn’t need a signature win, but UT will be playing the second best team on their entire schedule on Friday. This will carry a strong impact with the voters with a solid win, by which I mean 14 points or more. More than 21 points and Texas will steal a LOT of points from both Cal and Utah. They will definitely take a significant number of points from both teams with any win of 7 or more points. This is just naturally to be expected when only Texas of that group of teams is playing this week. Positive news unanswered means positive vote pulls for the Horns, if they win with any margin.
The question is, will they take away enough points? And won't Cal have the opportunity to get those points back? I think that very clearly the answers are a resounding yes and yes. Let's look at the background to where we are a little more clearly.
Here are last week's point totals for Cal and Texas in the polls:
AP
Cal 1,409
Texas 1,301
Utah 1,316
USA Today/ESPN
Cal 1,311
Texas 1,222
Utah 1,203
On the most recent weekend, USC was idle, Oklahoma won 35-0 at Baylor, Auburn won 21-13 at Alabama, Cal stomped Stanford 31-6, Texas was idle, Utah drilled BYU 52-21, and most importantly Michigan, Florida State and Wisconsin all lost. That left some significant points to be picked up by the teams around and above them.
Here are this week's point totals in the polls:
AP
Cal 1,413
Texas 1,323
Utah 1,340
USA Today/ESPN
Cal 1,299
Texas 1,260
Utah 1,246
What! In the AP poll Cal gained 4 points, while Texas gained 22. And in the USA Today/ESPN poll, Cal LOST 12 points, while Texas GAINED 38! That is a net gain for Texas of 68 total points, on a weekend where Cal won by 25 and Texas didn't even play a game. Yes, Utah picked up 67 total points, to all but ensure a top 6 finish, when they blasted BYU 52-21. But for Cal it was a resounding no vote, a real indication that the voters are starting to question just how strong a conference really is with only 3 of 10 teams with more than six wins after having played a relatively weak schedule. And that schedule was one that they were able to navigate with only 7 more wins than losses. Compare that to the Big 12, which has won 20 more games than lost.
What do those numbers tell us? It tell us exactly this - that people are starting to believe that Texas is NOT a fluke this year, and that Cal may in fact be overrated.
This is the time of year, and especially in a year with three unbeaten teams at the top of the polls, that the pollsters are going to scrutinize the relative strengths more closely, and in doing so, they are seeing that Texas has played a stronger schedule than Cal. There is compensation going on for treating UT so poorly after their loss to Oklahoma, where they held the Sooners to 16 fewer points than the Sooners have scored on any other team.
If Texas can gain 68 points on Cal while idle and with Cal drilling Stanford 31-7, what is going to happen this week if Texas wins by 14 or more points while Cal is idle, and while Utah is done playing games?
I’ll tell you what will occur. Texas will add the second best win to their schedule, on national television in front of a large audience, and inherit the nation’s toughest strength of schedule from Texas A&M. And that will turn even more voters their way than last week. Try maybe 75-100 more total points in the two polls.
And we know that almost all of those points will be coming at the expense of Cal and Utah, not the unbeaten teams above, who are unassailable, or the two-loss teams below, who don’t have any appreciable points to steal anyway.
Let’s assume that Texas only climbs by 50 total points with a solid win over A&M on Friday. And let’s just assume for argument that they steal half of those points from both Cal and Utah.
This is where that kind of move would take them in the BCS ratings:
California .8542 .8354 .8200 .8365
Texas .8449 .8590 .8500 .8513
That’s a very significant lead, and that is being conservative in just about every way. I believe that it is very much more likely that the margin for Texas will be even larger in the final BCS rankings. Oklahoma may still add another win over a 7-4 Iowa State or Colorado to their opponent’s resume, while only USC at UCLA will remain for Cal’s opponents, and that game is a wash for Cal, since they played both teams.
Cal only has a very average Southern Miss team to play, which might actually hurt them rather than help them. The voters will have already discounted a strong win over a much weaker team in this week’s poll, and even if Cal blows out Southern Miss it will be largely unnoticed by voters in the mass of end of the season stories that vastly surpass that game in general interest.
In short, I believe that analysis says that almost without a doubt Texas will pass Cal on Friday with a solid win over Texas A&M, and they will stay ahead through the final BCS rankings.
Bye-bye Cal.
Thanks for reading to the end of this ridiculously long post. I had no idea it would be this long when I started.
YCN special BCS analysis: Why Cal may well be Texas toast. Reply
--------------------------------------------------------------------------------
Just a warning first, this is an extremely lost post!
I have stated that I believe that Texas, with a good win over Texas A&M (a win by more than 7 points that isn't decided late in the game) will jump California in the BCS rankings this week and stay there. As a result, Cal will be left out of the BCS bowls unless either Southern Cal, OU or Auburn loses.
Meacham has asked that I post the analysis that made me conclude that Texas would jump California in the BCS rankings with a solid win over Texas A&M, and that California would not be able to get that top 4 ranking back in their final game against Southern Miss on 12/4. This would be somewhat of a disaster for Cal and the Rose Bowl. (so sorry)
Some of you may read this may have read some of the other BCS analysis I've done, so at least some of you are familiar with the way I work this thing. I'm going to do something a little different here, but I'm going to put my wizard cap on and dive into this analysis with all the reasoning powers that I can summon for this task.
I did the grunt work on this last night, in relatively hasty fashion, but I was entirely convinced that Texas was almost surely going BCS bowling with a good win over A&M. Today is where I get into serious detail about the why's and wherefore's of how this happens.
In order to do this analysis, it requires separate judgements based upon each of the computer rankings, and both of the polls. The polls are more of a common sense analysis rather than technical, as we all know the will of the pollsters sways in the wind.
I'll start with the computer rankings analysis.
For each individual computer rankings, only the teams ranked above and near the lowest ranked team of interest are shown. Those are the only teams that can have an impact on the relative rankings of the teams in question. Since we are dealing with evaluating individual computer rankings systems here, the rankings of the future opponents are from the respective computer rankings.
First I am going to list the current standings of teams relevant to this analysis in each of the six computer rankings that are used in the BCS ratings formula. After listing those standings, I am going to normalize all of the ratings systems used in order to have a standard "yardstick" by which to compare the relative ratings of each team as compared to the others. I viewed this as essential to getting the clearest possible picture of where each team really stands.
In order to normalize the teams, I am going to take the highest-rated team in each ranking and assign a value of 1.0000 to that team's rating in that poll. Every other team's rating as shown will be a number that represents the ratio of that team's rating to the rating of the top team in the poll.
As an example of how this works, let's say a rating system has the top team with a rating of .800 and another team at .720. I'll normalize the top team at 1.0000, and since the ratio of .720/.800 is .900, I'll assign a rating of .9000 to that team. I hope I've explained that clearly.
After I do that, I will get into a breakdown of each of the computer rankings to see where the teams should end up ranked after the remaining games. In those evaluations I will also list each team's future opponents, and their adjusted rankings in that computer rating system after all non-Division I-A teams have been removed from the ratings. (The Massey, Sagarin and Wolfe formulas include teams other than Division I-A.) And as previously stated, in these breakdowns I will use "standardized" ratings of each team for cross-comparison purposes.
Here are the six current computer rankings, with the rank listed first, then the rating (as used in that computer system), then the name of the team.
Anderson & Hester
1 .841 OU
2 .841 USC
3 .825 Auburn
4 .804 Utah
5 .787 California
6 .786 Texas
7 .784 Boise St
8 .783 Arizona St
Billingsley
1 334.347 OU
2 319.712 Auburn
3 319.166 USC
4 292.947 Texas
5 280.997 Boise St
6 278.160 Utah
7 274.438 Michigan
8 273.273 Georgia
9 271.557 California
10 271.475 LSU
Colley
1 .980034 OU
2 .960873 USC
3 .947769 Auburn
4 .893067 Texas
5 .890308 Arizona St
6 .888389 Utah
7 .865158 California
8 .863446 Boise St
Massey
1 5.460 USC
2 5.116 OU
3 4.961 Auburn
4 4.851 California
5 4.735 Utah
6 4.716 Arizona St
7 4.654 Boise St
8 4.651 Texas
Sagarin
1 100.50 OU
2 100.45 USC
3 94.12 Auburn
4 94.05 California
5 92.36 Texas
6 92.02 Utah
7 91.87 Boise St
8 91.04 Arizona St
Wolfe
1 8.984 OU
2 8.488 USC
3 7.918 Auburn
4 7.539 Texas
5 7.506 California
6 7.433 Utah
7 7.267 Boise St
8 7.208 Arizona St
So that is where all the computer rankings currently stand. For BCS calculation purposes, only the rankings for each computer rating system matter, but for this analysis the most important numbers are the ratings, which will be evaluated to see what kind of movement will occur in the rankings themselves. At issue are the following factors:
1. Where is each team currently ranked in this system, and how close is their rating to the other teams near them in the rankings?
2. What future opponents do these teams play, and where? It is important also to look at the current rankings of those future opponents individually for each computer ranking system, because they vary in each one.
3. Are any changes in the relative rankings of the teams likely to occur, and why?
And that is basically the analysis in a nutshell for the computer portion of this overall exercise. Now that I've covered that, here are the computer rankings using the adjusted numbers, and with future opponents and their rankings in those individual computers shown. After each individual rankings group is shown, I will analyze what I foresee for the final rankings in each one.
In any sort of analysis, in order to remove the white noise from the data, it is necessary to make some assumptions based upon what is MOST important and what isn't. In addition, it is necessary in any individual analysis to base your reasoning on a predetermined and limited set of outcomes. Otherwise the data becomes garbage, or the analysis starts to look like a Russian novel. Here are the assumptions that I have used, and go with them where you may.
1. I am assuming that all of these highly ranked teams will win all of their remaining games with teams that are ranked below them. That assumption may not make some people happy, so at the end of this analysis I will attempt to look at the amount of change and chaos any reasonably possible outcomes might create in the final BCS standings and bowl assignments.
2. I am assuming that with the large pool of opponent's opponents, that the actual impact on the individual computer ratings of each team will be relatively insignificant. I will only look at what effect playing future opponents is likely to have on the rankings of our teams of interest, in terms of adjusted strength of schedule.
3. I am going to assume the the world will not end prior to the bowl season being over. Keep your fingers crossed.
Here are the adjusted computer ratings:
Anderson & Hester (normalized ratings, not actual)
1 1.0000 OU - plays vs either 33 Iowa State or 27 Colorado in Big 12 Championship Game (12/4)
2 0.9999 USC - plays 28 Notre Dame (11/27), at 37 UCLA (12/4)
3 0.9812 Auburn - plays vs 12 Tennessee in SEC Championship Game (12/4)
4 0.9562 Utah - season ended
5 0.9360 California - plays at 63 Southern Miss (12/4)
6 0.9348 Texas - plays 13 Texas A&M (11/26)
7 0.9324 Boise St - plays at 97 Nevada (12/27)
8 0.9312 Arizona St - plays at 87 Arizona (12/26)
I'll go into greater detail in this first analysis, because then I can skip the boring stuff later.
Remember that the purpose here is to see where California and Texas wind up in the entire group of computer rankings relative to each other. The other teams are not in there for filler, but are necessary for this analysis to make good sense. It's clear from looking at the numbers that the top four teams are unassailable from below, given the large .0202 gap from #4 to #5, and the mediocre quality of Cal's final opponent. Boise State and Arizona State play very weak opponents, so they cannot move up. You wonder if maybe Texas couldn't rise the amount needed, given the fact they are playing a highly ranked opponent, and they also inherit that opponent's strength of schedule. But wondering about something isn't really fact-finding, so we will limit this to Texas and Cal. Texas trails Cal by a paltry .0012 here, and there's not a shadow of a doubt that the .0012 and considerably more is going to disappear this week with a Texas win. So much so, that with a weak Southern Miss team as Cal's remaining opponent, a Texas win is going to definitely mean UT #5, Cal #6.
Let me point out at this time that in the computer ratings systems, margin of victory has no bearing whatsoever on how teams get ranked. Texas gets the same benefit from beating A&M if they win by one in four overtimes or by 84 in a game shortened by the mercy rule. (just kidding)
Let me also state that the selection of teams shown only includes all the teams down to the last one were are interested in, and any other teams that could possibly change ranking relative to either Texas or Cal.
Billingsley (normalized ratings, not actual)
1 1.0000 OU - plays vs either 41 Iowa State or 29 Colorado in Big 12 Championship Game (Sat, 12/4)
2 0.9562 Auburn - plays vs 13 Tennessee in SEC Championship Game (Sat, 12/4)
3 0.9546 USC - plays 37 Notre Dame (Sat, 11/27), at 46 UCLA (Sat, 12/4)
4 0.8762 Texas - plays 25 Texas A&M (Fri, 11/26)
5 0.8404 Boise St - plays at 94 Nevada (12/27)
6 0.8320 Utah - season ended
7 0.8208 Michigan - season ended
8 0.8173 Georgia - plays 36 Georgia Tech (Sat, 11/27)
9 0.8122 California - plays at 56 Southern Miss (Sat, 12/4)
10 0.8120 LSU - plays vs 35 Arkansas (Fri, 11/26)
Texas is literally locked in at #4 in Billingsley's rankings. There is a giant gap between Texas and the top three, and a major margin down to Boise St, who finishes with a dud opponent. The next two teams have finished their seasons, and there isn't any way for any of the teams below that to make up ground. Cal is sitting way below at #9, with no prospect to climb above anyone with such an average closing opponent. But with LSU only .0002 behind Cal and playing a team ranked 21 places higher here, guess where Cal is going to go? Texas #4, Cal #10.
Colley (normalized ratings, not actual)
1 1.0000 OU - plays vs either 35 Iowa State or 29 Colorado in Big 12 Championship Game (Sat, 12/4)
2 0.9804 USC - plays 30 Notre Dame (Sat, 11/27), at 39 UCLA (Sat, 12/4)
3 0.9671 Auburn - plays vs 12 Tennessee in SEC Championship Game (Sat, 12/4)
4 0.9113 Texas - plays 14 Texas A&M (Fri, 11/26)
5 0.9084 Arizona St - plays at 89 Arizona (12/26)
6 0.9065 Utah - season ended
7 0.8828 California - plays at 58 Southern Miss (12/4)
8 0.8810 Boise St - plays at 96 Nevada (12/27)
Once again, Texas is locked in place, too far behind #3, an leading a #5 that plays a weaker opponent, #6 is done for the season, and Cal plays a much weaker opponent. Cal isn't going anywhere either, not with Utah ahead by a good margin and Boise St playing a sad team. Texas #4, Cal #7
Massey (normalized ratings, not actual)
1 1.0000 USC - plays 31 Notre Dame (Sat, 11/27), at 36 UCLA (Sat, 12/4)
2 0.9370 OU - plays vs either 32 Iowa State or 28 Colorado in Big 12 Championship Game (Sat, 12/4)
3 0.9086 Auburn - plays vs 14 Tennessee in SEC Championship Game (Sat, 12/4)
4 0.8885 California - plays at 68 Southern Miss (12/4)
5 0.8672 Utah - season ended
6 0.8637 Arizona St - plays at 85 Arizona (12/26)
7 0.8524 Boise St - plays at 103 Nevada (12/27)
8 0.8518 Texas - plays 12 Texas A&M (Fri, 11/26)
This time Cal is locked at #4, while Texas can move up. Without a doubt UT will pass Boise St, who they trail by only .0006, and it wouldn't be shocking to see them get all the way up to #5. But that's just what might happen, not what will happen. I think it is very safe to split the middle, because Arizona State playing #85 Arizona will definitely drop their rating, while Texas playing #12 A&M will definitely cause theirs to climb a significant amount. Texas #6, Cal #4.
Sagarin (normalized ratings, not actual)
1 1.0000 OU - plays vs either 31 Iowa State or 30 Colorado in Big 12 Championship Game (Sat, 12/4)
2 0.9995 USC - plays 35 Notre Dame (Sat, 11/27), at 38 UCLA (Sat, 12/4)
3 0.9365 Auburn - plays vs 17 Tennessee in SEC Championship Game (Sat, 12/4)
4 0.9358 California - plays at 55 Southern Miss (12/4)
5 0.9190 Texas - plays 10 Texas A&M (Fri, 11/26)
6 0.9156 Utah - season ended
7 0.9141 Boise St - plays at 104 Nevada (12/27)
8 0.9059 Arizona St - plays at 88 Arizona (12/26)
Here we finally have a side-by-side ranking of Cal and Texas, where an upward move by Texas would do double damage, in that Cal would have to move down. There is nothing in this group of teams that would indicate that UT and Cal would end up at #4 and #5 in either order. While the margin of .0168 between Texas and Cal isn't all that large, the question of whether or not Texas would pass Cal isn't a clear one to determine. So for the sake of erring on the side of caution, I'll just leave them where they are. Texas #5, Cal #4.
Wolfe (normalized ratings, not actual)
1 1.0000 OU - plays vs either 30 Iowa State or 32 Colorado in Big 12 Championship Game (Sat, 12/4)
2 0.9448 USC - plays 35 Notre Dame (Sat, 11/27), at 42 UCLA (Sat, 12/4)
3 0.8813 Auburn - plays vs 15 Tennessee in SEC Championship Game (Sat, 12/4)
4 0.8392 Texas - plays 12 Texas A&M (Fri, 11/26)
5 0.8355 California - plays at 48 Southern Miss (12/4)
6 0.8274 Utah - season ended
7 0.8089 Boise St - plays at 96 Nevada (12/27)
8 0.8023 Arizona St - plays at 97 Arizona (12/26)
Down to the last computer rankings, and again it is side-by-side for Texas and Cal. There's no upside or downside for either team, and UT won't lose ground to Cal. Texas #4, Cal #5.
Now that I've finished this part of the analysis, let's see where these teams figure to end up in each of the computer rankings all in one spot, what their average computer rankings are, and what that means in terms of that component of the BCS rating formula.
Texas 5 4 4 6 5 5
Cal 6 10 7 4 4 5
Dropping the best and worst rankings for each team, we have
Texas 5 + 4 + 5 + 5 = 19
Cal 6 + 7 + 4 + 5 = 22
Skipping to the result, we have
Texas .8500
Cal .8200
Texas will have a lead on Cal in the computer rating of .0300 at the very minimum. Currently Texas leads Cal by .0200 in this category.
I'm done with the computers, and the analysis says that Texas will increase their lead in this component by at least .0100, which reduces the amount that they need to gain in the polls. So let's look at the polls, and we won't use normalized numbers here, since they are raw vote totals. There is no particular reason to look at anyone but Texas and California here, because it is the difference in points between only those two teams that is in question.
First, though, let's get an idea of where the BCS ratings are and would be with the computer ratings change factored in
Currently the BCS ratings are:
4 California .8695 .8518 .8300 .8504
5 Texas .8142 .8262 .8500 .8301
The numbers above left-to-right are: AP poll points percentage, USA Today/ESPN poll points percentage, computer rankings percentage, current BCS rating.
So with the change in the computer rankings, the numbers look like this:
4 California .8695 .8518 .8200 .8471
5 Texas .8142 .8262 .8500 .8301
Before the polls are factored in, Cal leads Texas by .0170. That seems like such a cold, abstract number, .0170. What can be done with that? And what does it mean?
In a typical case it means exactly this. The BCS rating is an average of three parts - the two polls and the computer rating. The formula takes the three parts and divides by three to get the percentage, expressed as a 4-place decimal of the total possible computer and poll ranking points that a team can have. If a team is unanimously #1 in both polls and the computers, then that team gets the holy grail of BCS rankings, a perfect score of 1.0000. No poll votes and no top 25 computer rankings and you get exactly 0.0000, which means thanks for stopping by.
There is no need to average the three parts to find the relative ranking of two teams to each other, just add the three parts.
4 California .8695 + .8518 + .8200 = 2.5413
5 Texas .8142 + .8262 + .8500 = 2.4904
That is a total difference of .0509. That is the total points lead that Cal has on Texas when you factor in the change in the computer rankings for their future opponents. Ordinarily that would be some pretty serious points, but not in this case. The reason for that is that Cal and Tex are right next to each other in the polls as the top two teams in the rankings with one loss. And there is where Texas has a big glaring advantage on California.
Let's look at those poll rankings again for Texas and Cal, but this time let's look at the whole top 6.
AP Poll
1 1603 USC
2 1541 OU
3 1536 Auburn
4 1413 California
5 1340 Utah
6 1323 Texas
USA Today/ESPN Poll
1 1510 USC
2 1440 OU
3 1436 Auburn
4 1299 California
5 1260 Texas
6 1246 Utah
Let's see by groupings where these votes are concentrated.
In the AP poll, all of the first place votes are worth a combined 1,625 points, 2nd place 1,560, 3rd place 1,495. Add those together and you get 4,680 points. USC, OU and Auburn combine for 4,680 points. So none of Cal, Utah and Texas has a ranking on any ballot above #4.
All of the 4th-6th place votes are worth a total of 4,095 points, and Cal, Utah and Texas have 4,076 of them, leaving only 19 total points for teams ranked below #6 from the 4-6 voting positions
In the USA Today/ESPN poll, those same 1st-3rd place points total 4,392 points. USC, OU and Auburn have a total of 4,386 points. That leaves 6 points in any form available for the teams ranked below them.
Those six points plus the 3,843 points that the 4th-6th place votes are worth totals 3,849 points, while Cal, Texas and Utah combine for 3,805 points. That is all but 44 of the total points available for the top 6 teams.
What that exercise is telling us is that almost all the available points that can affect Cal or Texas positively are going to come from the three team group of Utah, Texas and Cal. So what are the real implications of that? Here is what they are:
1. Utah's regular season is done; a bowl game is next.
2. Cal plays at mediocre Southern Miss on the final day of the season.
3. Texas plays against a ranked Texas A&M team on the day after Thanksgiving.
Quick question: Which team has by far the greatest upside here?
This isn't rocket science; it's poll science. A little bit voodoo, but mostly common sense. And we are certainly to the debatable portion of this issue. And if I am at fault in my analysis, this is where I will get attacked.
Texas has a long-standing reputation as a team that cannot get over the two-loss hump with Mack Brown. This is a widely held belief among the pollsters for sure; or else Texas would not have been slapped much more severely for losing to #2 Oklahoma than Cal was for losing to #1 USC. What has been worse for Texas is that they've been the team that can't beat the good teams, as their record in recent years against top 25 teams hasn't been good at all.
Things are different for Texas this year. Texas fought all the way to the end against OU, and although once again they lost, they fought to the last and held Oklahoma to 16 fewer points than any other opponent this year. They might have struggled in some other games, but so has Cal, and Utah has had a dramatically weaker schedule this year. If Texas was to beat A&M in solid fashion, how many points do you think they would steal away from Utah, and especially from Cal? And if they beat A&M by 21 points, the same amount that Utah beat A&M by, or by more points than that, how many points do you think they would take away from Utah and Cal?
Here is where my reasoning comes to fruition. Texas doesn’t need a signature win, but UT will be playing the second best team on their entire schedule on Friday. This will carry a strong impact with the voters with a solid win, by which I mean 14 points or more. More than 21 points and Texas will steal a LOT of points from both Cal and Utah. They will definitely take a significant number of points from both teams with any win of 7 or more points. This is just naturally to be expected when only Texas of that group of teams is playing this week. Positive news unanswered means positive vote pulls for the Horns, if they win with any margin.
The question is, will they take away enough points? And won't Cal have the opportunity to get those points back? I think that very clearly the answers are a resounding yes and yes. Let's look at the background to where we are a little more clearly.
Here are last week's point totals for Cal and Texas in the polls:
AP
Cal 1,409
Texas 1,301
Utah 1,316
USA Today/ESPN
Cal 1,311
Texas 1,222
Utah 1,203
On the most recent weekend, USC was idle, Oklahoma won 35-0 at Baylor, Auburn won 21-13 at Alabama, Cal stomped Stanford 31-6, Texas was idle, Utah drilled BYU 52-21, and most importantly Michigan, Florida State and Wisconsin all lost. That left some significant points to be picked up by the teams around and above them.
Here are this week's point totals in the polls:
AP
Cal 1,413
Texas 1,323
Utah 1,340
USA Today/ESPN
Cal 1,299
Texas 1,260
Utah 1,246
What! In the AP poll Cal gained 4 points, while Texas gained 22. And in the USA Today/ESPN poll, Cal LOST 12 points, while Texas GAINED 38! That is a net gain for Texas of 68 total points, on a weekend where Cal won by 25 and Texas didn't even play a game. Yes, Utah picked up 67 total points, to all but ensure a top 6 finish, when they blasted BYU 52-21. But for Cal it was a resounding no vote, a real indication that the voters are starting to question just how strong a conference really is with only 3 of 10 teams with more than six wins after having played a relatively weak schedule. And that schedule was one that they were able to navigate with only 7 more wins than losses. Compare that to the Big 12, which has won 20 more games than lost.
What do those numbers tell us? It tell us exactly this - that people are starting to believe that Texas is NOT a fluke this year, and that Cal may in fact be overrated.
This is the time of year, and especially in a year with three unbeaten teams at the top of the polls, that the pollsters are going to scrutinize the relative strengths more closely, and in doing so, they are seeing that Texas has played a stronger schedule than Cal. There is compensation going on for treating UT so poorly after their loss to Oklahoma, where they held the Sooners to 16 fewer points than the Sooners have scored on any other team.
If Texas can gain 68 points on Cal while idle and with Cal drilling Stanford 31-7, what is going to happen this week if Texas wins by 14 or more points while Cal is idle, and while Utah is done playing games?
I’ll tell you what will occur. Texas will add the second best win to their schedule, on national television in front of a large audience, and inherit the nation’s toughest strength of schedule from Texas A&M. And that will turn even more voters their way than last week. Try maybe 75-100 more total points in the two polls.
And we know that almost all of those points will be coming at the expense of Cal and Utah, not the unbeaten teams above, who are unassailable, or the two-loss teams below, who don’t have any appreciable points to steal anyway.
Let’s assume that Texas only climbs by 50 total points with a solid win over A&M on Friday. And let’s just assume for argument that they steal half of those points from both Cal and Utah.
This is where that kind of move would take them in the BCS ratings:
California .8542 .8354 .8200 .8365
Texas .8449 .8590 .8500 .8513
That’s a very significant lead, and that is being conservative in just about every way. I believe that it is very much more likely that the margin for Texas will be even larger in the final BCS rankings. Oklahoma may still add another win over a 7-4 Iowa State or Colorado to their opponent’s resume, while only USC at UCLA will remain for Cal’s opponents, and that game is a wash for Cal, since they played both teams.
Cal only has a very average Southern Miss team to play, which might actually hurt them rather than help them. The voters will have already discounted a strong win over a much weaker team in this week’s poll, and even if Cal blows out Southern Miss it will be largely unnoticed by voters in the mass of end of the season stories that vastly surpass that game in general interest.
In short, I believe that analysis says that almost without a doubt Texas will pass Cal on Friday with a solid win over Texas A&M, and they will stay ahead through the final BCS rankings.
Bye-bye Cal.
Thanks for reading to the end of this ridiculously long post. I had no idea it would be this long when I started.