This was originally posted on Orangebloods.com "Inside the 40 Acres," so it obviously focuses on the BCS from the perspective of Texas fans.
For some perspective to non-Texas fans, this was a difficult weekend for those of us that bleed burnt orange. I'm not going to try and spin it, I got the initial call wrong, and looking back, with the margin setting up to be excruciatingly close, I should have said it was too close to call (which ESPN's Brad Edwards wisely did). However, I was fairly clear once the Sagarin poll came out early Sunday morning (about 10 hours before the official BCS was announced) that Texas was in deep, deep trouble. So at least I "rescinded" my initial projection as soon as data came out to the contrary.
Please see the 2008 BCS Post Mortem for some thoughts on what made the call so challenging. The rest of this post includes the original comments I made that fateful Saturday night and Sunday morning.
Boy, it hurts to re-read it. Tough, tough weekend.
-- Jeff
Early Sunday Morning update
This is not a good morning {for Texas fans}. Sagarin is confirmed, and OU narrowly edged out Texas. I had hoped Texas could hold them off, but the SOS hits we took finally appear to have tweaked us. There are posters on this board that have used the Colley simulator to determine Texas will be #1 and OU #3. While I think that Texas will definitely be #1, I'm afraid OU will jump to #2 (another SOS tweak).
Wolfe is the key now, and unfortunately, someone on another message board seems to think that already came out pro-OU. I'm trying to confirm, usually that one doesn't come out, but given the way today is going, Wolfe going to OU would not surprise me. Still worried OU could jump Bama in Billingsley as well.
Best case for Texas now really is a push in the computers, and it looks like the higher probability is OU to have a 0.01 or 0.02 computer lead. If Wolfe falls pro-OU, you can count on at least a 0.02 computer lead.
I'm very disappointed. Anything can happen in the human polls, but this is looking like a real uphill battle. I will leave my original prediction for posterity's sake (below), and I'll have to stick with it, but it's not looking great. I think Texas is probably outside looking in, and Oklahoma is in much better shape.
Original Note from Saturday Night
1. Alabama
2. Texas
3. Oklahoma
4. Florida
This is what you call a "pick'em" in the BCS. This makes Cal/Texas look like a walk in the park. This may end up being comical as far as how small the margin will be between OU and Texas.
I will officially call the computers a push. I could see it go +0.01 Texas, and +0.01 OU. There's even some shot it could go +0.02 Texas (and even a remote shot it goes +0.02 OU). I wish SOS would have been more in our favor, but downsizing my computer projection down to a push seems fair to me.
So it's all on the human voters. And your guess is as good as mine. Seriously, your guess is as good as mine. I cannot caveat this projection enough. And I would not be surprised if Brad Edwards comes out tonight and says OU makes it to #2 (and he could be right).
Why do I think Texas will prevail?
1. OU did not get style points tonight. They got some last-minute garbage points, but not style points.
2. OU's defense was somewhat exposed, and Texas has held its last two opponents to single-digits.
3. Texas took care of business against A&M.
4. At least 80% of national pundits that have commented on the issue state that they think Texas should be ahead of OU.
5. Baylor gave Tech everything they could handle, taking a ton of luster off of OU's victory last week.
6. Tech moved themselves even more out of the discussion on the three way tie with the aforementioned struggle versus Baylor.
7. Anyone who gives OU a first place vote (and there were several last week), when Alabama went undefeated in the SEC, and Florida far more convincingly won their hostile road game, should seriously be called into question. Texas should gain some points as the Gators get promoted.
8. Texas beat OU by 10 points on a neutral field.
There are many good counter-arguments. I just think Texas has the best arguments. I don't know how many voters were even watching TV this week and hearing the media be pro-Texas. I don't know how willing they are to change their vote. I don't know what they thought when they watched the Bedlam game.
Couple of final thoughts:
a) do not take my projection as anything more than it is. This is a very hard call, and there is so much uncertainty on how human voters will behave. There is nothing scientific about this. Your guess is as good as mine. Please don't get your hopes up, this is just my guess.
b) again, this is far more difficult than Texas/Cal.
c) I've really appreciated all of the interest and support over the last few weeks.
d) Please don't kill me if I'm wrong. Because there is a good shot I am.
Could I have possibly caveated this projection any more? :)
BCS analysis for the college football fanatic
Sunday, November 30, 2008
Saturday, November 29, 2008
Friday night update before the big weekend
This was originally posted on Orangebloods.com "Inside the 40 Acres," so it obviously focuses on the BCS from the perspective of Texas fans.
Would've been nice to not have that 57 yarder go through the uprights, and yes, the CU/NU game was bigger than Ark/LSU because it was both a pro-Texas and anti-OU potential in one game. Ark/LSU was only potentially pro-Texas.
Still, when I looked at today's matchups, I really thought 0-3 was possible. Arkansas trailed pretty badly at one point, and that was quite a comeback (what a gutsy call on 4th and 1 as precious clock was burning down).
Tomorrow really does look better. If we can just get this weekend to be a wash on the strength-of-schedule front, it should keep the computers from deviating too far from our expectations. Remember to root for FAU, Syracuse, Rice and Mizzou. And obviously, root for Baylor and FSU. If you are feeling confident about the human vote, go ahead and root for OU, in a close one.
QUICK SUMMARY ON COMPUTERS
As long as we don't go 0-fer tomorrow, and there are no upsets, I continue to project a 0.01 Texas computer lead for the computers in the upcoming BCS (versus 0.06 now). The swing factors to watch will be:
1) Can Texas hold off OU in Sagarin (yikes, this'll be close, but my call is yes)
2) Can Texas hold off OU in Wolfe (another scary one, but my call is yes)
3) Will OU be one or two spots behind Texas in Colley (I think 2 spots, feel generally secure)
4) Is there any shot OU doesn't jump Bama in Billingsley (I think OU jumps them, to be conservative)
These are all so delicate. If one or more swing OU's way, they may have the slight computer lead. If they swing Texas' way, the lead may be 0.02 or 0.03.
IN CONCLUSION
Texas clearly did what they needed to last night (with "style"), and holding the last two opponents to single-digits suggests Texas looks more balanced than OU.
Texas has also gotten a lot of love from the media. There is no possible way to quantify the effect in the human polls. I have to believe that Texas will pick up some ground in the human polls, but we still don't know how dominant OU will be in their win over OSU... that is, if they beat OSU.
Remember with my computer projection, we need 10 coaches and 18 Harris voters (about 1/6 of each poll's voters) to be willing to flip Texas ahead of OU. If we get some breaks in the computers, it may only be 5 coaches and a dozen or so Harris voters.
My gut tells me if OU does not beat OSU by at least double-digits, Texas has a really good shot to pick up the votes. I won't "call my shot" until we see the final in Stillwater.
Boy oh boy, if we see a close OU victory, I won't be able to sleep tomorrow night.
Would've been nice to not have that 57 yarder go through the uprights, and yes, the CU/NU game was bigger than Ark/LSU because it was both a pro-Texas and anti-OU potential in one game. Ark/LSU was only potentially pro-Texas.
Still, when I looked at today's matchups, I really thought 0-3 was possible. Arkansas trailed pretty badly at one point, and that was quite a comeback (what a gutsy call on 4th and 1 as precious clock was burning down).
Tomorrow really does look better. If we can just get this weekend to be a wash on the strength-of-schedule front, it should keep the computers from deviating too far from our expectations. Remember to root for FAU, Syracuse, Rice and Mizzou. And obviously, root for Baylor and FSU. If you are feeling confident about the human vote, go ahead and root for OU, in a close one.
QUICK SUMMARY ON COMPUTERS
As long as we don't go 0-fer tomorrow, and there are no upsets, I continue to project a 0.01 Texas computer lead for the computers in the upcoming BCS (versus 0.06 now). The swing factors to watch will be:
1) Can Texas hold off OU in Sagarin (yikes, this'll be close, but my call is yes)
2) Can Texas hold off OU in Wolfe (another scary one, but my call is yes)
3) Will OU be one or two spots behind Texas in Colley (I think 2 spots, feel generally secure)
4) Is there any shot OU doesn't jump Bama in Billingsley (I think OU jumps them, to be conservative)
These are all so delicate. If one or more swing OU's way, they may have the slight computer lead. If they swing Texas' way, the lead may be 0.02 or 0.03.
IN CONCLUSION
Texas clearly did what they needed to last night (with "style"), and holding the last two opponents to single-digits suggests Texas looks more balanced than OU.
Texas has also gotten a lot of love from the media. There is no possible way to quantify the effect in the human polls. I have to believe that Texas will pick up some ground in the human polls, but we still don't know how dominant OU will be in their win over OSU... that is, if they beat OSU.
Remember with my computer projection, we need 10 coaches and 18 Harris voters (about 1/6 of each poll's voters) to be willing to flip Texas ahead of OU. If we get some breaks in the computers, it may only be 5 coaches and a dozen or so Harris voters.
My gut tells me if OU does not beat OSU by at least double-digits, Texas has a really good shot to pick up the votes. I won't "call my shot" until we see the final in Stillwater.
Boy oh boy, if we see a close OU victory, I won't be able to sleep tomorrow night.
Tuesday, November 25, 2008
BCS Projection/Analysis Week 11
This was originally posted on Orangebloods.com "Inside the 40 Acres," so it obviously focuses on the BCS from the perspective of Texas fans.
PROJECTION SUMMARY
With all of the data now available, I'm going to move in a slight more pro-Texas direction with my computer projections. While I think OU will move up in Colley Matrix, I don't think they will move to #2 as I previously expected. I think Florida will remain ahead of OU, and keep OU down to #3. I am also more comfortable with Texas holding on to its advantage in the Wolfe computer (now that the data has been disclosed). Sagarin remains a wildcard, and frankly, if that one turns on us, a lot comes unhinged. I am still sticking with Elo-chess keeping Texas ahead of OU, and fortunately, we'll know that result fairly early on Sunday morning. If that one turns on us, we have problems.
My base case is now for Texas to hold a 0.01 computer advantage, versus my previous forecast for a dead even computer score. This sounds like a trivial number, but that basically means Texas will only have to convert 10 coaches and 18 Harris poll voters to switch Texas and OU in order to stay ahead of OU in next week's BCS standings.
FLORIDA LOSS TO FSU?
I've also had more time to study the potential impact of a Florida loss on the BCS rankings. Given that OU likely outranks Florida and Texas on several ballots (especially in the coaches poll), all else equal, a Florida loss should result in more relative benefit to Texas than Oklahoma.
One may ask why? If a voter had the following ballot: 1) Bama 2) OU 3) Fla 4) Tex, Texas trails OU by two points. If Florida loses, the same ballot (assuming no funny business) should change to: 1) Bama 2) OU 3) Tex, whereby Texas only trails by one point. Imagine this phenomenon occuring on several different ballots, and it can start to add up.
There's a lot of guesswork in this exercise, but I estimate that a Florida loss could shave OU's relative point advantage in the coaches poll from the current 42 to about 26, and for the Harris poll, from the current 21 to about 15.
What would this mean? If you believe my base case for the computers (and I don't blame you if you don't given all of the uncertainty), then Texas would only need to convert 5 coaches and 9 Harris voters to switch. And if you believe my bull case (where OU doesn't jump Bama in the Billingsley computer) you would only need one coach and 2 Harris voters to switch Texas ahead of OU.
I don't want to get people too excited. The chances of FSU beating Florida seem unfavorable. And you shouldn't view an FSU win as a solution by itself. You should view it as a steroid for the system - it can help Texas a lot, but by itself, it can't keep Texas #2. Texas, in some shape or form, is going to need help from the humans.
SUMMARY THOUGHTS
I am very encouraged by the strong support UT is getting in the media. I was disappointed that ESPN's pro-Texas sentiment late Saturday night did not help Texas in the polls that came out Sunday morning, but perhaps there was not enough time for the arguments to be made in the voters' minds. They will have all week to be reminded that Texas beat OU head to head on a neutral field by 10 points, and then more importantly, they will get to see each team audition for votes.
I think this thing is going to be close. I regret to say it leans in OU's favor, but there are still plenty of variables remaining. If you tell me Texas beats A&M by 3 or more TDs, and OU only beats OSU by 6 points, I think it will be incredibly, incredibly interesting.
Do not forget key games this week. We need to at least tread water on strength-of-schedule relative to OU.
Pro Texas/Pro OU (likelihood)
Colorado/Nebraska (unlikely)
Missouri/Kansas (likely)
Syracuse/Cincy (unlikely)
FAU/FIU (likely)
UTEP/ECU (unlikely)
Rice/Houston (good shot)
Ark/LSU (push)
------DETAIL OF MY COMPUTER FORECAST SCENARIOS------
Base case (55%): 0.01 computer lead for Texas
Human shift needed: 10 coaches and 18 Harris voters switch Tx/OU
Bull case (30%): 0.02 computer lead for Texas
Key change: RB breaks our way, and OU doesn't jump Bama (they get 24 rather than 25)
Human shift needed: 5 coaches and 12 Harris voters switch TX/OU
Bear case (15%): 0.02 computer DEFICIT for Texas
Key changes: OU jumps Florida in CM (OU gets 24, not 23), and JS reverses Texas and OU (OU gets 25, Texas gets 24)
Human shift needed: 21 coaches and 40 Harris voters switch TX/OU
Original from Saturday night
1. Alabama
2. Florida
3. Texas
4. Oklahoma
The margin betweeen #2 Florida and #3 Texas may be razor thin, and Texas may edge out Florida to remain #2.
I put the most pro-OU human vote I could into my analysis, and couldn't quite get OU to jump Texas. But it could also be damn close.
I project a 0.96 computer score for Texas, and a 0.90 computer score for OU. There's a chance OU could make up more ground than that. Today was a disaster for Texas in the computer polls. Texas' unique opponents compared to OU went a lousy 1-3 (Rice won, FAU/UTEP/Ark lost), while OU's opponents did a much better 3-2 (TCU/Cincy/KSU won, Wash/Chat lost).
Even worse, this 0.06 advantage for Texas is likely to be compromised further next week. Bear in mind, my projection for the ultimate BCS standings just before the Big XII championship is seeded was for Texas to hold a 0.03 edge in the computers. Frankly, I don't know where that stands, and after such a difficult day, it's probably best to wait and see how the precise computer scores and human votes look tomorrow.
Today couldn't have gone much worse. Fortunately, there is one week left. Let's just hope we keep some goodwill with the voters tonight. But that's quite a thumping they watched tonight.
I am curious what you all think. The human vote is such a big swing factor. Does OU killing Tech so badly, in some ways work against them, and diminish Tech as a third-tier opponent, and get people fixated back on the head-to-head matchup? Or do people really think this OU team is "different" than the one that played Texas on a neutral field in October?
I would have told you Texas had the edge coming into today, but that edge seems to have been lost. Don't get me wrong, it's not impossible. But today could not have gone much worse.
PROJECTION SUMMARY
With all of the data now available, I'm going to move in a slight more pro-Texas direction with my computer projections. While I think OU will move up in Colley Matrix, I don't think they will move to #2 as I previously expected. I think Florida will remain ahead of OU, and keep OU down to #3. I am also more comfortable with Texas holding on to its advantage in the Wolfe computer (now that the data has been disclosed). Sagarin remains a wildcard, and frankly, if that one turns on us, a lot comes unhinged. I am still sticking with Elo-chess keeping Texas ahead of OU, and fortunately, we'll know that result fairly early on Sunday morning. If that one turns on us, we have problems.
My base case is now for Texas to hold a 0.01 computer advantage, versus my previous forecast for a dead even computer score. This sounds like a trivial number, but that basically means Texas will only have to convert 10 coaches and 18 Harris poll voters to switch Texas and OU in order to stay ahead of OU in next week's BCS standings.
FLORIDA LOSS TO FSU?
I've also had more time to study the potential impact of a Florida loss on the BCS rankings. Given that OU likely outranks Florida and Texas on several ballots (especially in the coaches poll), all else equal, a Florida loss should result in more relative benefit to Texas than Oklahoma.
One may ask why? If a voter had the following ballot: 1) Bama 2) OU 3) Fla 4) Tex, Texas trails OU by two points. If Florida loses, the same ballot (assuming no funny business) should change to: 1) Bama 2) OU 3) Tex, whereby Texas only trails by one point. Imagine this phenomenon occuring on several different ballots, and it can start to add up.
There's a lot of guesswork in this exercise, but I estimate that a Florida loss could shave OU's relative point advantage in the coaches poll from the current 42 to about 26, and for the Harris poll, from the current 21 to about 15.
What would this mean? If you believe my base case for the computers (and I don't blame you if you don't given all of the uncertainty), then Texas would only need to convert 5 coaches and 9 Harris voters to switch. And if you believe my bull case (where OU doesn't jump Bama in the Billingsley computer) you would only need one coach and 2 Harris voters to switch Texas ahead of OU.
I don't want to get people too excited. The chances of FSU beating Florida seem unfavorable. And you shouldn't view an FSU win as a solution by itself. You should view it as a steroid for the system - it can help Texas a lot, but by itself, it can't keep Texas #2. Texas, in some shape or form, is going to need help from the humans.
SUMMARY THOUGHTS
I am very encouraged by the strong support UT is getting in the media. I was disappointed that ESPN's pro-Texas sentiment late Saturday night did not help Texas in the polls that came out Sunday morning, but perhaps there was not enough time for the arguments to be made in the voters' minds. They will have all week to be reminded that Texas beat OU head to head on a neutral field by 10 points, and then more importantly, they will get to see each team audition for votes.
I think this thing is going to be close. I regret to say it leans in OU's favor, but there are still plenty of variables remaining. If you tell me Texas beats A&M by 3 or more TDs, and OU only beats OSU by 6 points, I think it will be incredibly, incredibly interesting.
Do not forget key games this week. We need to at least tread water on strength-of-schedule relative to OU.
Pro Texas/Pro OU (likelihood)
Colorado/Nebraska (unlikely)
Missouri/Kansas (likely)
Syracuse/Cincy (unlikely)
FAU/FIU (likely)
UTEP/ECU (unlikely)
Rice/Houston (good shot)
Ark/LSU (push)
------DETAIL OF MY COMPUTER FORECAST SCENARIOS------
Base case (55%): 0.01 computer lead for Texas
Human shift needed: 10 coaches and 18 Harris voters switch Tx/OU
TEAM AH RB CM KM JS PW Score
TEX 23 22 25 24 25 24 0.96
OU 22 25 23 25 24 23 0.95
Bull case (30%): 0.02 computer lead for Texas
Key change: RB breaks our way, and OU doesn't jump Bama (they get 24 rather than 25)
Human shift needed: 5 coaches and 12 Harris voters switch TX/OU
Bear case (15%): 0.02 computer DEFICIT for Texas
Key changes: OU jumps Florida in CM (OU gets 24, not 23), and JS reverses Texas and OU (OU gets 25, Texas gets 24)
Human shift needed: 21 coaches and 40 Harris voters switch TX/OU
Original from Saturday night
1. Alabama
2. Florida
3. Texas
4. Oklahoma
The margin betweeen #2 Florida and #3 Texas may be razor thin, and Texas may edge out Florida to remain #2.
I put the most pro-OU human vote I could into my analysis, and couldn't quite get OU to jump Texas. But it could also be damn close.
I project a 0.96 computer score for Texas, and a 0.90 computer score for OU. There's a chance OU could make up more ground than that. Today was a disaster for Texas in the computer polls. Texas' unique opponents compared to OU went a lousy 1-3 (Rice won, FAU/UTEP/Ark lost), while OU's opponents did a much better 3-2 (TCU/Cincy/KSU won, Wash/Chat lost).
Even worse, this 0.06 advantage for Texas is likely to be compromised further next week. Bear in mind, my projection for the ultimate BCS standings just before the Big XII championship is seeded was for Texas to hold a 0.03 edge in the computers. Frankly, I don't know where that stands, and after such a difficult day, it's probably best to wait and see how the precise computer scores and human votes look tomorrow.
Today couldn't have gone much worse. Fortunately, there is one week left. Let's just hope we keep some goodwill with the voters tonight. But that's quite a thumping they watched tonight.
I am curious what you all think. The human vote is such a big swing factor. Does OU killing Tech so badly, in some ways work against them, and diminish Tech as a third-tier opponent, and get people fixated back on the head-to-head matchup? Or do people really think this OU team is "different" than the one that played Texas on a neutral field in October?
I would have told you Texas had the edge coming into today, but that edge seems to have been lost. Don't get me wrong, it's not impossible. But today could not have gone much worse.
Sunday, November 23, 2008
BCS Projection/Analysis Week 10
This was originally posted on Orangebloods.com "Inside the 40 Acres," so it obviously focuses on the BCS from the perspective of Texas fans.
Conclusion First
Texas' chances to win the Big XII took it on the chin this week. OU gained a staggering 59 points on Texas in the coaches poll, and an also stunning 122 points in the Harris poll. Texas' current 0.06 advantage in the computers is the only reason they remain ahead of OU in the BCS. Unfortunately, I think that most (if not all) of this computer advantage will be going away when next weekend's action is completed. Texas therefore needs to pick up meaningful votes in the human polls if they want to win the three way tie. I'm encouraged that several in the media are on Texas' "side", but I don't know if that could possibly translate into anything in the polls to move the meter. If OU and Texas both take care of business, I hate to say it, but OU will probably be going to KC. I believe about 1/4 of the voters need to switch Texas and OU for Texas to get the edge, and that's not likely unless OU really struggles against OSU, and there is a groundswell of support for Texas. I'm rooting for OSU, and I think most people on this board have already figured out that is the right thing to root for at this stage.
All is not lost, there are many other possibilities:
1) What if Baylor beats Tech? Texas controls its destiny (without question).
2) What if FSU beats Florida? Many votes are suddenly thrown into the mix, and Texas is a much stronger beneficiary. Let's just say, it would be a lot more interesting.
3) What if OSU beats OU (after all, it is in Stillwater)? Texas is a heavy favorite to go to the NC championship (regardless of what Tech does)
4) Even if OU beats OSU, what if they lose to Missouri? Texas is almost guaranteed to go to the national championship.
5) What if UF loses to FSU, but beats Bama? Texas is a strong contender to play OU in the NC.
6) What if Bama loses to Auburn, but beats Florida? Texas is a solid contender to play OU in the NC.
Still lots of interesting possibilities. And the 3-way tie, while challenging, is not impossible. Texas needs to take care of business against A&M. We'll see where it goes from there.
Detail on BCS Forecast
This is an early update, as we don't have concrete numbers from A&H, KM or PW computer polls yet. That said, I regret to say that if one wants to be realistic/conservative about this, we should probably kiss the computer advantage good-bye (my base case below).
Base case: Texas and OU have identical computer average. This means Texas needs to convert 13 coaches and 26 Harris voters (about 25% of all voters) to switch OU and Texas on their ballots.
Texas' best-case: Texas has 0.02 computer lead over OU (get key wedges in a couple of important polls, and hang on to leads in A&H, JS, CM and PW). We'd only need to convert 5 coaches and 12 Harris voters if this happened.
Texas' worst case: a couple more computers turn toward OU. Texas suddenly has a 0.03 DEFICIT. We'd have to convert 24 coaches and 48 Harris voters (getting close to half of all voters). Basically, Texas would be screwed.
I've included a full appendix discussing the computer polls, and I'd really welcome any discussion on how people think these shake out. I'll update the discussion and projections if we see anything jump out on the computers that have yet to publish.
WHY IS THIS HAPPENING?
On the computer side, OU gaining on Texas was inevitable. In fact, I now believe OU's schedule is just as strong as Texas. Unfortunately, things took a turn for the worst this weekend when Texas' opponents went a lousy 1-3, while OU's Div I-A opponents went 3-1 (counting KSU). I am also being conservative with expecations because I think there is an uphill battle on this weekend's "battleground" games. I detail these below:
Project Tex opp 3-3
FAU vs. FIU (give it to FAU)
UTEP at ECU (yuck... ECU wins)
Ark vs. LSU (LSU's imploding, but Ark is playing poorly... LSU wins)
Rice vs. UH (UH is dangerous, but hope Rice can win)
CU at Neb (yuck... Neb wins)
Mizz vs. KU (Mizz should win)
Project OU opp 2-0
Cin vs. Syra (be realistic, Cin wins)
Neb vs. CU (yuck... Neb wins)
-------------COMPUTER APPENDIX, FOR THOSE WHO CARE-------------
Here's how I get to a dead even (0.96 vs. 0.96) computer score for Texas and Oklahoma. (* means data not yet available)
Team *AH RB CM *KM JS *PW
Texas 23 22 25 24 25 24
Okla 22 25 24 25 24 23
Poll level detail for those who care:
Anderson & Hester (VULNERABLE)
This has been a fairly stable poll, but we don't have the latest version yet to scrutinize how much ground OU has made on Texas. Bama and Utah have been entrenched at the top for a while on this one. For now, my temptation is to leave it static. 25--Bama 24--Utah 23--Tex 22--OU
Richard Billingsley
Not getting too scientific here. You can see the type of gains teams register when they beat a certain caliber opponent. OU picked up lots of points when they beat Tech, and they'll pick up a bunch more if they beat OSU. While OU already leads Texas as we speak, my fear is that they will jump Alabama. Even worse, I'm afraid Florida may sneak up from the rear and jump Texas. That's two wedges, and that sucks. It's going to be close on Florida, but to be conservative, I project: 25--OU 24--Bama 23--Fla 22-Tex
Colley Matrix
Texas has had a lock on this rating's top spot for several weeks, and although OU is gaining fast, I think Texas has a big enough lead to remain #1. Unfortunately, I think OU will jump Bama, Florida and Utah, taking away what had been a big gap of advantage for us. 25--Tex 24--OU 23--Fla 22--Bama 21--Utah
Kenneth Massey
I don't even see the point of studying this one too much, as OU has already taken the top spot, and Texas is #2. Texas simply can't play catch-up with OU in any poll, so I think it's a given OU stays #1. With TTU directly behind us, I'm not too worried about a possible jump by anyone, but we don't have the data yet, so we'll keep this one statick. 25--OU 24--TX 23--TTU 22--Ala 21--Utah
Jeff Sagarin (VULNERABLE)
OU picked up major ground in this one (remember Elo-Chess is used--similar to the rating system of chess players). It doesn't purportedly take into account recency of the game. Texas has held up better week by week in this poll than I expect, and I am going to take a leap of faith that OU can't replicate the gains they had last week because a) they are playing a lesser opponent, and b) OU is now higher rated, so it becomes less "unusual" for them to beat a top 12 team at this point. Who knows, maybe Texas can get a wedge here (you never know). If this one doesn't work in our favor, a lot comes unhinged. 25--Tex 24--OU 23--Ala 22--TTU
Peter Wolfe (VULNERABLE)
Without the data, can't comment too much on how strong Texas' lead over OU is. But I don't like Tech as a wedge between Texas and OU. At a minimum, OU will probably jump Tech. 25--Ala 24--Tex 23--OU 22-TTU 21-Utah
Conclusion First
Texas' chances to win the Big XII took it on the chin this week. OU gained a staggering 59 points on Texas in the coaches poll, and an also stunning 122 points in the Harris poll. Texas' current 0.06 advantage in the computers is the only reason they remain ahead of OU in the BCS. Unfortunately, I think that most (if not all) of this computer advantage will be going away when next weekend's action is completed. Texas therefore needs to pick up meaningful votes in the human polls if they want to win the three way tie. I'm encouraged that several in the media are on Texas' "side", but I don't know if that could possibly translate into anything in the polls to move the meter. If OU and Texas both take care of business, I hate to say it, but OU will probably be going to KC. I believe about 1/4 of the voters need to switch Texas and OU for Texas to get the edge, and that's not likely unless OU really struggles against OSU, and there is a groundswell of support for Texas. I'm rooting for OSU, and I think most people on this board have already figured out that is the right thing to root for at this stage.
All is not lost, there are many other possibilities:
1) What if Baylor beats Tech? Texas controls its destiny (without question).
2) What if FSU beats Florida? Many votes are suddenly thrown into the mix, and Texas is a much stronger beneficiary. Let's just say, it would be a lot more interesting.
3) What if OSU beats OU (after all, it is in Stillwater)? Texas is a heavy favorite to go to the NC championship (regardless of what Tech does)
4) Even if OU beats OSU, what if they lose to Missouri? Texas is almost guaranteed to go to the national championship.
5) What if UF loses to FSU, but beats Bama? Texas is a strong contender to play OU in the NC.
6) What if Bama loses to Auburn, but beats Florida? Texas is a solid contender to play OU in the NC.
Still lots of interesting possibilities. And the 3-way tie, while challenging, is not impossible. Texas needs to take care of business against A&M. We'll see where it goes from there.
Detail on BCS Forecast
This is an early update, as we don't have concrete numbers from A&H, KM or PW computer polls yet. That said, I regret to say that if one wants to be realistic/conservative about this, we should probably kiss the computer advantage good-bye (my base case below).
Base case: Texas and OU have identical computer average. This means Texas needs to convert 13 coaches and 26 Harris voters (about 25% of all voters) to switch OU and Texas on their ballots.
Texas' best-case: Texas has 0.02 computer lead over OU (get key wedges in a couple of important polls, and hang on to leads in A&H, JS, CM and PW). We'd only need to convert 5 coaches and 12 Harris voters if this happened.
Texas' worst case: a couple more computers turn toward OU. Texas suddenly has a 0.03 DEFICIT. We'd have to convert 24 coaches and 48 Harris voters (getting close to half of all voters). Basically, Texas would be screwed.
I've included a full appendix discussing the computer polls, and I'd really welcome any discussion on how people think these shake out. I'll update the discussion and projections if we see anything jump out on the computers that have yet to publish.
WHY IS THIS HAPPENING?
On the computer side, OU gaining on Texas was inevitable. In fact, I now believe OU's schedule is just as strong as Texas. Unfortunately, things took a turn for the worst this weekend when Texas' opponents went a lousy 1-3, while OU's Div I-A opponents went 3-1 (counting KSU). I am also being conservative with expecations because I think there is an uphill battle on this weekend's "battleground" games. I detail these below:
Project Tex opp 3-3
FAU vs. FIU (give it to FAU)
UTEP at ECU (yuck... ECU wins)
Ark vs. LSU (LSU's imploding, but Ark is playing poorly... LSU wins)
Rice vs. UH (UH is dangerous, but hope Rice can win)
CU at Neb (yuck... Neb wins)
Mizz vs. KU (Mizz should win)
Project OU opp 2-0
Cin vs. Syra (be realistic, Cin wins)
Neb vs. CU (yuck... Neb wins)
-------------COMPUTER APPENDIX, FOR THOSE WHO CARE-------------
Here's how I get to a dead even (0.96 vs. 0.96) computer score for Texas and Oklahoma. (* means data not yet available)
Team *AH RB CM *KM JS *PW
Texas 23 22 25 24 25 24
Okla 22 25 24 25 24 23
Poll level detail for those who care:
Anderson & Hester (VULNERABLE)
This has been a fairly stable poll, but we don't have the latest version yet to scrutinize how much ground OU has made on Texas. Bama and Utah have been entrenched at the top for a while on this one. For now, my temptation is to leave it static. 25--Bama 24--Utah 23--Tex 22--OU
Richard Billingsley
Not getting too scientific here. You can see the type of gains teams register when they beat a certain caliber opponent. OU picked up lots of points when they beat Tech, and they'll pick up a bunch more if they beat OSU. While OU already leads Texas as we speak, my fear is that they will jump Alabama. Even worse, I'm afraid Florida may sneak up from the rear and jump Texas. That's two wedges, and that sucks. It's going to be close on Florida, but to be conservative, I project: 25--OU 24--Bama 23--Fla 22-Tex
Colley Matrix
Texas has had a lock on this rating's top spot for several weeks, and although OU is gaining fast, I think Texas has a big enough lead to remain #1. Unfortunately, I think OU will jump Bama, Florida and Utah, taking away what had been a big gap of advantage for us. 25--Tex 24--OU 23--Fla 22--Bama 21--Utah
Kenneth Massey
I don't even see the point of studying this one too much, as OU has already taken the top spot, and Texas is #2. Texas simply can't play catch-up with OU in any poll, so I think it's a given OU stays #1. With TTU directly behind us, I'm not too worried about a possible jump by anyone, but we don't have the data yet, so we'll keep this one statick. 25--OU 24--TX 23--TTU 22--Ala 21--Utah
Jeff Sagarin (VULNERABLE)
OU picked up major ground in this one (remember Elo-Chess is used--similar to the rating system of chess players). It doesn't purportedly take into account recency of the game. Texas has held up better week by week in this poll than I expect, and I am going to take a leap of faith that OU can't replicate the gains they had last week because a) they are playing a lesser opponent, and b) OU is now higher rated, so it becomes less "unusual" for them to beat a top 12 team at this point. Who knows, maybe Texas can get a wedge here (you never know). If this one doesn't work in our favor, a lot comes unhinged. 25--Tex 24--OU 23--Ala 22--TTU
Peter Wolfe (VULNERABLE)
Without the data, can't comment too much on how strong Texas' lead over OU is. But I don't like Tech as a wedge between Texas and OU. At a minimum, OU will probably jump Tech. 25--Ala 24--Tex 23--OU 22-TTU 21-Utah
Thursday, November 20, 2008
General BCS update
This was originally posted on Orangebloods.com "Inside the 40 Acres," so it obviously focuses on the BCS from the perspective of Texas fans.
I've spent a few hours studying the latest computer polls, and I'm trying to look ahead to the final BCS before the Big XII title game would be potentially "seeded."
The ultimate question is: what would be the relative advantage (if any) Texas would have in the computers over OU, given that OU will make up some considerable ground?
Right now, I project Texas will hold a 0.03 computer advantage (current advantage is 0.11, to give context). If anyone has specific questions on individual polls, definitely ask, and I can walk you through my logic, but here's the high-level takeaways:
I think Texas is going to get slaughtered in Billingsley, and that will definitely be the low thrown out for Texas, and the high thrown out for OU. Andersen Hester is really tough to call, but I think Texas may keep the edge, with Florida serving as a wedge.
The other four polls seem to favor Texas, especially the ones that model out all of the teams (not just Division 1A). It's stunning to see how poorly Washington and Chattanooga rank in those more expanded rankings, and I think that will be an albatross for OU even as their schedule strengthens quite a bit.
Texas still looks to have the strongest SOS in the country when all is said and done, and while OU will close the gap, I do think about a 0.03 computer poll advantage to Texas will hold out. Clearly, these rankings are sensitive, and there will be some key games that may help tweak it, but 0.03 seems like the right number.... for now.
So what does that mean?
I think a 0.03 computer lead is really tough for OU to overcome. Essentially, using the current human vote lead for Texas as the reference point, you would need to see 30 coaches and 54 Harris voters reverse their vote and put OU ahead of Texas. Not one or the other, but both.
To give you an idea, that is half of the total 61 coaches and the 113 Harris voters that participate in the vote. Another way to think about it, since many on this board look at the "points," this would mean OU would need both a 60 point and 108 point gain in coaches poll and Harris poll, respectively.
Texas this last week picked up 31 points in the coaches poll and 57 points in the Harris poll, so OU would basically have to reverse that move, and then reverse it again.
I think it's safe to say that the odds favor Texas, "all else equal."
What could destroy this projection?
If OU slips one notch higher in a computer, it could cut the 0.03 computer lead to 0.02, and make the human vote tighter.
The more problematic outcome could be if OU absolutely throttles Tech and OSU, and there is a groundswell of support where voters don't just reverse Texas and OU, but voters also put OU ahead of Florida. Or there is a more unanimous consensus that OU really is "better."
This could be the most fascinating BCS rankings since 2004, when Cal and Texas were duking out for the at large. Bear in mind, while the human vote actually favored Cal that year, it was the computers that spoke loud and clear and overruled the humans. We may see a similar phenomenon work in Texas' favor this year.
Then again, after spending all this time looking at it and talking about it, Murphy's law says Tech is going to beat OU and make the whole dialogue irrelevant.
I've spent a few hours studying the latest computer polls, and I'm trying to look ahead to the final BCS before the Big XII title game would be potentially "seeded."
The ultimate question is: what would be the relative advantage (if any) Texas would have in the computers over OU, given that OU will make up some considerable ground?
Right now, I project Texas will hold a 0.03 computer advantage (current advantage is 0.11, to give context). If anyone has specific questions on individual polls, definitely ask, and I can walk you through my logic, but here's the high-level takeaways:
I think Texas is going to get slaughtered in Billingsley, and that will definitely be the low thrown out for Texas, and the high thrown out for OU. Andersen Hester is really tough to call, but I think Texas may keep the edge, with Florida serving as a wedge.
The other four polls seem to favor Texas, especially the ones that model out all of the teams (not just Division 1A). It's stunning to see how poorly Washington and Chattanooga rank in those more expanded rankings, and I think that will be an albatross for OU even as their schedule strengthens quite a bit.
Texas still looks to have the strongest SOS in the country when all is said and done, and while OU will close the gap, I do think about a 0.03 computer poll advantage to Texas will hold out. Clearly, these rankings are sensitive, and there will be some key games that may help tweak it, but 0.03 seems like the right number.... for now.
So what does that mean?
I think a 0.03 computer lead is really tough for OU to overcome. Essentially, using the current human vote lead for Texas as the reference point, you would need to see 30 coaches and 54 Harris voters reverse their vote and put OU ahead of Texas. Not one or the other, but both.
To give you an idea, that is half of the total 61 coaches and the 113 Harris voters that participate in the vote. Another way to think about it, since many on this board look at the "points," this would mean OU would need both a 60 point and 108 point gain in coaches poll and Harris poll, respectively.
Texas this last week picked up 31 points in the coaches poll and 57 points in the Harris poll, so OU would basically have to reverse that move, and then reverse it again.
I think it's safe to say that the odds favor Texas, "all else equal."
What could destroy this projection?
If OU slips one notch higher in a computer, it could cut the 0.03 computer lead to 0.02, and make the human vote tighter.
The more problematic outcome could be if OU absolutely throttles Tech and OSU, and there is a groundswell of support where voters don't just reverse Texas and OU, but voters also put OU ahead of Florida. Or there is a more unanimous consensus that OU really is "better."
This could be the most fascinating BCS rankings since 2004, when Cal and Texas were duking out for the at large. Bear in mind, while the human vote actually favored Cal that year, it was the computers that spoke loud and clear and overruled the humans. We may see a similar phenomenon work in Texas' favor this year.
Then again, after spending all this time looking at it and talking about it, Murphy's law says Tech is going to beat OU and make the whole dialogue irrelevant.
Subscribe to:
Posts (Atom)