Donovan, It seems Jimbo Fisher agrees with you...

[Spence]

On SOS.

http://espn.go.com/college-football/story/_/id/8624250/florida-state-seminoles-jimbo-fisher-wants-move-back-pre-bcs-system

[donovan]

I chuckled at the way the direction of this went. I think when you step back and look at the whole system, where we came from and where we are now, it is nuts. I suspect, as has been said here many times, we probably are where we are not because of anyone really thinking this is better, but because it generates more money.

[Spence]

I agree. The problem with SOS is there isn't enough information to fairly evaluate teams. If you are a good team who happens to play a few team that aren't very good, know one really knows the strength of the team.

[billybud]

What Jimbo objected to were the computer algorithmns...while opaque, they are difficult to puzzle though...

Sagarin admits that he would rank FSU about 20 points higher if he used his more valid algrithmn....

The discepancy between the human polls (Coaches-Harris) that both have FSU #6 and the computers is vast. What isn't known is how the polls treat FCS..and FSU, because of WVU's and Air Force's cancellations had to schedule two very weak teams.

Jimbo notes that Ole Miss is 5-5, has zip for good wins...actually zip for teams over .500

FCS Central Arkansas
2-8 UTEP
2-8 Tulane
2-8 Auburn
4-6 Arkansas

YET..Ole Miss is rated higher. It appears that Ole Miss is rated higher on who thy lost to than who they beat.

But Vegas knows who is who....Todd Fuhrman (used to run Caesar's Palace) has FSU power ranked over all but Alabama and Oregon...has Notre Dame and KSU as being dogs on a neutral site.

But...FSU only has one "good" win and a weak schedule...so it is what it is....But Jimbo is coach and he isn't a computer guy....

[billybud]

And...how many of us are watching these SEC matches this weekend?

Western Carolina at Alabama

Georgia Southern at Georgia

Wofford at South Carolina

Jacksonville State at Florida

Alabama A & M at Auburn

Samford at Kentucky

Sam Houston State at Texas A & M

[Eric]

I'm willing to give the SEC a pass here. They seem to start conference play earlier than most conferences, so this is similar to their opening week. I think it's an easy way to give most teams an easy Senior Day.

[Spence]

It also gives them a better TV presence early in the season and helps their teams up the top 25 polls early, when you can make big moves. It is a pretty smart strategy.

[billybud]

It really gives them an off day before rivalry weekend...before Florida plays FSU, Aggies play Mizzou,Alabama plays Auburn, Georgia plays GT, etc.

[billybud]

Phil Steele, who is a huge proponent of the use of computers in evaluating sports and somebody who uses 9 different computer models that he is constantly adjusting, agreed whole heartedly with Jimbo.

He said flat out, any computer that produces a result of 5-5 Ole Miss being above FSU is clearly flawed and cannot be relied upon. Also noted those same computers have I think 4 teams with 4 losses above us and a second 5 loss team above us. There are two computers with those results.

He said he thinks next year, even with it being the last year the computers will be used, that:

1. Margin of victory should be returned to the formula, but capping it at 21 points. I.e. winning by 50 is the same as winning by 21. The idea being that prevents teams from trying to run up the score and also prevents inherent favortism towards teams that play a specific style (think Oregon's style vs. Bama's style).

2. The two computer polls that currently have FSU ranked behind 5 loss teams should be forced to change their formulas to produce more accurate results or should be removed completely from the BCS.

And this is coming from somebody who really likes the use of computers in sports. Pretty telling.

Boise hated the computers....and I do not, but I do question a couple of algorithmns.

[Spence]

I agree, because I have seen both FSU and Ole Miss play, that FSU is a better team. But, as you have said in the past, a 5 win Mississippi could be better than a one loss or undefeated team. Whether or not a computer is accurate or not does not necessarily have anything to do with their win/loss record or strength of schedule. Florida State is getting hammered in some rankings because of the FCS games. Granted Florida State didn't intent to have two FCS teams, but that is beside the point. If Florida State was Wyoming then the computers would likely hammer them for overall schedule, but the voters would also hammer them because they are Wyoming. The humans would use their schedule as proof they shouldn't be ranked higher. It is a different scenerio altogether with Florida State, though, because FSU is a traditionally strong program with a historical record of playing and beating many great teams in the past. The human voters tend to put them higher in the polls - deserved or not - because of that. Same has happened to Ohio State and lots of other traditionally good football programs. Whether or not Florida State is a great or good team doesn't matter. That is the irony of the SOS debate. We all know SOS should be applied in some way as a way to compare teams. What we don't know is the method because it is impossible to compare a team's strength when they play in a conference that doesn't have a couple national powerhouse teams in it. SOS is used as a weapon against non traditional powers.

What makes this funny is that rarely does this argument ever apply to a major power. Vandy ducked out of a game with next year with Ohio State only a couple weeks ago. Ohio State was able to schedule San Diego State in their place, but they could have been in the same position pretty easily. If no one had the game open Ohio State would have been forced to take an FCS team. San Diego state probably isn't going to be quite as good as Vandy, but it is an OK sub.

I don't agree that we should adjust computer models because they don't represent what we believe to be true. I think they should be judged base on results. It is possible that a 5-5 team could be better than a one loss team. Maybe not in this particular case, but the only way to know that for sure is to play and that probably isn't going to happen this year.

[donovan]

I do think I get it. I have a lot of references that say I don't. Boise did not like computers, because they did not favor them. I get that. Humans in their mind do exactly what computers do. The argument, rightfully made is computers will be consistent with the programing they have and humans, thankfully, do not bear that burden. So the question is asked, why doesn't the BCS or SEC or ESPN or the Libertarians, set the criteria that is to be used, how it should be be weighted. The discussion is absolutely valid. But no, they rely on individuals to give them the end result without having a clue of what they use and end up where they are. Remember Mr. Congrove was blackballed from the BCS because he was not willing to change his model to conform to whatever the BCS wanted. That's not right and will never be right.

I suspect when the NC game is formulated, it will be the one that gets the most exposure and brings in the most money. Notre Dame, and they may be the best team, I don't know, if they go undefeated..there is the money machine. If Oregon and Kansas State stub their toe, not necessarily even lose, Bama is back in. That's the money game.

[Derek]

And this is coming from somebody who really likes the use of computers in sports. Pretty telling.

Boise hated the computers....and I do not, but I do question a couple of algorithmns.

In my Humble Opinion....this is exactly right. I feel the same way about them. I like them, but there is no reason that Ole Miss should be ahead of FSU...change your formula or go away.

[Spence]

I do think I get it. I have a lot of references that say I don't. Boise did not like computers, because they did not favor them. I get that. Humans in their mind do exactly what computers do. The argument, rightfully made is computers will be consistent with the programing they have and humans, thankfully, do not bear that burden. So the question is asked, why doesn't the BCS or SEC or ESPN or the Libertarians, set the criteria that is to be used, how it should be be weighted. The discussion is absolutely valid. But no, they rely on individuals to give them the end result without having a clue of what they use and end up where they are. Remember Mr. Congrove was blackballed from the BCS because he was not willing to change his model to conform to whatever the BCS wanted. That's not right and will never be right.

I suspect when the NC game is formulated, it will be the one that gets the most exposure and brings in the most money. Notre Dame, and they may be the best team, I don't know, if they go undefeated..there is the money machine. If Oregon and Kansas State stub their toe, not necessarily even lose, Bama is back in. That's the money game.

They keep trying to set the criteria. I think that is the problem. Use them or don't use them, but please don't change them every time there is an answer that we don't like. The computers really have no chance to "get it right" because there isn't enough infomation available for them to be consistent. There aren't enough common opponents to get a fair sampling. Humans probably can come closer, but only if they can keep from showing a negative bias. When the humans vote honestly, I believe they get a pretty good poll. The problem comes when the games are played and the humans try to manipulate the top part of their poll. That is what happened to Georgia a few years ago and they played Hawaii instead of the NCG. That kind of manipulation is wrong.

[Derek]

I do think I get it. I have a lot of references that say I don't. Boise did not like computers, because they did not favor them. I get that. Humans in their mind do exactly what computers do. The argument, rightfully made is computers will be consistent with the programing they have and humans, thankfully, do not bear that burden. So the question is asked, why doesn't the BCS or SEC or ESPN or the Libertarians, set the criteria that is to be used, how it should be be weighted. The discussion is absolutely valid. But no, they rely on individuals to give them the end result without having a clue of what they use and end up where they are. Remember Mr. Congrove was blackballed from the BCS because he was not willing to change his model to conform to whatever the BCS wanted. That's not right and will never be right.

I suspect when the NC game is formulated, it will be the one that gets the most exposure and brings in the most money. Notre Dame, and they may be the best team, I don't know, if they go undefeated..there is the money machine. If Oregon and Kansas State stub their toe, not necessarily even lose, Bama is back in. That's the money game.

They keep trying to set the criteria. I think that is the problem. Use them or don't use them, but please don't change them every time there is an answer that we don't like. The computers really have no chance to "get it right" because there isn't enough infomation available for them to be consistent. There aren't enough common opponents to get a fair sampling. Humans probably can come closer, but only if they can keep from showing a negative bias. When the humans vote honestly, I believe they get a pretty good poll. The problem comes when the games are played and the humans try to manipulate the top part of their poll. That is what happened to Georgia a few years ago and they played Hawaii instead of the NCG. That kind of manipulation is wrong.

Exactly. I believe they CAN be more accurate than computers....but are often not. There is just to much "homer" bias, from me, and every other voter.

I don't know that there is a perfect solution. But it's what makes CFB so much better than Pro to me.

[Spence]

I don't think the homer bias causes problems. It is the negative bias. Homer bias generally works itself up. If a homer from 50 different markets all vote their teams (or the coaches from every conference vote their leader) it is usually a wash. It is where voters start holding other schools down is when it hurts the honesty of the poll.