collegefbfan8898's Objective Poll

[donovan]

I think that is my point. If only the top 60 schools in SOS are elgible...then have a top 60 school division; but this nonsense of will play you and you can not reap the reward is nuts..and why the mid majors bought into that I will never know...well I do know but and so did Essau.

[..fanatic]

Computer models are generally reflective of what the programmer is trying to accomplish. Very few can keep out their bias...

COMPLETE ELIMINATION of bias has to be the ONLY objective. It is the only thing that can make a computer ranking better than a human ranking. If it is biased at all, it serves absolutely no purpose and is not a true computer ranking as it has been corrupted by human opinion. The formula must start with facts, constantly use nothing but facts, and allow only facts to be achieved in the answer.

Therefore, strength of schedule can not be predetermined based on conference (that enters an opinion). SOS should NOT be entered from a secondary source (even the NCAA) - your formula should determine this. Every aspect, except actual game results, should be your creation. You can't tell your formula to weight 80% here and 20% there - that is opinionating what is more inmportant and by how much. Winning percentage can not be the driving force in SOS - Michigan's loss to Appalachian State is worse than Boston College losing to Florida State. Enter nothing but facts and your unbiased formula will give you nothing but facts. Whether you or anyone else agrees with the results is unimportant.

[donovan]

I can not think of one computer program I have ever used....save balancing numbers.....that did not have programmers bias. None. This rages from Microsoft Word....tons of bias...to medical models...even more....

[billybud]

I have a question...

Bama is 6-3, as is FSU...

FSU beat Bama on a neutral field.

Start SEC rant, now...a one, a two, a three...

Bama's best win is over 6-3 Tennessee....FSU has beaten both Boston College and Bama...if it's "wins", Bama doesn't belong.

In fact, all SEC teams that are 6-3 are ranked....and 6-3 Tennessee is ranked higher than
6-3 Cal who beat the Vols...Kentucky didn't play and moved up the rankings (because 6-3 South Carolina became 6-4 (formerly #6).

And freeking Tennessee being ranked?...The Vols have had 45, 59, and 41 points scored against them in their losses...(FSU also lost three...but in the last two minutes to Miami who had a miracle drive in the last 1:59. FSU was on Clemson's 20 with a minute left only to lose by 6, and FSU lost by three to Wake.)

I think that Tennessee, as well as others, ride the SEC swell of the polls.

I saw a poster brag about having a bunch of SEC teams bowl eligible...well, Doh!

That's because of two reasons:

1. The SEC has played the fewest non-conference games vs BCS teams.

2. The SEC has played just 7 non-conference road games all season. 7 out of 40 non-conference games for just 17%, BY FAR the lowest of any conference.

It's kind of hard to NOT be bowl eligable when you are basically guaranteeing four wins...

Rant off...

[Spence]

COMPLETE ELIMINATION of bias has to be the ONLY objective. It is the only thing that can make a computer ranking better than a human ranking. If it is biased at all, it serves absolutely no purpose and is not a true computer ranking as it has been corrupted by human opinion. The formula must start with facts, constantly use nothing but facts, and allow only facts to be achieved in the answer.

Therefore, strength of schedule can not be predetermined based on conference (that enters an opinion). SOS should NOT be entered from a secondary source (even the NCAA) - your formula should determine this. Every aspect, except actual game results, should be your creation. You can't tell your formula to weight 80% here and 20% there - that is opinionating what is more inmportant and by how much. Winning percentage can not be the driving force in SOS - Michigan's loss to Appalachian State is worse than Boston College losing to Florida State. Enter nothing but facts and your unbiased formula will give you nothing but facts. Whether you or anyone else agrees with the results is unimportant.

I agree that complete emimination of bias should be the goal. I just don't think it is possible to completely take bias out even in a computer poll. That isn't to imply that the people who set up the formulas are trying to insert bias, but by setting a set of criteria to rank teams you are leaving out other criteria. Which means that even though all teams are judged on the same criteria, they are not being judged by some factors that may change the order of the ranking. Anyway, that is my basis for saying that all polls and rankings have pre built bias, even if it isn't intended.

That said you have to judge teams in some fashion and removing as much bias as possible is the best way to do it. The problem with that is one persons judgement of what a fair assessment would be is not someone elses. I believe that the computers and the human polls should have equal billing in the BCS and that the BCS committee should pick reputable rankings services and then stick with their results, not make them adjust their formulas to conform with common opinion. I may not like their results, but I conceed that I may not be right. The reason you use so many sources is that by putting so many rankings together the thought is that you eliminate regional bias and form a common view. I think for the most part it works. The two best teams may not always play for the championship, but two of the top 10 teams always do and I think that is as good as you would get in a 10 team playoff.

Billybud, I agree with you. I think when you beat the number 2 team and another ranked team you should be recognized in the T25. I put Florida State in my top 25 for that reason and neither KY or Tennessee made it. Boston College was recognized as a top 5 team in every poll, so if you beat that team and have other victories over a ranked team and a similar win loss total, you move up. Just my opinion, but fair is fair.

[..fanatic]

I can not think of one computer program I have ever used....save balancing numbers.....that did not have programmers bias. None. This rages from Microsoft Word....tons of bias...to medical models...even more....

Give me an example of what you mean by "bias" because I'm not seeing the impossibility of eliminating all bias. "Bias" to me would be an assumption that you should give more credit to a conference, for example. - In that case, you are prejudging a factor and you are blatantly biased.

But how is a spreadsheet bias before you tell it to do anything? It's blank.

One crude form of an unbiased formula.
1) Wait until every team has played 3 games.
2) Every team starts at "0".
3) Enter how much each team won or lost every game by.
4) Divide by the number of games played
5) Multiply by the number of wins
You have a crude ranking.

Team "A" won 3-out-3 by an average of 24 points = rating of 72
Team "B" won 2-out-of-3 with a total positive spread of 17 points (17 points divided 3 games x 2 wins = 11.33 points.

SOS can be determined by addding the rating of all of the teams played and dividing by the number of games (3, in this case).

I'm not putting this in a spreadsheet right now, so I don't know what the heck you'll get for answers. It's probably total crap. But all I'm trying to demonstrate is that formulas and answers can be totally unbiased and no computer program can make it not so.

Therefore, if you use you the right facts (without assuming weight to different criteria) and avoid "prejudging" what should be more important, you can create an unbiased ranking formula that makes sense. The trick is to find the facts and the formula that work in combination. You can judge its accuracy by seeing how it performs in picking pointspreads. Is it 50% right or 75% right or 25% right. Can you make it 80% right? How? That would seem to be the process. You can't judge it by putting it together and then asking yourtself if the rankings "look" right. You have to see if it performs with any reliability. And, all the while, maintaining an unbiased approach. If, through application, you find that weighting a factor improves the reliability in picking winners or spreads, than you are not necessarily inputting "bias" because this wasn't prejudged, it was proven.

[Dossenator]

Billybud...one error...Ark is 6-3 and they are not ranked.

No way to have a computer ranking that is not influenced by humans. Someone has to create the formula for the computers to use to determine who is #1-#25. And who determines what numbers are more important then others...what percentage to weight this over that....No way the system will ever be perfect. No way to ever get the human touch out of the equation. Only way a pole will ever completely be unbiased is if God himself delivers such a machine....and I have heard that he is a Razorback fan....so, there you go.

[billybud]

Yeah...I forgot Arkansas...as did the raters.

Guess it was having a couple of teams score 40 plus points on the Razorbacks.

[..fanatic]

No way to have a computer ranking that is not influenced by humans. Someone has to create the formula for the computers to use to determine who is #1-#25. And who determines what numbers are more important then others...what percentage to weight this over that.

Why does everyone assume there has to be bias built in? Bias means, for example, you give the SEC more weight than the Sun Belt so you build the formula to do so. But an unbiased formula would provide that result. Just because humans create the ranking formulas does not inherently mean they are biased. This is like a conspiracy theory. If you don't have built-in favoritism for one team, or group of teams, over another - you don't have bias. Plain and simple. That's where voters are biased. They THINK one conference or group of teams is stronger than another so they rely on their bias to rank the teams. Computers do not automatically assume such things, and if you don't tell it to, it won't be biased.

[Dossenator]

I don't mean that at all fanatic.... Who determines how much strength of schedule is going to count, how much wins are going to count, etc. If you favor wins more in the formula then somes teams benefit from that....if you favor SOS....then maybe a totally different set of teams benefit from that. You can come up with 100 programs that look to be fair and everyone of them will give you different teams at the top. Who is to say which is the best way? This is the human part I am talking about. I for one think the human part should never be taken out. Like someone stated before....App St beating Michigan is a worse loss then FSU beating Boston College....computers don't truly see the difference.

[..fanatic]

I for one think the human part should never be taken out. Like someone stated before....App St beating Michigan is a worse loss then FSU beating Boston College....computers don't truly see the difference.

??? How do the computers not see the difference in that scenario? They are producing power rankings from their results. Of course, they see the difference. One school lost to a team with a power rating 35 points below theirs. The other school lost to a team whose power ratings are about even.

Also, you don't sit around and decide you're going to give SOS a certain percentage of impact in the rankings.

If a team with a power rating of 85 beats a team with a power rating of 60, and a team with a power rating of 90 beats a team with a power rating of 65, the second one obviously has a higher SOS because (based on that one game) it beat a team with a higher power rating. The SOS is a product of the results, it's not a separate formula aimed at changing the results. That's why the BCS removed it' SOS quotient from its rankings - the mathemeticians tiold them it was duplicitous.

[Dossenator]

fantatic...where do the power rankings come from? A whole other thing now. Are they created by humans or computers again....but what's the difference. A human has to come up with the eqaution that decides the power rankings. I see how power rankings then could show the difference in the two losses, but who is to say the power rankings are correct, fair, and unbiased.

I guess I am just not understanding anything you guys are saying. If SOS is not a percentage of how the computers pick the BCS standings then I am completely confused. fanatic are you saying all teams are ranked solely on SOS and nothing else. That is what you just said...that you can't decide to give SOS a certain percentage of impact on the rankings....so you are saying it means everything or that is means nothing. Not sure I follow. Is strength of schedule not factored in to the current computer rankings? If so what percentage are they given in the equation? Please explain.

My point is there is no way to really rank teams and be fair. In the end we don't know is Ark can beat Boston College, or OU can beat Oregon, or if Ohio St can beat Florida, and etc.....you get my point. The only way to know who the best team in the country is to have all 119 teams play every team. That is impossible. It is all a big guess since teams only will face 12 to 14 teams in a given season. I don't think the computers help to really solve the problem.

I think we should drop the computers completely and just go back to voters. Because yet again we will have another year with the BCS where 2 or 3 teams that could or maybe should be given a shot at the NC that will be left out....how are the computers doing anything that the voters can do on their own? You know look over teams that should get a shot....

[donovan]

I can not think of one computer program I have ever used....save balancing numbers.....that did not have programmers bias. None. This rages from Microsoft Word....tons of bias...to medical models...even more....

Give me an example of what you mean by "bias" because I'm not seeing the impossibility of eliminating all bias. "Bias" to me would be an assumption that you should give more credit to a conference, for example. - In that case, you are prejudging a factor and you are blatantly biased.

As I read your case I said to myself...ok...it is self answering.

Let me try to figure out examples..

1. You have a column of numbers... 1 to 9. you add them up. you get 45. Seems unbiased and may well be if you are counting apples in boxes and you count correctly. So in this case it can be.

2. You count apples in boxes and come up with the numbers 1 to 9...you get 45.....but..someone told you..only count the apples with stems..if no stem...do not count. Bias...does not mean it is wrong...bias is not an evil word..means..criteria for counting was prejudiced by condition acceptable.

3. In a conference..you take wins and losses....get a percentage and you have a winner...but the decision was made to count all wins and loses equal...now we accept this bias..but it is still there.

3. On the national rankings...we have all kind of input that is based upon opinion. SOS OOC rating of teams....the math is done right..but the basis is not with out prejudice.

My point..which is never clear is...people that say their polls have not bias or prejudice
or helmet stuff...whatever that is...fail to recognize they do and account for it...nothing wrong with that..just be not deceived.

There are fields..medical for one...if you do not know the philosophy upon which computer protocol is designed...it could be fatal....and sometimes is...critical...