Math Experts... Check This Computer Poll...

[WoVeU]

It looks good. I see a few places that look different from other polls I have looked at it, which can be good...or not so.

As you are developing a numerical system if you want to get a good comparison to the other guy's rankings out there I'd use a spread sheet and bar graph to compare, not just 1-25, but your points. Easier than using derivatives, limits, statistical moments, or even percent differences to shed light on what you are setting on. You are going to get a one-sided Gaussian curve (Normal Curve) for each system (Massey, Sagarin, etc.) you compare yours with.

*Note if you do this, you can also work on reverse engineering the other guys systems if you wanted to, if you include the factors they report they do and each would need a coefficient that you can manipulate and watch the curve move.

Also, if you are trying to avoid scores as input...home and away is still a good input.

[billybud]

The problem with the whole winning percentage and opponent's winning percentage is that by themselves (without a quality measure) they give you an inaccurate reading.

A 9-2 Arkansas State, on the surface, looks like 9-2 Oklahoma. Oklahoma's 6-6 opponent, Texas A&M, might be tougher than Arkansas State's 8-4 opponent Florida International, Oklahoma's 6-5 opponent, Iowa State, might be tougher than Arkansas State's 8-4 opponent, Louisianna Lafayette.

[billybud]

Auburn is an example of record vs the toughness of the games played.

Many would look at 7-5 Auburn and, on that record, underrate their team. Auburn (SOS #11)lost their five games to teams that are in the top 20.

7-5 Auburn might be a lot better than 7-5 Western Kentucky or 7-4 Wyoming.....SOS is important and is too often undervalued.

[RazorHawk]

I am not sure that you can develop a formula that will accurately rank the teams. I think there may be some things that can only be determined subjectively.
It is fun though to try.

[Spence]

You could if it was possible to get enough information to make it work.

[Dossenator]

How can you have a 10-2 South Carolina ranked ahead of a 10-2 Arkansas when Arkansas completely dominated SC this year. I was at that game...SC was not even competitive most of the game. Ark two losses came on the road to # 1 and #2. Pretty sure there are another 120 teams that would have 2 additional losses if they had to play Bama and LSU on the road.

I know no one but me saw Ark dominate SC. It was on the same time as Bama vs LSU. But you guys saw how Bama and LSU dominated Ark...well that's how lopsided SC looked against AR (on a side note that was AR 5th straight win over SC).

[WoVeU]

CFB,

Sorry for not noting about the tiers showing up in your system. But if you use W-L, 1st and 2nd Order...that is the sortation you are likely to converge on eventually.
*What your data is trying to tend toward (in the last form especially) is SOS. But it doesn't do so across all winning percentage but with in each, like 10-2 and 9-3. For your approach to have a lower record team to outrank a higher record team, you need notable upsets. Or a team playing a really weak schedule as compared to their W-L neighbors...that is Houston here.

If you can, falsify some data. As you are showing Top 25 do it with a game that involves only Top 25 teams. A very notable one should be WVU - LSU. Give LSU the loss and WVU the win and watch the movement in the results. Then make another adjustment, WVU (still showing a win against LSU) but have them losing to Marshall. Then compare that result to the current rankings.
This should show what a big upset would do for a team even if their record was the same.

If am scheduling and the BCS used this method I would look at the Non-BCS conferences, especially the larger ones, C-USA and the MAC. And try to play the best team from those. As I know they will play 9 or 10 conference games and have a good record. And still, their opponents in conference will play the same conference games.

[Dossenator]

SC only mustered 207 total yards....the final score was AR 44 - SC 28. AR did drop two TD passes that kept the game close. They did win by 16 and on their last two drives called the dogs off...run clock and punt. They could have put up more points. Of course it is more than a numbers game. I was at the game and the game never felt like it was in jeopardy of AR losing. SC simply could not move the ball. After the game Spurier just said AR was simply a better team...I think he is frustrated losing to AR five straight times. Same way AR is frustrated losing to Bama four straight.

[Dossenator]

Come on, Doss. I revamped one factor and applied more weight to it and now Arkansas is 5 spots ahead of the Gamecocks. You could have at least given me a pat on the back.

I"ll give you two pats, and throw in a Wooo Pig Sooooieee! I was not trying to say Ark needed to go up 5 spots...just that both teams have the same record, AR beat SC, and AR's two losses came to #1 and #2 (just did not see how SC could be ranked ahead of AR).

[Spence]

That is the kind of things I miss when putting the info in. Also, one site I use didn't have every team's win/loss record updated. I kept looking at it and wondered why it was that way with Arkansas and South Carolina. So, a small change was made, and that was the result. A couple of South Carolina wins looked better, but they did miss playing some of the SEC West's best. Arkansas wasn't as fortunate. So, that was taken into account.

Thanks again.

You can't account for every situation. There just isn't enough info to have any ranking be consistent without a change in how scheduling it done.