1. Florida State
2. Alabama
3. Missouri
4. Oregon
5. Ohio State
6. Stanford
7. Auburn
8. Virginia Tech
9. Clemson
10. Miami (FL)
11. Texas Tech
12. Baylor
13. Northern Illinois
14. UCLA
15. LSU
16. Oklahoma
17. Fresno State
18. Texas A&M
19. UCF
20. Oregon State
21. Michigan
26. South Carolina
28. Nebraska, Louisville and Oklahoma State

I like these better.

Can't really cherry puck...the computers are 1/3 of the rankings...the Harris and Coaches Poll are 2/3...but you can just go with 1/3 of the ranking if that makes you happy"...

(10-20-2013 09:41 PM)Maize Wrote:  Can't really cherry puck...the computers are 1/3 of the rankings...the Harris and Coaches Poll are 2/3...but you can just go with 1/3 of the ranking if that makes you happy"...

Everyone knows the polls are biased. Everyone complains about it. I really posted it to show how significantly different the rankings are when the computers are statistically and mathematically ranking teams without any human bias.

Hey wait, you put USCar behind you even though they beat you at your place!!!!!!! A little like putting UL above UCF, hmmmm.

(10-20-2013 10:08 PM)AusTxPony Wrote:  Hey wait, you put USCar behind you even though they beat you at your place!!!!!!! A little like putting UL above UCF, hmmmm.

Because SCe has lost two games. Do some homework. That said, these theoretical polls are troll bait.

(10-20-2013 09:41 PM)Maize Wrote:  Can't really cherry puck...the computers are 1/3 of the rankings...the Harris and Coaches Poll are 2/3...but you can just go with 1/3 of the ranking if that makes you happy"...

I am getting old...i just figured out what cherry puck meant.

Actually meant cherry pick...the point being you can't leave out 2/3 of the equation to suit your needs.

I'm still trying to figure out how NIU is so high other than the obvious fact that they do not have a loss. Their SOS is horrible and will get worse as they finish their even weaker MAC schedule. The best FBS school they faced is a 4 and 3 Iowa and is the only over .500 school they faced.

(10-21-2013 07:28 AM)sfink16 Wrote:  I'm still trying to figure out how NIU is so high other than the obvious fact that they do not have a loss. Their SOS is horrible and will get worse as they finish their even weaker MAC schedule. The best FBS school they faced is a 4 and 3 Iowa and is the only over .500 school they faced.

Don't some of the computers take into account the final ranking from last year as a guide for this season? There has to be a starting point and maybe NIU was high and has never really dropped.

Once they play more MAC games, their SOS will drop...they played Iowa and Purdue so far...Eastern Michigan, UMass, and Western Michigan are a combined 2-20.

(10-21-2013 05:50 AM)Maize Wrote:  Actually meant cherry pick...the point being you can't leave out 2/3 of the equation to suit your needs.

I believe the OP's intention was clear from the thread title. He wanted to show only the computer rankings the BCS uses to show how objective formulas rank the teams.

(10-21-2013 07:36 AM)Inigo Wrote:  

(10-21-2013 05:50 AM)Maize Wrote:  Actually meant cherry pick...the point being you can't leave out 2/3 of the equation to suit your needs.

I believe the OP's intention was clear from the thread title. He wanted to show only the computer rankings the BCS uses to show how objective formulas rank the teams.

Objective or baked in crap.

I asked a couple of post ago how NIU is rated so high by the computers with such a lousy SOS and OOC schedule (Iowa and Purdue, the best OOC teams are 5 and 9 combined).

After seeing such a SOS and OOC schedule, how in the world does the JS service ranked NIU 5th and Ohio State 8th? Both are undefeated but OSU a much better schedule.

(10-21-2013 07:35 AM)HP-TBDPITL Wrote:  

(10-21-2013 07:28 AM)sfink16 Wrote:  I'm still trying to figure out how NIU is so high other than the obvious fact that they do not have a loss. Their SOS is horrible and will get worse as they finish their even weaker MAC schedule. The best FBS school they faced is a 4 and 3 Iowa and is the only over .500 school they faced.

Don't some of the computers take into account the final ranking from last year as a guide for this season?

Yep. And that effect still lingers in the BCS computer rankings right now. It will be another 3 weeks or so before the computer rankings are based solely on this year's results. Until then you can't take them seriously.

OK, Quo, that makes sense for NIU (and Louisville), but why is Fresno so high. Even we beat them badly last year and they weren't ranked highly. However, that said, this is why scheduling light OOC makes sense, any 6-0, 7-0 gets you ranked, no matter the SOS.

(10-21-2013 11:58 AM)AusTxPony Wrote:  OK, Quo, that makes sense for NIU (and Louisville), but why is Fresno so high. Even we beat them badly last year and they weren't ranked highly. However, that said, this is why scheduling light OOC makes sense, any 6-0, 7-0 gets you ranked, no matter the SOS.

Fresno is high because the schools around them with 1 loss also have relatively soft schedules. E.g., UCF is around 85, Michigan is around 90 and Oregon State is around 100. So even though Fresno's schedule is really awful, around 120 right now, the 1-loss schools are down there too, enough for the loss to be decisive. Plus, like Fresno none of those teams were ranked at the end of last year so they aren't getting a carry-over boost in this year's computers.

(10-20-2013 09:45 PM)MagicKnightmare Wrote:  Everyone knows the polls are biased. Everyone complains about it. I really posted it to show how significantly different the rankings are when the computers are statistically and mathematically ranking teams without any human bias.

If there was no human bias, then the six different computer polls would not sport six different answers. There is bias in how each rating is determined, what metrics to use, which piece of data should count for how much, and which pieces of data to use. The only unbiased portion of the computer rankings is the actual data that is computated. But that is no different than the human polls.

All polls have bias unless it is a pure wins and losses poll, such as used in the pros. Anything outside of that, has bias.

(10-21-2013 07:35 AM)HP-TBDPITL Wrote:  Don't some of the computers take into account the final ranking from last year as a guide for this season? There has to be a starting point and maybe NIU was high and has never really dropped.

I believe some do up until usually the sixth or seventh game. By now, the data from last year should be gone. As for Northern Iowa, they have played five road games. I would bet the house no other undefeated team has played 5 road games. If the computers rank road games as heavily as the RPI does, that us enough to move their SOS up nearly 29%. Plus they have only played one less Big Ten team that Ohio State. Put that together, and it makes sense.

(10-21-2013 04:37 PM)adcorbett Wrote:  

(10-20-2013 09:45 PM)MagicKnightmare Wrote:  Everyone knows the polls are biased. Everyone complains about it. I really posted it to show how significantly different the rankings are when the computers are statistically and mathematically ranking teams without any human bias.

If there was no human bias, then the six different computer polls would not sport six different answers. There is bias in how each rating is determined, what metrics to use, which piece of data should count for how much, and which pieces of data to use. The only unbiased portion of the computer rankings is the actual data that is computated. But that is no different than the human polls.

All polls have bias unless it is a pure wins and losses poll, such as used in the pros. Anything outside of that, has bias.

Even pure wins and losses has bias, in that it implicitly assumes that every game is the same - e.g., a win over Alabama is no more or less meaningful than a win over Jackson State.

I think what those who call the computers "unbiased" mean is that (a) the algorithms are impartial such that unless there is a line of code that says "if team is from SEC, give them 3 bonus points" they do not systematically favor any team or conference, and (b) the algorithm is applied the same way to all teams over the course of the season.

(10-21-2013 05:22 PM)quo vadis Wrote:  Even pure wins and losses has bias, in that it implicitly assumes that every game is the same - e.g., a win over Alabama is no more or less meaningful than a win over Jackson State.

I think what those who call the computers "unbiased" mean is that (a) the algorithms are impartial and unless there is a line of code that says "if team is from SEC, give them 3 bonus points" do not systematically favor any team or conference, and (b) the algorithm is applied impartially to all teams over the course of the season.

True but bias can be built in without naming teams or conferences. I'll give a ridiculous example just to prove a point. What if the line of code reads if "if it is snowing during the game a point is deducted for the winning team because it's unfair to give full value for the win". It's the same code for all teams but unlikely to affect teams that play in warmer climate.

I realize that code like that would be ridiculous but silly code can be added to any program to make things work the way the coder wants it to work.

(10-21-2013 05:22 PM)quo vadis Wrote:  

(10-21-2013 04:37 PM)adcorbett Wrote:  All polls have bias unless it is a pure wins and losses poll, such as used in the pros. Anything outside of that, has bias.

Even pure wins and losses has bias, in that it implicitly assumes that every game is the same - e.g., a win over Alabama is no more or less meaningful than a win over Jackson State.

Actually a win over Alabama means exactly the same as a win over Jackson St in an unbiased world. You are describing the exact type of bias the OP aims to eliminate. No on ever said bias is inherently inaccurate: it's just subjective and thus open to question. Pure wins and losses take the bias you just introduced out of the equation. I didn't say wins and losses as the only measurement is the best way to measure teams, I just said it is the only measurement that lacks any sort of bias, since the object is to win the game. For example in the NFL there is no difference between a win over Jacksonville or Minnesota as opposed to Denver or San Francisco. They affect your seeding just the same - save for some tiebreaking scenarios, primarily those involving division opponents. However even then the quality of who you beat is irrelevant.