One Loss Team Comparison

[Landlord]

As long as people are convinced "SEC is overrated because ESPN", there will be no reasoning that will sway them. I can point to unbiased systems till I'm blue in the face--systems that do not watch ESPN and have no understanding of conference affiliation or even team name--no one will bother.

These systems are made and influenced by people, who are influenced a great deal by ESPN.

It's not possible to make unbiased systems in a vaccuum that doesn't have ESPN influence.

 

[JTrain]

As long as people are convinced "SEC is overrated because ESPN", there will be no reasoning that will sway them. I can point to unbiased systems till I'm blue in the face--systems that do not watch ESPN and have no understanding of conference affiliation or even team name--no one will bother.

These systems are made and influenced by people, who are influenced a great deal by ESPN.

It's not possible to make unbiased systems in a vaccuum that doesn't have ESPN influence.

Yes, it is. Unbiased means the computer does not know anything about the teams or conferences. Just scores. Unless the scores were somehow significantly different between conferences on a regular basis (in other words, if a program could look at 100 SEC scores and 100 Big 12 scores with no teams and be able to tell based only on the numbers which conference was which), there is no way to intentionally skew the system towards the creator's favorite team or conference.

A lot of these creators are pretty brilliant guys and many of them have published papers on their methodology. They would tell you it would be extremely difficult or perhaps impossible to program an unbiased rating system to favor any conference, even if they wanted to. Some of them are very accessible via email or Twitter: Kenneth Massey, Wes Colley and perhaps Peter Wolfe (who also was friends with the late David Rothman and probably knows most of the ins and outs of his system). You could certainly debate the point with them if you like.

Not to mention all of the top systems were formulated years and years before ESPN created the SEC Network.

 

[Kernal]

As long as people are convinced "SEC is overrated because ESPN", there will be no reasoning that will sway them. I can point to unbiased systems till I'm blue in the face--systems that do not watch ESPN and have no understanding of conference affiliation or even team name--no one will bother.

These systems are made and influenced by people, who are influenced a great deal by ESPN.

It's not possible to make unbiased systems in a vaccuum that doesn't have ESPN influence.

Yes, it is. Unbiased means the computer does not know anything about the teams or conferences. Just scores. Unless the scores were somehow significantly different between conferences on a regular basis (in other words, if a program could look at 100 SEC scores and 100 Big 12 scores with no teams and be able to tell based only on the numbers which conference was which), there is no way to intentionally skew the system towards the creator's favorite team or conference.

A lot of these creators are pretty brilliant guys and many of them have published papers on their methodology. They would tell you it would be extremely difficult or perhaps impossible to program an unbiased rating system to favor any conference, even if they wanted to. Some of them are very accessible via email or Twitter: Kenneth Massey, Wes Colley and perhaps Peter Wolfe (who also was friends with the late David Rothman and probably knows most of the ins and outs of his system). You could certainly debate the point with them if you like.

Not to mention all of the top systems were formulated years and years before ESPN created the SEC Network.

There are no "unbiased systems" because systems don't program themselves. The selection and weighting of variables among the millions of possibilities introduces the beliefs of the programmer. You get out what you put in, nothing more.

 

[JTrain]

As long as people are convinced "SEC is overrated because ESPN", there will be no reasoning that will sway them. I can point to unbiased systems till I'm blue in the face--systems that do not watch ESPN and have no understanding of conference affiliation or even team name--no one will bother.

These systems are made and influenced by people, who are influenced a great deal by ESPN.

It's not possible to make unbiased systems in a vaccuum that doesn't have ESPN influence.

Yes, it is. Unbiased means the computer does not know anything about the teams or conferences. Just scores. Unless the scores were somehow significantly different between conferences on a regular basis (in other words, if a program could look at 100 SEC scores and 100 Big 12 scores with no teams and be able to tell based only on the numbers which conference was which), there is no way to intentionally skew the system towards the creator's favorite team or conference.

A lot of these creators are pretty brilliant guys and many of them have published papers on their methodology. They would tell you it would be extremely difficult or perhaps impossible to program an unbiased rating system to favor any conference, even if they wanted to. Some of them are very accessible via email or Twitter: Kenneth Massey, Wes Colley and perhaps Peter Wolfe (who also was friends with the late David Rothman and probably knows most of the ins and outs of his system). You could certainly debate the point with them if you like.

Not to mention all of the top systems were formulated years and years before ESPN created the SEC Network.

There are no "unbiased systems" because systems don't program themselves. The selection and weighting of variables among the millions of possibilities introduces the beliefs of the programmer. You get out what you put in, nothing more.

Unbiased system means no data outside game results. It does not mean value-free. Some systems value margin of victory more than others, for example. Some don't include it at all. Whatever the value system, it is applied evenly among all results without reference to outside information, including conference. A ten point win is a ten point win every time, and the team name is just a string of characters to the program--a row in a database.

http://www.masseyratings.com/theory/

http://www.masseyratings.com/faq.php

http://homepages.cae.wisc.edu/~dwilson/rsfc/rate/rothman.html

http://www.colleyrankings.com/matrate.pdf

 

[skersfan]

We don't need to worry about it. Take it to the bank.

Gawd I hope I have to eat crow on this.

 

[VA Husker Fan]

As long as people are convinced "SEC is overrated because ESPN", there will be no reasoning that will sway them. I can point to unbiased systems till I'm blue in the face--systems that do not watch ESPN and have no understanding of conference affiliation or even team name--no one will bother.

These systems are made and influenced by people, who are influenced a great deal by ESPN.

It's not possible to make unbiased systems in a vaccuum that doesn't have ESPN influence.

Good grief. And you got a +1 for this. I really hope you are joking. Either way, please explain how you figure this could possibly happen. I need another laugh today.

 

[SandhillshuskerW]

#11 Kansas State - Lost to #4; Beat #19

#12 Baylor - Lost to #20; Beat #10

#14 Arizona - Lost to #27; Beat #5

#10 TCU - Lost to #12; Beat #19, #33

#9 Georgia - Lost to unranked; Beat #22, #31

A better question is what has Georgia done to deserve to be ranked ahead of any of these teams?

There are some people that get tired of hearing this excuse, but it's true.........Georgia is in the SEC. That's the only reason why they are given that kind of respect. If anyone disagrees with that, just look at Missouri and Texas A&M and tell me that those teams would be given the same respect they have been given this year if they were still in the Big 12.

I agree with most people though that say, it will start to sort its way out in the next couple of weeks. I'm sure we will still be upset in some cases though if some SEC teams are still ranked above us with two losses.

 

[Amac3309]

#11 Kansas State - Lost to #4; Beat #19

#12 Baylor - Lost to #20; Beat #10

#14 Arizona - Lost to #27; Beat #5

#10 TCU - Lost to #12; Beat #19, #33

#9 Georgia - Lost to unranked; Beat #22, #31

A better question is what has Georgia done to deserve to be ranked ahead of any of these teams?

There are some people that get tired of hearing this excuse, but it's true.........Georgia is in the SEC. That's the only reason why they are given that kind of respect. If anyone disagrees with that, just look at Missouri and Texas A&M and tell me that those teams would be given the same respect they have been given this year if they were still in the Big 12.
I agree with most people though that say, it will start to sort its way out in the next couple of weeks. I'm sure we will still be upset in some cases though if some SEC teams are still ranked above us with two losses.

Not according to Joe Schad

@schadjoe: Only Top 25 teams in Top 25 in both scoring O and scoring D: Auburn, Mich. St., UGA, Kansas St., Ohio St., Nebraska, Marshall.

@schadjoe: Only Top 25 teams in Top 25 in both scoring O and scoring D with strength of schedule in Top 45: Auburn and Georgia.

@schadjoe: UGA underrated? #1 in nation w +13 turnover margin, #8 scoring o, #19 scoring d, #14 3rd down d, #14 rush o, #13 rush d, #7 comp. pct.

BTW The huskers ranks in those areas are

12 scoring o, 22 scoring d, 7 rush o, 29 rush d, 6 opp 3rd down, 59 turnover margin

I'm not huge on the SEC bias thing, while I think it exists I think it's more predicated on the preseason polls and the way the SEC ingeniously does their scheduling. That being said it drives me nuts how the National media doesn't see or acknowledge how they feed it just like Schad is

So one is in the top 10 and one drops a spot below a 2 loss SEC school after a blowout, who is overrated and who is underrated, and I would argue their strength of schedule is pretty freakin close!

 

[Kernal]

As long as people are convinced "SEC is overrated because ESPN", there will be no reasoning that will sway them. I can point to unbiased systems till I'm blue in the face--systems that do not watch ESPN and have no understanding of conference affiliation or even team name--no one will bother.

These systems are made and influenced by people, who are influenced a great deal by ESPN.

It's not possible to make unbiased systems in a vaccuum that doesn't have ESPN influence.

Yes, it is. Unbiased means the computer does not know anything about the teams or conferences. Just scores. Unless the scores were somehow significantly different between conferences on a regular basis (in other words, if a program could look at 100 SEC scores and 100 Big 12 scores with no teams and be able to tell based only on the numbers which conference was which), there is no way to intentionally skew the system towards the creator's favorite team or conference.

A lot of these creators are pretty brilliant guys and many of them have published papers on their methodology. They would tell you it would be extremely difficult or perhaps impossible to program an unbiased rating system to favor any conference, even if they wanted to. Some of them are very accessible via email or Twitter: Kenneth Massey, Wes Colley and perhaps Peter Wolfe (who also was friends with the late David Rothman and probably knows most of the ins and outs of his system). You could certainly debate the point with them if you like.

Not to mention all of the top systems were formulated years and years before ESPN created the SEC Network.

There are no "unbiased systems" because systems don't program themselves. The selection and weighting of variables among the millions of possibilities introduces the beliefs of the programmer. You get out what you put in, nothing more.

Unbiased system means no data outside game results. It does not mean value-free. Some systems value margin of victory more than others, for example. Some don't include it at all. Whatever the value system, it is applied evenly among all results without reference to outside information, including conference. A ten point win is a ten point win every time, and the team name is just a string of characters to the program--a row in a database.

http://www.masseyratings.com/theory/

http://www.masseyratings.com/faq.php

http://homepages.cae.wisc.edu/~dwilson/rsfc/rate/rothman.html

http://www.colleyrankings.com/matrate.pdf

Yeah, I've read about the different computer rankings previously.

Even if you use only data from games, which data do you use? How do you weight it? Is rushing defense more important than pass-efficiency defense, etc., etc., etc.,? Somebody has to decide how to do all that. That's why there are a thousand different polls.

There is no agreed-upon weighting and selecting of game data for accurately ranking teams. If there were, there would be no speculation. There would be only one poll. There would be no need for a committee to pick the playoff teams. The fact is, nobody knows with any certainty how to rank teams. The only legitimate method is wins and losses, since that is the only game result that matters, but then you must also value strength of schedule, which is another variable nobody really knows how to value. RPI comes close, but even then someone is deciding whether direct opponents are worth only 25%, opponents' opponents 50% and opp opp opps 25%, or whatever.

If it were possible to accurately rank teams there would be no need to play the games. The whole idea of ranking teams is contrary to the idea of sport, which is all about direct competition. Would the NFL rank teams? Not in a million years. Because it becomes a beauty pageant. Always. Especially when money is involved.

With regards to ESPN influencing things. Well, of course they do. Who watches CFB and isn't exposed to ESPN? Nobody, that's who. Even if you only watch games on there, you're exposed to their broadcasters musings and opinions.

 

[NebraskaShellback]

http://www.sbnation.com/lookit/2014/10/25/7071597/brent-musburger-on-medias-sec-bias-deal-with-it-theyre-the-best

Brent Musburger addressed claims that ESPN and SEC Network employees are biased... kind of."Deal with it, they're the best." He doesn't quite say he isn't biased, he just says that the SEC happens to have the best teams. Those two things are not mutually exclusive.



To be fair, the stats back him up. Earlier, Chris Fowler had called claims of ESPN's SEC bias "stupid, uninformed stuff."

 

[VA Husker Fan]

The computers don't watch ESPN and aren't influenced. The programmers probably do watch, but they set the parameters of their programs before the season, and most of them have actually stayed set for many, many years, and don't change based on what ESPN or anyone else says.

The main problem with the computer ratings is that 12 games isn't really a good enough sampling. I just flipped a coin 12 times, it came up tails 8 times. Should I assume that tails is twice as likely to come up as heads? If I flipped 100 times, I'm pretty sure I'm come a lot closer to 50/50. 1000 times, even closer. The other problem with the computers is that for many of them, we don't even know the parameters they use, and whether they make sense. ESPN influence? No.

Regarding Georgia, Joe Schad gives yet another example of how to make a case for a team based on a conveniently selected set of data. Mostly meaningless. I would put Georgia more toward the bottom of 1 loss major conference teams at this point.

 

[theneguy95]

Do the computers do a preseason poll?

 

[Mavric]

Just to see how my list lined up with the committee.

#3 Auburn - Lost to #1; Beat #11, #16
#10 Notre Dame - Lost to #2; Beat #29
#4 Mississippi - Lost to #16; Beat #3
#6 Alabama - Lost to #7; Beat #20
#5 Oregon - Lost to #14; Beat #8, #25
#8 Michigan State - Lost to #5; Beat #17
#9 Kansas State - Lost to #4; Beat #19
#13 Baylor - Lost to #20; Beat #10
#12 Arizona - Lost to #27; Beat #5
#7 TCU - Lost to #12; Beat #19, #33
#15 Nebraska - Lost to #8; Beat none
#14 Arizona State - Lost to #25; Beat #27, #29
#11 Georgia - Lost to unranked; Beat #22, #31
#17 Utah - Lost to unranked; Beat #25, #27
#16 Ohio State - Lost to unranked; Beat none
#24 Duke - Lost unranked; Beat none
#23 East Carolina - Lost to unranked; Beat none

So there are three teams that are way out of order between the two - Notre Dame is significantly lower to the Committee, TCU is significantly higher and Georgia is significantly higher.

Notre Dame seems to be downgraded for not having a really good win.

TCU's "good" loss seems to matter more than Baylor and Arizona's good wins. Georgia doesn't seem to get downgraded much for a "bad" loss.

 

[admo]

Can add FSU to the one loss group if they do not make a comeback win. If they do end up losing I think several 1 loss teams should bump them. Not an impressive schedule and they have struggled several times this year. I'm just not impressed.

Got 5 minutes to keep fingers crossed Louisville holds on.

 

[Landlord]

Can add FSU to the one loss group if they do not make a comeback win. If they do end up losing I think several 1 loss teams should bump them. Not an impressive schedule and they have struggled several times this year.

They've got a win over #10 and their loss would be to a team ranked higher than #25.

That's a better resume than Nebraska, Arizona State, Georgia, Utah, Ohio State, Duke, East Carolina, Notre Dame, and arguably Michigan State, Alabama, Oregon, Baylor and TCU as well.