Sagarin ranks 4-3 USC #10 in the country

[ichawk24]

#10 in main ranking, #6 in predictor.

Using any single computer ranking is problematic, and there are lots of threads around here based entirely on a single Sagarin ranking. Though its imperfect, using a large sample of computer rankings is better.

USC #21 in the Massey Compilation:

http://www.masseyratings.com/cf/compare.htm

 

[theIowaHawk]

Are you really surprised? They are loaded with talent, and recent history has proven that good coaches with the team behind them at USC win games.

 

[Slappy Pappy]

Computers are only as good as the programer. Everyone knows that Sagarin puts his thumb on the scale for his favorites.

 

[ichawk24]

Are you really surprised? They are loaded with talent, and recent history has proven that good coaches with the team behind them at USC win games.

This is a non sequitur for a computer ranking. Computer rankings use game results to rank, not 3 year old recruiting rankings.

 

[iowalaker]

The amount of "talent"each team has, should do nothing for any ranking once a season has actually started.

Talent is a subjective measure. This is what hurts Iowa every time we are good. We might be a better football team, but our perceived "talent" is always less then other schools.

Stanford for example was perceived to have more talent then Iowa at the beginning of the year and now. That is why they are ranked ahead of us right now, even though we have a better record, and a common opponent, that tips the scales mightily in our favor.

 

[LincolnHawkeye]

It might help that in their 4 wins, they've outscored their opponents 198-53, while beating Arizona St and Utah pretty soundly, their 3 losses have been by 10, 5, and 10. Two of those losses came against top 10 ranked teams

 

[theIowaHawk]

The amount of "talent"each team has, should do nothing for any ranking once a season has actually started.

Talent is a subjective measure. This is what hurts Iowa every time we are good. We might be a better football team, but our perceived "talent" is always less then other schools.

Stanford for example was perceived to have more talent then Iowa at the beginning of the year and now. That is why they are ranked ahead of us right now, even though we have a better record, and a common opponent, that tips the scales mightily in our favor.

I think that is crazy. Of course talent matters. If the Knicks start 10-0 are they suddenly ranked 1 over a 3 loss Cavs team? Of course not. You seem to think it would be "unfair" to rank a 2-loss Bama over an undefeated Iowa.

What do you think rankings are for? Aren't they to determine the best team in the country? Or even the one who will be champion? That is skewed now because of a 4-team playoff, but do you think Iowa is a better team than Alabama? Better than Stanford? Because of Northwestern? How many games out of 10 would Iowa beat Stanford? Or, more logically, how many games out of 10 would Northwestern have beaten Stanford?

Everything is a subjective measure, right up until the playoff/NC.

You can see what I'm talking about with the ridiculous rankings for the AAC. They have 3 ranked teams...solely because they are undefeated. So if that is what goes in to a computer, it has to rank them highly, but writers and coaches do not. These are teams that the polls are "worse" than Memphis:

Previously #9 Florida State who lost on the actual wire;
Ole Miss
Duke
Pittsburgh
UCLA
Miss St
A&M
UNC
USC
BYU
Georgia
Wisconsin
Northwestern
Cal

Either the AAC is better than half the P5, or this is skewed because of their wins....obviously. It is like the opposite of the SEC, the bias is ever in favor of "undefeated", regardless of who the competition is. The ol' Boise State method.

So ask yourself: Which of those 14 schools (and all others below them) would lose to Memphis, Houston, and Temple 6 times out of 10?

Temple's SOS is 83, they beat Penn State. Memphis is 46 (just above ISU, two below Iowa) and beat Ole Miss. Houston is 87 and beat, uh, Louisville by 3.

 

[Chuck C]

I think that is crazy. Of course talent matters. If the Knicks start 10-0 are they suddenly ranked 1 over a 3 loss Cavs team? Of course not. You seem to think it would be "unfair" to rank a 2-loss Bama over an undefeated Iowa.

What do you think rankings are for? Aren't they to determine the best team in the country? Or even the one who will be champion? That is skewed now because of a 4-team playoff, but do you think Iowa is a better team than Alabama? Better than Stanford? Because of Northwestern? How many games out of 10 would Iowa beat Stanford? Or, more logically, how many games out of 10 would Northwestern have beaten Stanford?

Everything is a subjective measure, right up until the playoff/NC.

You can see what I'm talking about with the ridiculous rankings for the AAC. They have 3 ranked teams...solely because they are undefeated. So if that is what goes in to a computer, it has to rank them highly, but writers and coaches do not. These are teams that the polls are "worse" than Memphis:

Previously #9 Florida State who lost on the actual wire;
Ole Miss
Duke
Pittsburgh
UCLA
Miss St
A&M
UNC
USC
BYU
Georgia
Wisconsin
Northwestern
Cal

Either the AAC is better than half the P5, or this is skewed because of their wins....obviously. It is like the opposite of the SEC, the bias is ever in favor of "undefeated", regardless of who the competition is. The ol' Boise State method.

So ask yourself: Which of those 14 schools (and all others below them) would lose to Memphis, Houston, and Temple 6 times out of 10?

Temple's SOS is 83, they beat Penn State. Memphis is 46 (just above ISU, two below Iowa) and beat Ole Miss. Houston is 87 and beat, uh, Louisville by 3.

So why even play any games then? Just give USC and Alabama conference championships at the beginning of the year and call it good. The thing is results DO matter, and the fact is both USC and Alabama have lost, that Should affect them.

 

[soybean]

Are you really surprised? They are loaded with talent, and recent history has proven that good coaches with the team behind them at USC win games.

But they are 4-3...is a three loss team more talented than an undefeated one? Ranking based on perceived talent is exactly what we are trying to distance ourselves from with this new playoff scheme.

 

[theIowaHawk]

But they are 4-3...is a three loss team more talented than an undefeated one? Ranking based on perceived talent is exactly what we are trying to distance ourselves from with this new playoff scheme.

Well first, the word "talent", doesn't cross those lines per se. Of course a more talented team can lose to a more (even much more) talented team...that is a big reason we watch football.

Was Appalachian State really more talented than #5 Michigan in 2007? Of course not. Hell, they went on to lose to Wofford and GA Southern, both of which probably had less talent than Appalachian State.

I don't think Playoff distances us from talent, in fact I think it narrows it. It is MUCH more likely the more talented team wins the NC than a surprise Boise State knocking off Oklahoma with a statue of liberty proposal.

But, it really comes down to what you think a ranking is "for". If the season is over and USC is 9-3 and Memphis is undefeated, should Memphis be ranked higher? Probably, there aren't any games left to alter the standings. But at this point in the season, where do you think each should "finish" relative to the other?

Obviously is USC wins the next 5 and Memphis loses 3 they end with the same record....and NOBODY ranks Memphis higher, nor thinks they would beat USC in a bowl game. I think the ranking is "for" determining who is the best team. Which is why it is F****** RIDICULOUS that a #12 Michigan team losing to a #7 MSU is gaping absurdity. Why would the #12 team be expected to beat #7 in a ranking of who is best? They shouldn't, definitionally. When they damned near did/should have, they lose ranking to Notre Dame/Stanford/OSU, all teams who should have won?

When #1 beats #2 in close fashion, say, in 2006, #1 vs. #2 ....a game decided by 3 points, why should Michigan drop?

 

[ichawk24]

Idiot jackass derails the thread with stupidity.

 

[soybean]

Well first, the word "talent", doesn't cross those lines per se. Of course a more talented team can lose to a more (even much more) talented team...that is a big reason we watch football.

Was Appalachian State really more talented than #5 Michigan in 2007? Of course not. Hell, they went on to lose to Wofford and GA Southern, both of which probably had less talent than Appalachian State.

I don't think Playoff distances us from talent, in fact I think it narrows it. It is MUCH more likely the more talented team wins the NC than a surprise Boise State knocking off Oklahoma with a statue of liberty proposal.

But, it really comes down to what you think a ranking is "for". If the season is over and USC is 9-3 and Memphis is undefeated, should Memphis be ranked higher? Probably, there aren't any games left to alter the standings. But at this point in the season, where do you think each should "finish" relative to the other?

Obviously is USC wins the next 5 and Memphis loses 3 they end with the same record....and NOBODY ranks Memphis higher, nor thinks they would beat USC in a bowl game. I think the ranking is "for" determining who is the best team. Which is why it is F****** RIDICULOUS that a #12 Michigan team losing to a #7 MSU is gaping absurdity. Why would the #12 team be expected to beat #7 in a ranking of who is best? They shouldn't, definitionally. When they damned near did/should have, they lose ranking to Notre Dame/Stanford/OSU, all teams who should have won?

When #1 beats #2 in close fashion, say, in 2006, #1 vs. #2 ....a game decided by 3 points, why should Michigan drop?

I'm sorry, IowaHawk, I'm sure its my fault but I can't make heads or tails of your reasoning.

 

[Cushawk]

I think that is crazy. Of course talent matters. If the Knicks start 10-0 are they suddenly ranked 1 over a 3 loss Cavs team? Of course not. You seem to think it would be "unfair" to rank a 2-loss Bama over an undefeated Iowa.

What do you think rankings are for? Aren't they to determine the best team in the country? Or even the one who will be champion? That is skewed now because of a 4-team playoff, but do you think Iowa is a better team than Alabama? Better than Stanford? Because of Northwestern? How many games out of 10 would Iowa beat Stanford? Or, more logically, how many games out of 10 would Northwestern have beaten Stanford?

Everything is a subjective measure, right up until the playoff/NC.

You can see what I'm talking about with the ridiculous rankings for the AAC. They have 3 ranked teams...solely because they are undefeated. So if that is what goes in to a computer, it has to rank them highly, but writers and coaches do not. These are teams that the polls are "worse" than Memphis:

Previously #9 Florida State who lost on the actual wire;
Ole Miss
Duke
Pittsburgh
UCLA
Miss St
A&M
UNC
USC
BYU
Georgia
Wisconsin
Northwestern
Cal

Either the AAC is better than half the P5, or this is skewed because of their wins....obviously. It is like the opposite of the SEC, the bias is ever in favor of "undefeated", regardless of who the competition is. The ol' Boise State method.

So ask yourself: Which of those 14 schools (and all others below them) would lose to Memphis, Houston, and Temple 6 times out of 10?

Temple's SOS is 83, they beat Penn State. Memphis is 46 (just above ISU, two below Iowa) and beat Ole Miss. Houston is 87 and beat, uh, Louisville by 3.

Sorry, but most of this does not make any sense. How many times would Northwestern beat Stanford if they played 10 times? Who cares? They played once and Northwestern beat them badly. Also, how many of those teams on your list would Memphis beat? I have no idea. Nobody does. They played one team on that list (Ole Miss) and beat them soundly. So why should they be penalized because people perceive they wouldn't beat other teams on that list?
And one last thing. Having talent does not equate to having a good team. There are a ton of factors that go into building a team. Coaching, cohesion and other intangibles that cannot be measured. So while you are praising USC for having more talent, they are getting beat on the field by inferior talent. Didn't Norm always say "Give me 11 football players over 11 athletes"? And have you heard the mantra "Hard work beats talent when talent doesn't work hard"?

 

[EvilMonkeyInTheCloset]

Are you really surprised? They are loaded with talent, and recent history has proven that good coaches with the team behind them at USC win games.

Excellent use of sarcasm.

 

[EvilMonkeyInTheCloset]

P.S. f*** USC.

And do you know why?...........



Because f*** USC. Deal with it.

 

[spiderland]

Apparently, we are supposed to use the past ("recent history shows...") and the perceived future ("if USC wins their last 5 and goes 9-3..") to determine a team's current ranking.

Anything other than their present state, I guess.

 

[UNIowaHawk]

TRP has USC at 6 and Iowa at 33.

 

[Slappy Pappy]

Football "experts" are loathe to admit their preseason biases were wrong, and thus will discount teams they hadn't picked in favor of their original favorites at every opportunity.

 

[ichawk24]

TRP has USC at 6 and Iowa at 33.

Which is why I said in the OP:

Using any single computer ranking is problematic... Though its imperfect, using a large sample of computer rankings is better.

 

[DodgerHawki]

I think that is crazy. Of course talent matters. If the Knicks start 10-0 are they suddenly ranked 1 over a 3 loss Cavs team? Of course not. You seem to think it would be "unfair" to rank a 2-loss Bama over an undefeated Iowa.

What do you think rankings are for? Aren't they to determine the best team in the country? Or even the one who will be champion? That is skewed now because of a 4-team playoff, but do you think Iowa is a better team than Alabama? Better than Stanford? Because of Northwestern? How many games out of 10 would Iowa beat Stanford? Or, more logically, how many games out of 10 would Northwestern have beaten Stanford?

Everything is a subjective measure, right up until the playoff/NC.

You can see what I'm talking about with the ridiculous rankings for the AAC. They have 3 ranked teams...solely because they are undefeated. So if that is what goes in to a computer, it has to rank them highly, but writers and coaches do not. These are teams that the polls are "worse" than Memphis:

Previously #9 Florida State who lost on the actual wire;
Ole Miss
Duke
Pittsburgh
UCLA
Miss St
A&M
UNC
USC
BYU
Georgia
Wisconsin
Northwestern
Cal

Either the AAC is better than half the P5, or this is skewed because of their wins....obviously. It is like the opposite of the SEC, the bias is ever in favor of "undefeated", regardless of who the competition is. The ol' Boise State method.

So ask yourself: Which of those 14 schools (and all others below them) would lose to Memphis, Houston, and Temple 6 times out of 10?
Temple's SOS is 83, they beat Penn State. Memphis is 46 (just above ISU, two below Iowa) and beat Ole Miss. Houston is 87 and beat, uh, Louisville by 3.

Do you not get how rankings like Sagarin works? USC is ranked so high because they have demolished teams in their wins (including Utah and Arizona State) and lost games by close margins to good teams. Beating a team by 30 points shows more quality than beating the same team by 2 points. It's why Iowa took such a big jump in Sagarin after winning by 30 at Northwestern.

There is no bias in the Sagarin ratings, other than measuring who teams played, the quality of that team, where it was played and the result of that game.

 

[theIowaHawk]

Apparently, we are supposed to use the past ("recent history shows...") and the perceived future ("if USC wins their last 5 and goes 9-3..") to determine a team's current ranking.

Anything other than their present state, I guess.

So do you believe a ranking is simply a snapshot in time? Such as, on October 27, 2015 Iowa was the 10th best in the country, regardless of any other dates?

 

[icantfindausernamethatisntused]

There seems to be a lot of confusion about the Sagarin ratings. He used to have a component that looked only to Wins/Losses (I believe this was called Pure_ELO) but this year he has done away with that.

For his rankings, wins and losses are less important than other factors such as yards per play. In terms of predicting future performance, this is a better system than looking at wins and losses. A classic example is when Iowa State beat Nebraska a number of years ago when Nebraska had 9 turnovers. In that game, Nebraska dominated the line of scrimmage and gained yards at-will against ISU, but there were a number of fluke fumbles that kept Nebraska from scoring and set up Iowa State points. ISU squeaked out a win in that game, but I doubt anyone that watched that game would have thought that Iowa State was the better team.

With this in mind, when we look at USC's schedule, they have only been out-gained in one game and that was against Stanford. Notably, they outgained Notre Dame by over 100 yards although they ended up losing that game. Statistics show, generally, that the best predictive outcome of future games is whoever can gain more yards than their opposition. If we want a metric that will only predict future performance, the Sagarin ratings provides a very useful tool.

However, if we want to reward teams for wins and losses (which at the end of the day is what sport is all about) then this is not a useful tool. The Sagarin ratings system is not designed for this.

 

[theIowaHawk]

Sorry, but most of this does not make any sense. How many times would Northwestern beat Stanford if they played 10 times? Who cares? They played once and Northwestern beat them badly. Also, how many of those teams on your list would Memphis beat? I have no idea. Nobody does. They played one team on that list (Ole Miss) and beat them soundly. So why should they be penalized because people perceive they wouldn't beat other teams on that list?
And one last thing. Having talent does not equate to having a good team. There are a ton of factors that go into building a team. Coaching, cohesion and other intangibles that cannot be measured. So while you are praising USC for having more talent, they are getting beat on the field by inferior talent. Didn't Norm always say "Give me 11 football players over 11 athletes"? And have you heard the mantra "Hard work beats talent when talent doesn't work hard"?

Who is being penalized? You have to make subjective decisions because there simply is no objective, the point of the OP.

So if Temple and Houston are undefeated at the end of the season and go to play in the AAC championship they should/will be both top 5? If every other undefeated team stumbles they should be #1/#2? That is the absurdity. You don't believe they are #1/#2, and you are basing it on things other than their record, of course.

I never said talent equates to a good team, in fact I acknowledged the opposite and said that USC's own program has shown that with the team behind a good coach (see: Orgeron) they win.

This isn't difficult, it just depends on what you think the purpose of AP/Coaches rankings is. Try this simple question: If USC is 9-3, where do you rank them? That means they beat Cal, Oregon, UCLA and Arizona. If you rank them then, why wouldn't you rank them ahead of those teams? The answer is usually simple: because you don't think they can beat them. It is entirely dependent on what you think they can do, right?

So what do you think Memphis can do? Oh you, "have no idea"? Bullshit. Which of the following do YOU believe they would beat:
Ohio State
Baylor
Clemson
LSU
TCU
Michigan State
Alabama
Stanford
Notre Dame
Iowa
Florida
Oklahoma State
Utah
Oklahoma
Michigan
Florida State
Duke
Pittsburgh
UCLA
Mississippi State
A&M
Cal
UNC
USC
BYU
Georgia
Wisconsin
Northwestern

Do you think they would beat any of those? Than why would they be in your top 25? Because they haven't lost? Hell, neither has Coastal Carolina nor Mount Union.

 

[theIowaHawk]

Idiot jackass derails the thread with stupidity.

Wait, I derailed a thread complaining about USC and the questionable use of sole computer rankings?

 

[MadDogC]

They rank Baylor awfully high for their 103rd toughest schedule in the country, which tells me sagarins computer jacks off when teams go hurry-up against inferior opponents and run up scores.

 

[keosahawkeye]

So why even play any games then? Just give USC and Alabama conference championships at the beginning of the year and call it good. The thing is results DO matter, and the fact is both USC and Alabama have lost, that Should affect them.

Has this question been answered yet?

I understand flukes happen. But the the games USC has lost weren't due to kick-6s and black swan event type of plays.

The Trojans simply aren't objectively that good as a team. Congrats to Sagarin for coming up with an algorithm that tranforms dogsh!t into gold.

 

[spiderland]

So do you believe a ranking is simply a snapshot in time? Such as, on October 27, 2015 Iowa was the 10th best in the country, regardless of any other dates?

In my opinion, the ranking should be based on the team's entire body of work over the course of this season, to this point in time, relative to other teams' similar bodies of work.

I would not drop Iowa out of the Top 25 because they've finished poorly the last few years (recent past), and I believe a similar swoon is imminent, or because their recruiting classes (talent) have been ranked so low.

I also would not rank them in the top 2-3 in anticipation that they will go 5-0 to finish out an undefeated regular season.

I would rank them #10 based on a body of work that includes 7 wins against 0 losses, with at least 3 of those wins coming against quality opponents, and 2 of them on the road.

I do not believe a body of work that includes 3 losses and only 4 wins qualifies as Top 10. If and when that changes to 9-3, that will be a different story.

 

[Herkmeister]

Computers are only as good as the programer. Everyone knows that Sagarin puts his thumb on the scale for his favorites.

Maybe his grandma is from Waterloo.

 

[theIowaHawk]

In my opinion, the ranking should be based on the team's entire body of work over the course of this season, to this point in time, relative to other teams' similar bodies of work.

I would not drop Iowa out of the Top 25 because they've finished poorly the last few years (recent past), and I believe a similar swoon is imminent, or because their recruiting classes (talent) have been ranked so low.

I also would not rank them in the top 2-3 in anticipation that they will go 5-0 to finish out an undefeated regular season.

I would rank them #10 based on a body of work that includes 7 wins against 0 losses, with at least 3 of those wins coming against quality opponents, and 2 of them on the road.

I do not believe a body of work that includes 3 losses and only 4 wins qualifies as Top 10. If and when that changes to 9-3, that will be a different story.

See, now this is reasonable.

Now what would you do with the AAC teams? Especially knowing that only one of the three can even possibly end undefeated. Is a 1-loss AAC team better than a 2 loss P5? 3 loss?

 

[theIowaHawk]

Has this question been answered yet?

I understand flukes happen. But the the games USC has lost weren't due to kick-6s and black swan event type of plays.

The Trojans simply aren't objectively that good as a team. Congrats to Sagarin for coming up with an algorithm that tranforms dogsh!t into gold.

Of course, which is why rankings are simplistic and pointless to begin with. The ranking should simply be at the end of the season, if ever. The rankings do nothing, they don't base matchups off of them, they are for ridiculous threads like this!

 

[EvilMonkeyInTheCloset]

So do you believe a ranking is simply a snapshot in time? Such as, on October 27, 2015 Iowa was the 10th best in the country, regardless of any other dates?

On October 26th, 1985, Iowa was the #1 team in college football. Now that's a snapshot.

 

[theIowaHawk]

On October 26th, 1985, Iowa was the #1 team in college football. Now that's a snapshot.

Yes it was, and it was apparently wrong on a season-best approach as opposed to a snapshot approach. They obviously weren't better than #8 Ohio State, nor #13 UCLA. At least Ronnie Harmon wasn't.

 

[cg8284]

A thread about a COMPUTER ranking and some of you are arguing over intangible items. I think you're a little lost.

 

[EvilMonkeyInTheCloset]

Yes it was, and it was apparently wrong on a season-best approach as opposed to a snapshot approach. They obviously weren't better than #8 Ohio State, nor #13 UCLA. At least Ronnie Harmon wasn't.

Except Iowa finished ahead of Ohio State in the rankings.....

 

[theIowaHawk]

Weird, right?

 

[Blackhawk6]

Who is being penalized? You have to make subjective decisions because there simply is no objective, the point of the OP.

So if Temple and Houston are undefeated at the end of the season and go to play in the AAC championship they should/will be both top 5? If every other undefeated team stumbles they should be #1/#2? That is the absurdity. You don't believe they are #1/#2, and you are basing it on things other than their record, of course.

I never said talent equates to a good team, in fact I acknowledged the opposite and said that USC's own program has shown that with the team behind a good coach (see: Orgeron) they win.

This isn't difficult, it just depends on what you think the purpose of AP/Coaches rankings is. Try this simple question: If USC is 9-3, where do you rank them? That means they beat Cal, Oregon, UCLA and Arizona. If you rank them then, why wouldn't you rank them ahead of those teams? The answer is usually simple: because you don't think they can beat them. It is entirely dependent on what you think they can do, right?

So what do you think Memphis can do? Oh you, "have no idea"? Bullshit. Which of the following do YOU believe they would beat:
Ohio State
Baylor
Clemson
LSU
TCU
Michigan State
Alabama
Stanford
Notre Dame
Iowa
Florida
Oklahoma State
Utah
Oklahoma
Michigan
Florida State
Duke
Pittsburgh
UCLA
Mississippi State
A&M
Cal
UNC
USC
BYU
Georgia
Wisconsin
Northwestern

Do you think they would beat any of those? Than why would they be in your top 25? Because they haven't lost? Hell, neither has Coastal Carolina nor Mount Union.

Clever listing of "traditional powers", the bottom seven of whom are currently ranked below Temple, Houston and Memphis. Any reason you didn't throw in Oregon? Or Texas? Also, to help make your point, you omitted Mississippi from your list of "powers" untouchable by Memphis. Maybe untouchable...but very BEATABLE, like 37-24.

 

[icantfindausernamethatisntused]

Has this question been answered yet?

I understand flukes happen. But the the games USC has lost weren't due to kick-6s and black swan event type of plays.

The Trojans simply aren't objectively that good as a team. Congrats to Sagarin for coming up with an algorithm that tranforms dogsh!t into gold.

Perhaps a simplification of the Sagarin ratings will help: it is essentially a metric that is designed to provide "accurate" projections of future matchups. Essentially the logic flows like this: (1) to win, you need to score more points than the other team, (2) in order to score points you need to gain yards efficiently, and (3) in order to prevent points, you need to prevent other teams from gaining yards.

So far this season, USC has been very good at #2 and #3, which is why they rank so high in the Sagarin ratings. Unfortunately for them, being good at #2 and #3 has not translated into a good record. However, statistical studies will show that #2 and #3 are generally the best indicators of future performance.

USC does not "deserve" to be ranked in the top 10 or even in the top 25 at this point, because at the end of the day all that matters in sport is winning. They haven't done that well enough this year to be in the top 25. However, if I were asked to bet on a certain game, I would look to something like the Sagarin ratings rather than to a ranking system based on where teams "deserve" to be ranked.

 

[theIowaHawk]

Clever listing of "traditional powers", the bottom seven of whom are currently ranked below Temple, Houston and Memphis. Any reason you didn't throw in Oregon? Or Texas? Also, to help make your point, you omitted Mississippi from your list of "powers" untouchable by Memphis. Maybe untouchable...but very BEATABLE, like 37-24.

I omitted Mississippi, obviously because they lost to one of them. It wasn't a list of traditional powers, it was a list directly from the AP top 25 + vote getters. Those would be the teams these AAC teams are "competing" with for their rankings.

 

[theIowaHawk]

Perhaps a simplification of the Sagarin ratings will help: it is essentially a metric that is designed to provide "accurate" projections of future matchups. Essentially the logic flows like this: (1) to win, you need to score more points than the other team, (2) in order to score points you need to gain yards efficiently, and (3) in order to prevent points, you need to prevent other teams from gaining yards.

So far this season, USC has been very good at #2 and #3, which is why they rank so high in the Sagarin ratings. Unfortunately for them, being good at #2 and #3 has not translated into a good record. However, statistical studies will show that #2 and #3 are generally the best indicators of future performance.

USC does not "deserve" to be ranked in the top 10 or even in the top 25 at this point, because at the end of the day all that matters in sport is winning. They haven't done that well enough this year to be in the top 25. However, if I were asked to bet on a certain game, I would look to something like the Sagarin ratings rather than to a ranking system based on where teams "deserve" to be ranked.

This, this is part of what I've been advocating. It depends on what your rankings are "for". If there are for predicting who would win, than record is less important than other factors, but of course it is a factor. W/Ls are more of a secondary factor to all sorts of factors, such as yards gained for a simple example. If you gain more yards you are more likely to win.

But if it is about a "snapshot" of who has the most "impressive record" thus far, than none of that really matters and USC should be ranked lowly.

 

[bouma1415]

And if you base future results on yards per play, etc., you run the risk of being Ferentzed...that is letting off the gas when your team is ahead. A USC team may keep passing, and a KF team runs a standard running plays trying to run the clock out. Another reason computer rankings can be misleading.