Memphis losing made this look a lot better but Navy at 8 is your new outlier.

Michigan, Navy, and North Carolina are all way too high.  Baylor is too low.  USC is too high. 

Posted by ncmusician_7 on 11/3/2015 6:34:00 PM (view original):
I think the problem may be how SOS schedule is usually calculated.  I have some questions for the0nlyis. If a team wins a game against a 8-2 team and a 2-8 team, will they get the same "credit" (all other things equal) as beating two 5-5 teams?  Also, do you have a margin of victory cap for each game?

the0nlysis, I'm still waiting on a response to this

Posted by ncmusician_7 on 11/15/2015 5:03:00 PM (view original):

Posted by ncmusician_7 on 11/3/2015 6:34:00 PM (view original):
I think the problem may be how SOS schedule is usually calculated.  I have some questions for the0nlyis. If a team wins a game against a 8-2 team and a 2-8 team, will they get the same "credit" (all other things equal) as beating two 5-5 teams?  Also, do you have a margin of victory cap for each game?

the0nlysis, I'm still waiting on a response to this

It's the same way SOS was calculated in the BCS pre 2004.

Personally I do not agree that a schedule of 8-2 and 2-8 teams are better than a schedule of 2 5-5 teams.  Hence why I like the overall w-l record.

Margin of victory is not a true factor into the equation, since I don't include it and I don't add/subtract or input scoring off and scoring def into the same equation until the very end when everything is put together.

Navy is still high not sure if I think they are actually top 10 worhy.

Oklahoma St is still pretty low, that will change either for the good or bad with their remaining schedule.

I'd switch NU and Toledo(NU was 26th)

Ranks by conf:
AAC                 3
ACC                 3
Big 10             5
Big 12             4
Independant 1
Mac                1
Pac 12           3
SEC               5

Posted by the0nlyis on 11/15/2015 9:44:00 PM (view original):

Posted by ncmusician_7 on 11/15/2015 5:03:00 PM (view original):

Posted by ncmusician_7 on 11/3/2015 6:34:00 PM (view original):
I think the problem may be how SOS schedule is usually calculated.  I have some questions for the0nlyis. If a team wins a game against a 8-2 team and a 2-8 team, will they get the same "credit" (all other things equal) as beating two 5-5 teams?  Also, do you have a margin of victory cap for each game?

the0nlysis, I'm still waiting on a response to this

It's the same way SOS was calculated in the BCS pre 2004.

Personally I do not agree that a schedule of 8-2 and 2-8 teams are better than a schedule of 2 5-5 teams.  Hence why I like the overall w-l record.

Margin of victory is not a true factor into the equation, since I don't include it and I don't add/subtract or input scoring off and scoring def into the same equation until the very end when everything is put together.

Well then, my feedback would be that I don't particularly like it.  I don't see that this is an improvement on some of the other computer based rankings.  By calculating SOS, off, def this way, I believe you are giving too much value for blowout wins versus mediocre/bad teams and not enough value for wins against good teams.  Also, I don't think yardage should be used if you aren't going to also use turnovers (I actually wouldn't use either of them).  I think it would be better to place a value on each game result individually (then add them up).

Beating an 8-2 team (and a 2-8 team) is generally much more difficult than beating two 5-5 teams.  The fact that you don't account for that difference in difficulty is a clear problem with the ranking and may account for some of the odd rankings. 

The fact that you don't even agree with the rankings (on say Navy) shows there are clear problems with the ranking.  I mean if you don't even trust them how can anyone else?

I disagree with one ranking and maybe by ~8 spots too high.

And:

lets say chances of beating 8-2 team are 50% and than chances of beating the 2-8 team are 90% thats 70% of winning both

lets say chances of beating the 5-5 teams are 75% each thats 75% chance of winning.

is that 5% chance of winning that much harder? personally no.

Posted by the0nlyis on 11/18/2015 4:25:00 PM (view original):
I disagree with one ranking and maybe by ~8 spots too high.

And:

lets say chances of beating 8-2 team are 50% and than chances of beating the 2-8 team are 90% thats 70% of winning both

lets say chances of beating the 5-5 teams are 75% each thats 75% chance of winning.

is that 5% chance of winning that much harder? personally no.

In ur example, that would be 45% vs 56.25%, not 70% vs 75%.

I wholly disagree with your percentages.  It is generally significantly more difficult to beat an 8-2 team than a 5-5 team.  SIGNIFICANTLY. 

Look at it by rankings what is easier beating the 10th ranked and 100th ranked team or beating the 54th and 56th ranked team.  In both cases the average rank is 55, but which one is a more difficult 55 to achieve. 

Posted by ncmusician_7 on 11/18/2015 4:59:00 PM (view original):

Posted by the0nlyis on 11/18/2015 4:25:00 PM (view original):
I disagree with one ranking and maybe by ~8 spots too high.

And:

lets say chances of beating 8-2 team are 50% and than chances of beating the 2-8 team are 90% thats 70% of winning both

lets say chances of beating the 5-5 teams are 75% each thats 75% chance of winning.

is that 5% chance of winning that much harder? personally no.

In ur example, that would be 45% vs 56.25%, not 70% vs 75%.

ah yes,multiplication instead of averages, thats my bad.

Posted by moranis on 11/18/2015 5:08:00 PM (view original):
I wholly disagree with your percentages.  It is generally significantly more difficult to beat an 8-2 team than a 5-5 team.  SIGNIFICANTLY. 

Look at it by rankings what is easier beating the 10th ranked and 100th ranked team or beating the 54th and 56th ranked team.  In both cases the average rank is 55, but which one is a more difficult 55 to achieve. 

is a 8-2 appalachian state harder to beat than an 8-2 Michigan?

No, everyone would agree Michigan is a tougher team to beat.  Heck many cases could be made for a 4-6 Indiana or a 4-6 Minnesota being a tougher team to play then App State.

Hence why I think using overall win-loss of opponents is the best simplistic method to measure strength of schedule.

that is why I said generally and then changed it up to ranks.  8-2 App St. won't be ranked 10th, but 8-2 Michigan might be. 

You ignored that part.  So I'll ask again, what is more difficult beating the 10th and 100th ranked teams or beating the 54th and 56th ranked teams?

Posted by moranis on 11/18/2015 9:30:00 PM (view original):
that is why I said generally and then changed it up to ranks.  8-2 App St. won't be ranked 10th, but 8-2 Michigan might be. 

You ignored that part.  So I'll ask again, what is more difficult beating the 10th and 100th ranked teams or beating the 54th and 56th ranked teams?

I think it depends again.  Sure the 10th ranked team is clearly the toughest opponent, but the 100th is also the worst.

So from the USA today 1-128 rankings it would be:

Michigan State and Rice verse playing Texas Tech and NC State

I go with Tech and State being the harder games.  Sure MSU is easily the toughest game, but Rice is a cupcake.  Tech and NC State are both easily capable of putting up a fight.

Going off my rankings it would be:

Michigan and Syracuse verse Southern Mississippi and Arkansas State.

There I go with Michigan and Syracuse being tougher.

Since it depends a lot thats why I like the simplistic overall opponent win% to determine sos and I think for the most part it comes out to be correct, there is really only 1 controversial team in Navy.  If I wanted a non-controversial poll I wouldn't have made one and just used the AP Poll.  This is supposed to provide insight besides just eye test and bias.