College Preseason Top 25

Dedicated to College Football and Draft discussion.
post

Re: College Preseason Top 25

Postby DreadNaught » Tue Sep 10, 2019 9:54 am

Ken Carson wrote:
NavyBuc wrote:
Can't blame Clemson for their schedule being bad. They scheduled Texas A&M nonconference, which is pretty powerful. But it's not Clemson's fault the rest of the ACC sucks this year and FSU has gone off the deep end.

Speaking of schedules, LSU already beat #12 Texas, and still has to play #2 Alabama, #8 Auburn, #9 Florida, #16 Texas A&M, and should they get through that gauntlet, likely #3 Georgia (unless they lose to Florida). Sweet Jesus, that's not fun.

It should be illegal to refer to rankings prior to week 7. Those numbers mean absolutely nothing right now. Did Michigan look like the #7 team in the country?


Agreed. This is a rational/logical conclusion, which is why the playoff committee doesn't even meet til later in the season.

But rankings create discussion amongst fans which is obviously great for the game. So they aren't going anywhere as meaningless as they are this early.
Image
User avatar
DreadNaught
 
Posts: 16399
Joined: Thu Jan 15, 2015 9:18 am
Has thanked: 700 times
Been thanked: 851 times

Re: College Preseason Top 25

Postby Zarniwoop » Tue Sep 10, 2019 10:13 am

DreadNaught wrote:
Ken Carson wrote:It should be illegal to refer to rankings prior to week 7. Those numbers mean absolutely nothing right now. Did Michigan look like the #7 team in the country?


Agreed. This is a rational/logical conclusion, which is why the playoff committee doesn't even meet til later in the season.

But rankings create discussion amongst fans which is obviously great for the game. So they aren't going anywhere as meaningless as they are this early.



I wrote a research paper a few years ago about bias in polling that got published in an academic Sports Journal.....while the main focus was on bias in the coaches' poll*....one other thing I looked at was what people were always bitching about -- they thought pre-season and early week rankings are a bad idea because voters get "locked" into the order and won't let one team jump / fall based on what actually happens during the season....basically the idea that preseasons polls have too big of an influence on where teams end up.

Well that idea was basically debunked...because we compared the AP and Coaches poll (which include preseason rankings and rankings in the early week) to the ranking systems that started later in the year -- the biggest named one was the Harris poll which didn't rank teams til late October. We found no statistically significant differences between the rankings....in other words, the rankings that started later basically were identical to the rankings that started in pre-season.




***we found three significant sources of bias in the coaches poll:

1.) Coaches rank their own school higher on average than other coaches
2.) Coaches rank teams in their own conference on average higher than coaches from other conferences
3.) Coaches from the P-5 rank other P-5 programs higher than non P-5 programs compared to non P-5 coaches
User avatar
Zarniwoop
 
Posts: 11243
Joined: Thu Jan 15, 2015 12:23 pm
Has thanked: 1024 times
Been thanked: 738 times

Re: College Preseason Top 25

Postby Ken Carson » Tue Sep 10, 2019 10:18 am

Zarniwoop wrote:
Ken Carson wrote:The triple down. Got it.




you two really bring out the best in each other :P

In 10 months, Kress will quote this statement in order to prove that we do in fact bring out the best in each other thanks to your lack of blue text. Set your calendars...
Ken Carson
 
Posts: 5222
Joined: Wed Jan 14, 2015 12:33 pm
Has thanked: 5 times
Been thanked: 311 times

Re: College Preseason Top 25

Postby Ken Carson » Tue Sep 10, 2019 10:22 am

Zarniwoop wrote:
DreadNaught wrote:
Agreed. This is a rational/logical conclusion, which is why the playoff committee doesn't even meet til later in the season.

But rankings create discussion amongst fans which is obviously great for the game. So they aren't going anywhere as meaningless as they are this early.



I wrote a research paper a few years ago about bias in polling that got published in an academic Sports Journal.....while the main focus was on bias in the coaches' poll*....one other thing I looked at was what people were always bitching about -- they thought pre-season and early week rankings are a bad idea because voters get "locked" into the order and won't let one team jump / fall based on what actually happens during the season....basically the idea that preseasons polls have too big of an influence on where teams end up.

Well that idea was basically debunked...because we compared the AP and Coaches poll (which include preseason rankings and rankings in the early week) to the ranking systems that started later in the year -- the biggest named one was the Harris poll which didn't rank teams til late October. We found no statistically significant differences between the rankings....in other words, the rankings that started later basically were identical to the rankings that started in pre-season.




***we found three significant sources of bias in the coaches poll:

1.) Coaches rank their own school higher on average than other coaches
2.) Coaches rank teams in their own conference on average higher than coaches from other conferences
3.) Coaches from the P-5 rank other P-5 programs higher than non P-5 programs compared to non P-5 coaches

But aren’t you starting with a tainted sample? There is no world were you have a selection committee or poll created by people with no awareness of national college football rankings. I don’t see how you could realistically have a “control” to prevent bias leaking in.

Do you have the abstract or anything detailing the methodology?
Ken Carson
 
Posts: 5222
Joined: Wed Jan 14, 2015 12:33 pm
Has thanked: 5 times
Been thanked: 311 times

Re: College Preseason Top 25

Postby Zarniwoop » Tue Sep 10, 2019 10:39 am

Ken Carson wrote:
Zarniwoop wrote:

I wrote a research paper a few years ago about bias in polling that got published in an academic Sports Journal.....while the main focus was on bias in the coaches' poll*....one other thing I looked at was what people were always bitching about -- they thought pre-season and early week rankings are a bad idea because voters get "locked" into the order and won't let one team jump / fall based on what actually happens during the season....basically the idea that preseasons polls have too big of an influence on where teams end up.

Well that idea was basically debunked...because we compared the AP and Coaches poll (which include preseason rankings and rankings in the early week) to the ranking systems that started later in the year -- the biggest named one was the Harris poll which didn't rank teams til late October. We found no statistically significant differences between the rankings....in other words, the rankings that started later basically were identical to the rankings that started in pre-season.




***we found three significant sources of bias in the coaches poll:

1.) Coaches rank their own school higher on average than other coaches
2.) Coaches rank teams in their own conference on average higher than coaches from other conferences
3.) Coaches from the P-5 rank other P-5 programs higher than non P-5 programs compared to non P-5 coaches

But aren’t you starting with a tainted sample? There is no world were you have a selection committee or poll created by people with no awareness of national college football rankings. I don’t see how you could realistically have a “control” to prevent bias leaking in.

Do you have the abstract or anything detailing the methodology?



That is an interesting thought about the possibility of the early polls leaking bias into the later polls. I certainly recognize it as a potential methodological implication. I don't really think there is any real way to control for it - for example, even if the Harris poll voters tried not to look at the polls themselves, they certainly would see the ranking # when they watched games on TV.

I'll shoot you a PM for the other stuff
User avatar
Zarniwoop
 
Posts: 11243
Joined: Thu Jan 15, 2015 12:23 pm
Has thanked: 1024 times
Been thanked: 738 times

Re: College Preseason Top 25

Postby DreadNaught » Tue Sep 10, 2019 10:42 am

#SystematicBias

This is why UCF isn't equal.. :twisted:
Image
User avatar
DreadNaught
 
Posts: 16399
Joined: Thu Jan 15, 2015 9:18 am
Has thanked: 700 times
Been thanked: 851 times

Re: College Preseason Top 25

Postby Zarniwoop » Tue Sep 10, 2019 10:48 am

DreadNaught wrote:#SystematicBias

This is why UCF isn't equal.. :twisted:



What's interesting about this is because there were only two groups -- P5 coaches and non-P5 coaches...we basically just saw there was a significant difference. Basically P5 coaches ranked P5 teams higher. Non P5 coaches ranked non P5 teams higher. Because there was no "objective" data, one can't conclude whether it was the P5 coaches only that exhibited bias, the non-P5 coaches only exhibited bias or both did.

:)
User avatar
Zarniwoop
 
Posts: 11243
Joined: Thu Jan 15, 2015 12:23 pm
Has thanked: 1024 times
Been thanked: 738 times

Re: College Preseason Top 25

Postby NavyBuc » Tue Sep 10, 2019 11:14 am

Big statement game for UCF this weekend. Hope Costello plays so UCF can get Stamford’s best.
I like the Knights regardless by double digits.
NavyBuc
 
Posts: 2491
Joined: Sun Jan 18, 2015 9:07 am
Has thanked: 0 time
Been thanked: 73 times

Re: College Preseason Top 25

Postby Cheb » Tue Sep 10, 2019 11:48 am

Fascinating stuff, Zarni. Thanks for sharing.
Image
Cheb
 
Posts: 5407
Joined: Wed Jan 14, 2015 12:00 pm
Location: West Coast is best coast
Has thanked: 139 times
Been thanked: 519 times

Re: College Preseason Top 25

Postby Ken Carson » Tue Sep 10, 2019 1:00 pm

NavyBuc wrote:Big statement game for UCF this weekend. Hope Costello plays so UCF can get Stamford’s best.
I like the Knights regardless by double digits.

Vegas has UCF by 7.

A lot depends on how The Cardinal responds to getting boatraced by USC.

Also, interesting note... UCF was -10.5 against FAU last week. They covered that 3 times over and then some. Guessing Vegas have FAU a bump after they played a strong second half to cover vs Ohio State the week prior.
Ken Carson
 
Posts: 5222
Joined: Wed Jan 14, 2015 12:33 pm
Has thanked: 5 times
Been thanked: 311 times

Re: College Preseason Top 25

Postby Kress » Tue Sep 10, 2019 4:11 pm

Ken Carson wrote:
Zarniwoop wrote:you two really bring out the best in each other :P

In 10 months, Kress will quote this statement in order to prove that we do in fact bring out the best in each other thanks to your lack of blue text. Set your calendars...



Can I quote it as an example of your incredible wit? Because that's what you were trying to do there, right? Be witty? I never can tell, given your refusal to use wit font.
User avatar
Kress
 
Posts: 5965
Joined: Wed Jan 14, 2015 12:26 pm
Location: Cunning Linguist
Has thanked: 159 times
Been thanked: 331 times

Re: College Preseason Top 25

Postby BucaRican » Wed Sep 11, 2019 2:19 pm

Ken Carson wrote:
NavyBuc wrote:Big statement game for UCF this weekend. Hope Costello plays so UCF can get Stamford’s best.
I like the Knights regardless by double digits.

Vegas has UCF by 7.

A lot depends on how The Cardinal responds to getting boatraced by USC.

Also, interesting note... UCF was -10.5 against FAU last week. They covered that 3 times over and then some. Guessing Vegas have FAU a bump after they played a strong second half to cover vs Ohio State the week prior.


EASY day! UCF is going to play 3 QBs
BucaRican
 
Posts: 1305
Joined: Thu Jan 07, 2016 10:53 am
Has thanked: 6 times
Been thanked: 17 times

Previous

post

Return to College Football/Draft Talk

Who is online

Users browsing this forum: .irishHITMAN4rent and 4 guests