way too early polls

[ohstatefan]

May 27, 2014 11:36:13 GMT -5 @buckina said:

May 27, 2014 10:29:45 GMT -5 WANG said:

What OSF said. People put waaaaaaay too much stock in preseason polls and are unwilling to move people down a large number, even with poor performances.

For example, pretend that FSU starts this year #1 and runs the table, but due to a weak conference and bad OOC schedule (again, this is just hypothetical, not saying the ACC is bad) they have just an awful SOS. Additionally, they are unimpressive on offense and give up more points than they should. Voters, generally speaking, are going to be unlikely to move them off of their #1 spot simply because they started the season ranked there and never lost.

If voters were more fluid with their rankings from week to week, I wouldn't have a problem with preseason polls, but unfortunately that is not the case.

This. Remember when we backed into the NCG.

Not sure how tough are schedule is this year. Navy gave us a fit a few years back. MSU seems like the toughest game but we barely beat scUM last year.

There are several examples to consider.

2002 pre-season we were ranked #13. We didn't climb up to the #2 spot until week 12 when the only other undefeated team left was Miami. In other words if Oklahoma had kept winning that year, we wouldn't have been in the NC game.

2004 pre-season polls had Auburn #17. But USC and Oklahoma kept winning, so Auburn didn't get a chance at the title.

I shouldn't complain about this too much, since it's the big schools rich in history that usually benefit from such a system. They get noticed more which gives them a better chance of being ranked. We only suffer from not being in the SEC. We usually just need to win all our games to play for the title. The smaller schools or those who haven't been there in a long time also need a fair amount of luck.

[WANG]

May 27, 2014 11:36:13 GMT -5 @buckina said:

This. Remember when we backed into the NCG.

Not sure how tough are schedule is this year. Navy gave us a fit a few years back. MSU seems like the toughest game but we barely beat scUM last year.

And that is the thing, NOBODY knows how tough OSU's schedule will actually be this year. Sure, people can make their best educated guess, but until the season plays out that's all it is.

On paper, it looks like the OOC isn't terrible (Navy and VT should be okay team, UC isn't a complete throw away game, and Kent St is bad), but the conference schedule looks pretty light (toughest games will be @ MSU and probably @ PSU).

But who knows? Maybe scUM pulls it's head out of it's ass and is ranked at seasons end. Maybe Minny has a magical season and OSU is their only loss. It's too tough to say what is going to happen before the season starts.

[Deleted]

May 27, 2014 12:56:57 GMT -5 ohstatefan said:

May 27, 2014 11:36:13 GMT -5 @buckina said:

This. Remember when we backed into the NCG.

Not sure how tough are schedule is this year. Navy gave us a fit a few years back. MSU seems like the toughest game but we barely beat scUM last year.

There are several examples to consider.

2002 pre-season we were ranked #13. We didn't climb up to the #2 spot until week 12 when the only other undefeated team left was Miami. In other words if Oklahoma had kept winning that year, we wouldn't have been in the NC game.

2004 pre-season polls had Auburn #17. But USC and Oklahoma kept winning, so Auburn didn't get a chance at the title.

I shouldn't complain about this too much, since it's the big schools rich in history that usually benefit from such a system. They get noticed more which gives them a better chance of being ranked. We only suffer from not being in the SEC. We usually just need to win all our games to play for the title. The smaller schools or those who haven't been there in a long time also need a fair amount of luck.

Excellent points.

[Deleted]

This is done in all sports in some fashion. Pre season polls, predictions, barometers, whatever are done with all sports at every level. My local paper does it with local high schools. They have special edition inserts with writer predictions and coaches predictions. It's fun, it's part of it and they get a lot more right than wrong just as the cfb ones do.

[ohstatefan]

May 27, 2014 13:29:44 GMT -5 @theref said:

This is done in all sports in some fashion. Pre season polls, predictions, barometers, whatever are done with all sports at every level. My local paper does it with local high schools. They have special edition inserts with writer predictions and coaches predictions. It's fun, it's part of it and they get a lot more right than wrong just as the cfb ones do.

In the other sports or other levels of football, the results of those polls don't have as much of an impact (or any at all) on who gets to be in the playoffs. My complaint isn't that the polls exist, it's how they are used that is the problem.

[Deleted]

May 27, 2014 14:19:32 GMT -5 ohstatefan said:

May 27, 2014 13:29:44 GMT -5 @theref said:

This is done in all sports in some fashion. Pre season polls, predictions, barometers, whatever are done with all sports at every level. My local paper does it with local high schools. They have special edition inserts with writer predictions and coaches predictions. It's fun, it's part of it and they get a lot more right than wrong just as the cfb ones do.

In the other sports or other levels of football, the results of those polls don't have as much of an impact (or any at all) on who gets to be in the playoffs. My complaint isn't that the polls exist, it's how they are used that is the problem.

I get that. My main point is that while I agree it is flawed they still do a pretty good job of getting it right. Yes it is easily argued where in the polls teams belong but the top 25 is generally overall the best 25 teams

[WANG]

May 27, 2014 15:23:39 GMT -5 @theref said:

May 27, 2014 14:19:32 GMT -5 ohstatefan said:

In the other sports or other levels of football, the results of those polls don't have as much of an impact (or any at all) on who gets to be in the playoffs. My complaint isn't that the polls exist, it's how they are used that is the problem.

I get that. My main point is that while I agree it is flawed they still do a pretty good job of getting it right. Yes it is easily argued where in the polls teams belong but the top 25 is generally overall the best 25 teams

What OSF said, the problem is that the preseason polls, in the past, have had an effect on who makes the "playoffs" in college football. The new system will be less so that way, but it will still sway the voters a bit.

I don't have any data on it (and am too lazy too look it up) but I would love to see the starting top 25 versus the final top 25 over the last decade. You may be right, Ref, but I suspect that only about half the teams in the original top 25 are in the final 25.

[Deleted]

May 27, 2014 15:49:11 GMT -5 WANG said:

May 27, 2014 15:23:39 GMT -5 @theref said:

I get that. My main point is that while I agree it is flawed they still do a pretty good job of getting it right. Yes it is easily argued where in the polls teams belong but the top 25 is generally overall the best 25 teams

What OSF said, the problem is that the preseason polls, in the past, have had an effect on who makes the "playoffs" in college football. The new system will be less so that way, but it will still sway the voters a bit.

I don't have any data on it (and am too lazy too look it up) but I would love to see the starting top 25 versus the final top 25 over the last decade. You may be right, Ref, but I suspect that only about half the teams in the original top 25 are in the final 25.

Well all I have is time right now so I did some digging. Its about half way what you and I said. To save time I did top 20.

The BCS era 1998 to 2013 the top 20 pre-season finished in the top 20 -  54% of the time but is a little deceiving because in 12 of the 16 years it was over 60% and in 7 of the 16 it was  70%


From 1989 to to 1997 ( got tired of looking further) it was over 65%

From 1968 to 1988 the AP did only a top 20 pre-season as well as only a top 20 final Those years had a 68.7 % accuracy

Prior to 1968 there were several different types of polls like a preseason top 10 but a final top 20 or a pre top 5 and final 20 plus sevral more. The site that explained that claimed from inception (1936) it was 64.3% but I did not look into how they got that number.

Another interesting part I read was #8 thru 12 finished 12 or higher 79% of the time.


AP POLL MILESTONES:

All time, 1936 to present, Ohio State has appeared in the AP poll more times than anybody at 821 times which is 76.7% of the time with an average ranking of 8.5

We are #1 in top ten appearances with 553 (51% of the time)

We are #2 in top five appearances with 330

We are #3  in weeks ranked #1 with 94

We are one of only 16 schools to be ranked in at least 100 consecutive weeks.

[Deleted]

May 27, 2014 11:36:13 GMT -5 @buckina said:

May 27, 2014 10:29:45 GMT -5 WANG said:

What OSF said. People put waaaaaaay too much stock in preseason polls and are unwilling to move people down a large number, even with poor performances.

For example, pretend that FSU starts this year #1 and runs the table, but due to a weak conference and bad OOC schedule (again, this is just hypothetical, not saying the ACC is bad) they have just an awful SOS. Additionally, they are unimpressive on offense and give up more points than they should. Voters, generally speaking, are going to be unlikely to move them off of their #1 spot simply because they started the season ranked there and never lost.

If voters were more fluid with their rankings from week to week, I wouldn't have a problem with preseason polls, but unfortunately that is not the case.

This. Remember when we backed into the NCG.

Not sure how tough are schedule is this year. Navy gave us a fit a few years back. MSU seems like the toughest game but we barely beat scUM last year.

We never backed into anything. That is the argument that uneducated haters have used. We lost late to Illinois that year and we all thought we were done. But then every team in front of us lost, which allowed us back in the hunt. I just hate that argument.

[WANG]

Thanks for doing the research Ref!

[ohstatefan]

May 27, 2014 18:30:55 GMT -5 @theref said:

May 27, 2014 15:49:11 GMT -5 WANG said:

What OSF said, the problem is that the preseason polls, in the past, have had an effect on who makes the "playoffs" in college football. The new system will be less so that way, but it will still sway the voters a bit.

I don't have any data on it (and am too lazy too look it up) but I would love to see the starting top 25 versus the final top 25 over the last decade. You may be right, Ref, but I suspect that only about half the teams in the original top 25 are in the final 25.

Well all I have is time right now so I did some digging. Its about half way what you and I said. To save time I did top 20.

The BCS era 1998 to 2013 the top 20 pre-season finished in the top 20 -  54% of the time but is a little deceiving because in 12 of the 16 years it was over 60% and in 7 of the 16 it was  70%


From 1989 to to 1997 ( got tired of looking further) it was over 65%

From 1968 to 1988 the AP did only a top 20 pre-season as well as only a top 20 final Those years had a 68.7 % accuracy

Prior to 1968 there were several different types of polls like a preseason top 10 but a final top 20 or a pre top 5 and final 20 plus sevral more. The site that explained that claimed from inception (1936) it was 64.3% but I did not look into how they got that number.

Another interesting part I read was #8 thru 12 finished 12 or higher 79% of the time.


AP POLL MILESTONES:

All time, 1936 to present, Ohio State has appeared in the AP poll more times than anybody at 821 times which is 76.7% of the time with an average ranking of 8.5

We are #1 in top ten appearances with 553 (51% of the time)

We are #2 in top five appearances with 330

We are #3  in weeks ranked #1 with 94

We are one of only 16 schools to be ranked in at least 100 consecutive weeks.

I don't have quite as much time so I only looked at the AP polls from 2004.
Based on the criteria we had outlined above this poll was a success. Out of the top 20 in the preseason poll, 14 were ranked in the top 20 at the end. (5 were unranked and 1 fell to number 23). This makes them 70% correct based on the criteria above.

Only 7 teams finished 5 spots or fewer away from their preseason predictions. The lone exact pick was USC which finished and started #1.
Some of the larger misses were:
LSU started #4, finished #16
FSU started #5, finished #15
OSU started #9, finished #20
West Virginia started #10, finished unranked at 8-4
Florida started #11, finished unranked at 7-5
Kansas State started #12, finished unranked at 4-7
Clemson started #15, finished unranked at 6-5
Louisville started unranked, finished #6 at 11-1
Virginia Tech started unranked, finished #10 at 10-3
Boise State started unranked, finished #12 at 11-1
Auburn started #17, finished at #2 at 13-0

However, despite all those misses, the poll did much better than I could have done.

[WANG]

The question is, were the pollsters influenced by the preseason rankings? Did they rank some teams higher simply because they started out ranked higher? It's a chicken vs egg problem, so we'll never know.