Preseason coaches poll: how accurate?

[FeliSilvestris]

How accurate are preseason coaches polls?
A lot more accurate than you think, if you believe a study by a guy who has a Stanford Ph.D.
He looked at the winning percentage of the higher-ranked team (per preseason poll) in recent BOWL games.

The answer is 60% which is only slightly less than the winning percentage of the team favored by the closing line in the gambling markets.

But what is really surprising is that the (end of season) pre-bowl rankings actually do a bit worse than preseason rankings!

The author thinks that the reason is that as games are played there is a tendency to over-punish the losing team, without properly considering the strength of the opponent. This reduces the accuracy of the poll.

One could argue that 60% is relatively low, because just by flipping a fair coin one should get the winner right 50% of the time. However, these are BOWL games. Bowl games try to match very comparable teams (whether by invitation, or through conference tie-ins), and often succeed. So, predicting the winner of VERY EVEN match-ups 60% of the time is actually quite good.

http://www.footballstudyhall.com/2015/9/10/9295405/college-football-preseason-polls-accuracy

 

[thewildcat2011]

How accurate are preseason coaches polls?
A lot more accurate than you think, if you believe a study by a guy who has a Stanford Ph.D.
He looked at the winning percentage of the higher-ranked team (per preseason poll) in recent BOWL games.

The answer is 60% which is only slightly less than the winning percentage of the team favored by the closing line in the gambling markets.

But what is really surprising is that the (end of season) pre-bowl rankings actually do a bit worse than preseason rankings!

The author thinks that the reason is that as games are played there is a tendency to over-punish the losing team, without properly considering the strength of the opponent. This reduces the accuracy of the poll.

One could argue that 60% is relatively low, because just by flipping a fair coin one should get the winner right 50% of the time. However, these are BOWL games. Bowl games try to match very comparable teams (whether by invitation, or through conference tie-ins), and often succeed. So, predicting the winner of VERY EVEN match-ups 60% of the time is actually quite good.

http://www.footballstudyhall.com/2015/9/10/9295405/college-football-preseason-polls-accuracy

We are at war with Eurasia. We have always been at war with Eurasia.

 

[phatcat]

We are at war with Eurasia. We have always been at war with Eurasia.

Love that book. Reference it often

 

[Wrestlerneverdiesasacatfan]

The preseason polls are directly related to the population at schools. Higher population,higher ranking

 

[FeliSilvestris]

The preseason polls are directly related to the population at schools. Higher population,higher ranking

Did you actually read the linked article (or my brief summary of it)?
Of course you did NOT. Why don't you? You may find it enlightening!

 

[chariots]

By definition the coaches poll can only assign a favorite to games that include one of 25 teams. Many of those teams are demonstratively better than the majority. So if you only picked from them you would naturally have a larger percentage of mismatch games.

There is a large mass of mediocre teams that are harder to rank and therefore harder to predict ... the coaches poll doesn't have to worry about "picking" them. But even put at a severe disadvantage Vegas outperforms the coaches; speaking to the fact that the Coaches Poll is incoherent and the WORST RANKING SYSTEM year after year after year.

Feli, stop trying to validate their poll. It's a nonsense, worthless poll that contains less information than any other in existence.

 

[FeliSilvestris]

By definition the coaches poll can only assign a favorite to games that include one of 25 teams. Many of those teams are demonstratively better than the majority. So if you only picked from them you would naturally have a larger percentage of mismatch games.

Did you actually bother to read the linked article (or my brief summary of it)?
Of course you did NOT. That is completely OBVIOUS.
Why don't you? You may find it enlightening!
The author has a Stanford Ph.D., by the way...just in case that matters to you.

 

[Alaskawildkat]

We are at war with Eurasia. We have always been at war with Eurasia.

Here is a shot I took at Lenin Square during a break in the fighting while on Russia's Sakhalin Island located north of Hokkaido, Japan this summer.

CLICK PHOTO TO ENLARGE:

(The little guy pictured in the photo broke into a big grin when I said hello to him in my broken Russian. Last summer when I was at the Kremlin two grade school age boys broke into laughter when I attempted to answer in Russian their mother's request for directions to the Armory. I guess I have my two years of taking Russian at Northwestern to thank for those comedic consequences.)


With Saturday's upcoming contest between a Big Ten Team and BYU, the following article - on the lighter side- from the same source cited by the OP is worth a view:

http://www.vanquishthefoe.com/byu-f...u-secrets-behind-mormon-lds-missions-athletes