(12-03-2017 12:39 PM)quo vadis Wrote:  

(12-03-2017 12:37 PM)RUScarlets Wrote:  Would have been a shock if they went OSU.

I strongly thought Alabama deserved it, but also thought for political reasons, Ohio State would get in.

Alabama was ranked 1 or 2 all season and only dropped to 5 when they lost. Top three teams won and number 4 lost so Bama moves back into the play off. Last year Ohio St moved into the playoff over a 2 loss Penn St that won the Conference. Call it charma.

(12-04-2017 02:32 PM)The Cutter of Bish Wrote:  I think we're arguing semantics when we say "most deserving" versus the "best." The best team in the Big Ten last year, its champion, didn't make the playoff, but the most deserving, statistically speaking, did.

It's all a crappy risk with no baseline is what I'm saying. Wisconsin WAS the 4th best team going into championship week and ahead of Alabama. How does playing and losing to a top 10 team hurt them in a ranking? In basketball, it actually would help.

There can be a substantial difference between "most deserving" and "best." Almost nobody thought FSU was best when they were the only unbeaten a couple years ago. But they were the "most deserving."

(12-05-2017 04:36 PM)ken d Wrote:  

(12-04-2017 06:38 PM)johnbragg Wrote:  

(12-04-2017 09:16 AM)ken d Wrote:  The first seven are identical, and Penn State and USC take turns at #8 and #9. The only practical result that would be different from the selection committee's decisions would be that Notre Dame would get the last NY6 spot instead of Washington.

How much better could the committee have done?

It's not that the committee got it wrong. It's that, when the decision is made by a dozen people in a room, it's human nature for the side the decision goes against to say "No fair" and coming up with theories about how the system was rigged against them.

That's harder to do with a half-dozen computer formulas, a hundred-plus coaches and however many writers get AP poll ballots.

It's not that the committee could have done any better than it did. It's that the result of a complex formula "feels" more impartial than the result of a dozen people deciding.

(12-04-2017 02:14 PM)stever20 Wrote:  AP is being used in this in place of Harris.

I think the Harris Poll only existed because the AP refused to be a party to the BCS, and so they needed a substitute.

If computers are better (or fairer) than people, then why don't they all produce the same result? The very fact that they vary so wildly from each other demonstrates that they are anything but fair or objective.

Computers are objective. That doesn't mean they are "fair." They only consider the data input.

(12-05-2017 05:58 PM)bullet Wrote:  

(12-05-2017 04:36 PM)ken d Wrote:  

(12-04-2017 06:38 PM)johnbragg Wrote:  

(12-04-2017 09:16 AM)ken d Wrote:  The first seven are identical, and Penn State and USC take turns at #8 and #9. The only practical result that would be different from the selection committee's decisions would be that Notre Dame would get the last NY6 spot instead of Washington.

How much better could the committee have done?

It's not that the committee got it wrong. It's that, when the decision is made by a dozen people in a room, it's human nature for the side the decision goes against to say "No fair" and coming up with theories about how the system was rigged against them.

That's harder to do with a half-dozen computer formulas, a hundred-plus coaches and however many writers get AP poll ballots.

It's not that the committee could have done any better than it did. It's that the result of a complex formula "feels" more impartial than the result of a dozen people deciding.

(12-04-2017 02:14 PM)stever20 Wrote:  AP is being used in this in place of Harris.

I think the Harris Poll only existed because the AP refused to be a party to the BCS, and so they needed a substitute.

If computers are better (or fairer) than people, then why don't they all produce the same result? The very fact that they vary so wildly from each other demonstrates that they are anything but fair or objective.

Computers are objective. That doesn't mean they are "fair." They only consider the data input.

And the biases of the programmer. The result is a ranking which reflects the subjective opinion of the programmer as to what criteria are most important.

(12-06-2017 07:46 AM)ken d Wrote:  

(12-05-2017 05:58 PM)bullet Wrote:  

(12-05-2017 04:36 PM)ken d Wrote:  

(12-04-2017 06:38 PM)johnbragg Wrote:  

(12-04-2017 09:16 AM)ken d Wrote:  The first seven are identical, and Penn State and USC take turns at #8 and #9. The only practical result that would be different from the selection committee's decisions would be that Notre Dame would get the last NY6 spot instead of Washington.

How much better could the committee have done?

It's not that the committee got it wrong. It's that, when the decision is made by a dozen people in a room, it's human nature for the side the decision goes against to say "No fair" and coming up with theories about how the system was rigged against them.

That's harder to do with a half-dozen computer formulas, a hundred-plus coaches and however many writers get AP poll ballots.

It's not that the committee could have done any better than it did. It's that the result of a complex formula "feels" more impartial than the result of a dozen people deciding.

(12-04-2017 02:14 PM)stever20 Wrote:  AP is being used in this in place of Harris.

I think the Harris Poll only existed because the AP refused to be a party to the BCS, and so they needed a substitute.

If computers are better (or fairer) than people, then why don't they all produce the same result? The very fact that they vary so wildly from each other demonstrates that they are anything but fair or objective.

Computers are objective. That doesn't mean they are "fair." They only consider the data input.

And the biases of the programmer. The result is a ranking which reflects the subjective opinion of the programmer as to what criteria are most important.

Absolutely, and since different computers - Sagarin, Colley, etc. - produce different results, we know that computer programmers have different opinions about what is most important in ranking a team. That's why it's good to rely on multiple computers, like the BCS did, and throw out the best/worst rankings so as to control for outliers. It's why i look at the Massey Composite.

(12-05-2017 05:58 PM)bullet Wrote:  

(12-05-2017 04:36 PM)ken d Wrote:  

(12-04-2017 06:38 PM)johnbragg Wrote:  

(12-04-2017 09:16 AM)ken d Wrote:  The first seven are identical, and Penn State and USC take turns at #8 and #9. The only practical result that would be different from the selection committee's decisions would be that Notre Dame would get the last NY6 spot instead of Washington.

How much better could the committee have done?

It's not that the committee got it wrong. It's that, when the decision is made by a dozen people in a room, it's human nature for the side the decision goes against to say "No fair" and coming up with theories about how the system was rigged against them.

That's harder to do with a half-dozen computer formulas, a hundred-plus coaches and however many writers get AP poll ballots.

It's not that the committee could have done any better than it did. It's that the result of a complex formula "feels" more impartial than the result of a dozen people deciding.

(12-04-2017 02:14 PM)stever20 Wrote:  AP is being used in this in place of Harris.

I think the Harris Poll only existed because the AP refused to be a party to the BCS, and so they needed a substitute.

If computers are better (or fairer) than people, then why don't they all produce the same result? The very fact that they vary so wildly from each other demonstrates that they are anything but fair or objective.

Computers are objective. That doesn't mean they are "fair." They only consider the data input.

Depends on what you mean by "fair". I think "objective" = "fair". What would be unfair would be if a programmer like Sagarin had a line of code in his program that was deliberately biased towards or against a conference, such as "If conference = Sun Belt, then deduct 5 points from ranking". And I don't think they do that.

The better word for what you're aiming at here is "accurate". A process can be fair, in that the criteria used is chosen without animus towards anyone and with the honest belief that it is a very important factor, AND applied evenly to everyone, but be inaccurate, if those criteria are themselves the wrong ones.

For example, let's say I am interviewing people for a burger-flipping job, and i have the idea that the most important criterion in terms of how good a job someone will do is how tall they are, the taller the better. I believe this not because I hate shorter people, i just honestly believe that taller is better for this job. So I then measure every applicant for the job and rank-order them based on their height. That process would be fair, in that it is evenly applied to everyone, but it would result in the wrong conclusion about who is best qualified (inaccurate) because truth is, height has nothing to do with burger flipping ability.

(12-03-2017 03:53 PM)TerryD Wrote:  I think this puts to an end the "conventional wisdom" that Notre Dame has to join a conference because of the existence of the Playoff Committee.

This makes two years in a row that a non-conference champ or conference championship game participant made the playoffs.

The choice for number 4 has played out well for ND with the second place division winner at 11 and 1 of the P5 conferences making the play offs instead of the Conference Champion at 11 and 2. Last year Ohio State at 11 and 1 made it in over 11 and 2 Penn State and this year 11 and 1 Alabama made it into the play offs over Conference Champ Ohio State at 11 and 2. The way things have panned out ND doesn't have to be undefeated to get into the play off. They can get in at 11 and 1 too. The conference championship games might still help with the extra game though if a 12 and 1 conference champ was picked over ND at 11 and 1. Brand name, SOS, and who the conference champ lost to, and by how much might get so thought though.