Started By
Message
Playoff Selection - How to remove human bias?
Posted on 1/9/25 at 10:05 am
Posted on 1/9/25 at 10:05 am
Do you think there's a way to remove human bias from the selection process?
Could blind resumes work?
Would the masses be cool with a computer system doing the selecting?
Could blind resumes work?
Would the masses be cool with a computer system doing the selecting?
This post was edited on 1/9/25 at 10:17 am
Posted on 1/9/25 at 10:08 am to FAT SEXY
the system we had from 2014-2022 worked perfectly fine. The only people that object either don't understand it or have an axe to grind because their team was left out.
2023 was a total cluster.
2023 was a total cluster.
Posted on 1/9/25 at 10:10 am to FAT SEXY
quote:
Do you think there's a way to remove human bias from the selection process?
The only way to remove human bias is to make an algorithm that is strictly data driven. I'm talking, even taking the polls completely out of the equation. If there are problems that arise...modify the equation in the off-season and keep doing that. It will get better over time and eventually should be very effective.
Posted on 1/9/25 at 10:10 am to FAT SEXY
quote:
Would the masses be cool with a computer system doing the selecting?
We had this and it worked better than the committee, but the committee was a response to perceived SEC bias….by the computers. So to answer your question: no. The masses would not be cool with that because they think computers are biased towards the SEC.
Posted on 1/9/25 at 11:24 am to FAT SEXY
What’s wrong with this year’s system? I think it legitimately worked.
Teams left are the 5,6,7 and 8 seeds. Old system would have included 1-4 - none of which will win the championship this year.
Having teams like SMU and Boise make it interesting. Could you argue that Bama should be there instead of SMU? Sure -
Can you argue Bama would win the Natty in this field this year? No.
Teams left are the 5,6,7 and 8 seeds. Old system would have included 1-4 - none of which will win the championship this year.
Having teams like SMU and Boise make it interesting. Could you argue that Bama should be there instead of SMU? Sure -
Can you argue Bama would win the Natty in this field this year? No.
Posted on 1/9/25 at 11:44 am to FAT SEXY
Computers would look strictly at the data (depending there’s no bias somehow put in there).
Posted on 1/9/25 at 11:47 am to WG_Dawg
I disagree with the overall format trying to make 4 slots from 5 groups (& allowing a group to possess multiple slots).
Posted on 1/9/25 at 11:57 am to madmaxvol
quote:
I'm talking, even taking the polls completely out of the equation.
This is a touchy subject. What are people going to say when a team ranked 25th gets in over a team ranked 10th? (Yes, a little over exaggerated, but you get the idea.)
This post was edited on 1/9/25 at 12:10 pm
Posted on 1/9/25 at 12:01 pm to Pvt Hudson
quote:
What’s wrong with this year’s system?
This is evidence it’s not right:
quote:
Teams left are the 5,6,7 and 8 seeds.
Posted on 1/9/25 at 12:14 pm to djsdawg
Says either anybody can beat anybody, or rankings were wrong.
Posted on 1/9/25 at 12:18 pm to FAT SEXY
quote:
Would the masses be cool with a computer system doing the selecting?
A computer system will make selections based on an algorithm that is established by humans.
This is a scary world we live in, based on the things people say.
Posted on 1/9/25 at 12:20 pm to southernboisb
quote:
Says either anybody can beat anybody, or rankings were wrong.
It says the regular season and its championships are meaningless on a bigger scale than it used to be, though ND and OSU likely make a 4 team playoff in 2024
Posted on 1/9/25 at 12:23 pm to MtVernon
quote:
A computer system will make selections based on an algorithm that is established by humans. This is a scary world we live in, based on the things people say.
Would you feel better if AI designed it?
Posted on 1/9/25 at 12:25 pm to djsdawg
This is also just incorrect.
They were seeded 1-4 but they were not actually ranked 1-4.
The teams remaining were ranked #3, #4, #5, and #6
What needs to be fixed is the seeding. Putting Boise fricking State as #3 when they are not really #3 screws up all the brackets
They were seeded 1-4 but they were not actually ranked 1-4.
The teams remaining were ranked #3, #4, #5, and #6
What needs to be fixed is the seeding. Putting Boise fricking State as #3 when they are not really #3 screws up all the brackets
Posted on 1/9/25 at 12:38 pm to GeauxBurrow312
quote:
This is also just incorrect. They were seeded 1-4 but they were not actually ranked 1-4. The teams remaining were ranked #3, #4, #5, and #6
True, and A 4 team playoff would have been Oregon, uga, ND, and OSU.
Posted on 1/9/25 at 12:44 pm to Pvt Hudson
quote:
What’s wrong with this year’s system? I think it legitimately worked.
Teams left are the 5,6,7 and 8 seeds. Old system would have included 1-4 - none of which will win the championship this year.
Having teams like SMU and Boise make it interesting. Could you argue that Bama should be there instead of SMU? Sure -
Can you argue Bama would win the Natty in this field this year? No.
All you did was trade the traditional games that had meaning for teams ranked 8-18 to have meaning.
Ohio St vs Michigan game - completely meaningless.
SECCG - meaningless.
ACCCG - meaningless for an over-ranked SMU team, but magically meaningful for Clemson.
Then these teams win 1 game, or 2 games against other teams not that great and suddenly the system is working because they got in? Bull fricking shite.
You claim teams like Ohio St making it is somehow proof it's working and I say it's proof otherwise.
Posted on 1/9/25 at 12:47 pm to southernboisb
quote:
Would you feel better if AI designed it?
I would assume that is humor but around here who knows.
Posted on 1/9/25 at 12:50 pm to FAT SEXY
I don't know about human bias but we should get rid of the committee. On different occasions they've exercised manipulation to get certain outcomes they wanted before. That's not acceptable.
Posted on 1/9/25 at 12:58 pm to cattus
quote:
I don't know about human bias but we should get rid of the committee. On different occasions they've exercised manipulation to get certain outcomes they wanted before. That's not acceptable.
That’s the ranking system forever. No team should be ranked before they’ve played three games.
Posted on 1/9/25 at 1:00 pm to MtVernon
quote:
A computer system will make selections based on an algorithm that is established by humans.
This is a scary world we live in, based on the things people say.
I use to run a computer ranking system similar to SP+ about 10-15 years ago. I wrote it to determine if teams like Boise St actually belonged in the BCS. The advantage with computers is that the formula is applied to all teams equally.
It's doable, but the majority of computer polls are trash. They also change their formulas constantly because that's how models work. You improve as you go. If you notice something is off, you fix it.
Any good formula should be able to predict the winners of the games a minimum of 75% of the time. And I do not mean #1 should beat #2. I mean they should be able to pick the winner based on how they match up. And 75% isn't a high bar really, because many games aren't hard to predict.
People will want open source and then because of the above people will constantly argue over the formula.
In the end, computer rankings and data should only be a tool for humans.
The real issue with humans is that the people are a bunch of PC pussies who try to find ways to please everyone because "feelings" combined with greed among the schools/conferences wanting to make sure they get a piece of the pie, even if it means they didn't earn it on merit, aka shite like AQ status.
If humans would just be fricking honest and say - hey, you played a shite schedule and you lost to the only good teams you had and ranked them accordingly - no issues. Well, other than 12 teams is still too many.
Popular
Back to top
