StudentSectionMenace
Well-known member
- Joined
- Oct 27, 2022
- Messages
- 165
- Reaction score
- 287
- Points
- 63
After today’s bowl games I think it’s obvious that the Gophers would breeze through that weak conference.
I'm a huge college hockey fan and the Pairwise rankings aren't perfect but you rarely hear of many complaints. Helps a lot that everyone knows the criteria and no humans are part of the selections. The problem is you need more games and more cross-pollination with other conferences for a computer model like that to work well. The Pairwise in college hockey is not very reliable until 2/3 of the way through the season - 20+ games.None of the teams are that far ahead of the others. Gophers took Penn state to the wire,
Michigan beat Ohio State
Wisconsins had a 4th quarter lead on Oregon
Georgia lost to bama who lost to Michigan
Etc
Objectivity is what should get playoff bids. Not eye test. Eye test is fake.
Need a formula like ncaa hockey has. Need more auto bids and fewer at larges.
Go 16 with 9 auto bids for 9 conference champs. All seeded by a formula for the 7 at larges and the seeds.
High seed hosts first two rounds.
I agree it probably doesn’t work though the premise is interacting. The ratios are still pretty close in terms of conference series vs non which Just makes each data point weighted more heavily and you’d have to go back to 8 game conf schedules (4 non con with no FCS schools as your other games). Trouble becomes there’s a lot more teams so it’s going to get more and more volatile with only 12 games.I'm a huge college hockey fan and the Pairwise rankings aren't perfect but you rarely hear of many complaints. Helps a lot that everyone knows the criteria and no humans are part of the selections. The problem is you need more games and more cross-pollination with other conferences for a computer model like that to work well. The Pairwise in college hockey is not very reliable until 2/3 of the way through the season - 20+ games.
So I'm not sure what the solution is.
Not sure what the right metric is but it should be a metric that’s set at beginning of the yearI'm a huge college hockey fan and the Pairwise rankings aren't perfect but you rarely hear of many complaints. Helps a lot that everyone knows the criteria and no humans are part of the selections. The problem is you need more games and more cross-pollination with other conferences for a computer model like that to work well. The Pairwise in college hockey is not very reliable until 2/3 of the way through the season - 20+ games.
So I'm not sure what the solution is.
Only eight conference games, four non-con with two P-4 required to get that cross-pollination and make it easier to compare conferences. Plus, the variety will be more fun.I'm a huge college hockey fan and the Pairwise rankings aren't perfect but you rarely hear of many complaints. Helps a lot that everyone knows the criteria and no humans are part of the selections. The problem is you need more games and more cross-pollination with other conferences for a computer model like that to work well. The Pairwise in college hockey is not very reliable until 2/3 of the way through the season - 20+ games.
So I'm not sure what the solution is.
The way the SEC is getting savaged online has been awesome. So arrogant and then their teams go out and get beaten by teams they are supposed to beat.I have been getting a lot of belly laughs at the SEC’s expense on the internet and tik tok tonight.
My favorite quote so far:
“I understand South Carolina looks bigger stronger and faster than Illinois anytime they are not on the same field together”
My problem with that system is it still includes human polls. I want (a) no human votes and (b) no margin of victory component. Only thing that should matter is whether you win or lose, not how dominant you look (either to humans or models).Not sure what the right metric is but it should be a metric that’s set at beginning of the year
Old BCS formula with coaches poll, committee top 25, and computer top 25 would be most logical to start in year 1
RPI is your deal thenMy problem with that system is it still includes human polls. I want (a) no human votes and (b) no margin of victory component. Only thing that should matter is whether you win or lose, not how dominant you look (either to humans or models).
Yes it is. I like the way PWR uses RPi. I've always thought one of the elegant things about sports was that when you took the field, all that mattered was whether you won the game or not. Not whether you passed Kirk Herbsteit's eye test or impressed some spreadsheet.RPI is your deal then
I agree and I favor RPI over efficiency metricsYes it is. I like the way PWR uses RPi. I've always thought one of the elegant things about sports was that when you took the field, all that mattered was whether you won the game or not. Not whether you passed Kirk Herbsteit's eye test or impressed some spreadsheet.
Kind of a cross between Zen and the Art of Motorcycle Maintenance and Herb Brooks.I agree and I favor RPI over efficiency metrics
But I prefer the most deserving teams over the best teams
Yes!!! Particularly with what we have done in the portal this time...lost a couple players who could have been contributors (Daniels, Bangura--I 'get it; didn't see the field', but he could have been a contributor this year)...but added WAY more players who are likely to be contributorsThere are two ways to look at it, either the Big was way better than anyone gave it credit for, or the SEC sucks ass. We played really competitive football in a super tight conference, I think this program is built to be a winner/can go blow to blow with anyone if the chips fall the right way.