WAB (Wins Above Bubble) | Page 3 | The Boneyard

WAB (Wins Above Bubble)

Chief00 said:
This year the committee will use whatever metrics they can use to get the teams in that they want.
Expand ...
As opposed to prior years when the committee got teams that they didn't want in.
 
RuffRuff said:
What are we angling for here? To jump Duke in the seeding line and take the 1 in the East? I doubt a loss to UM is going to hurt them all that much if they sweep otherwise.
Expand ...
Of course you do.
 
businesslawyer said:
If you know how to plug in hypotheticals, which would be worth more: winning at Nova or beating St Johns at home?
Expand ...
Winning at Nova. Would be worth 0.74. Beating St. John's at home would be worth 0.54.
 
businesslawyer said:
Of course you do.
Expand ...
Do you think Duke losing to the number one team in the country, leader in every metric there is, is going to hurt Duke, on a neutral floor? They're expected to lose. What is up with this place - should vanilla frosting be put on every post?

If they win, I could see some saying they're the overall one.
 
So if we win out, and Duke loses to Michigan and otherwise wins out, at the end of the regular season their WAB would be slightly higher than ours. Meaning they would make up the deficit they would have to us on Sunday morning over the last two weeks of the regular season.
HooperScooper said:
Winning at Nova. Would be worth 0.74. Beating St. John's at home would be worth 0.54.
Expand ...
But what makes WAB feel right is if we go 1-1 in those 2 games (or any 2 games), it doesn’t matter which opponent you beat and which you lose to. Like — I don’t know — any other sport. If we win at Nova and lose to St Johns, we get .74 and lose .46, for a net of .28. If we lose at Nova and beat St Johns at home, we lose .26 and get .54, for a net of .28. Showing that a bubble team would be doing better than expected to split those two games. Isn’t math amazing?
 
.-.
businesslawyer said:
So if we win out, and Duke loses to Michigan and otherwise wins out, at the end of the regular season their WAB would be slightly higher than ours. Meaning they would make up the deficit they would have to us on Sunday morning over the last two weeks of the regular season.

But what makes WAB feel right is if we go 1-1 in those 2 games (or any 2 games), it doesn’t matter which opponent you beat and which you lose to. Like — I don’t know — any other sport. If we win at Nova and lose to St Johns, we get .74 and lose .46, for a net of .28. If we lose at Nova and beat St Johns at home, we lose .26 and get .54, for a net of .28. Showing that a bubble team would be doing better than expected to split those two games. Isn’t math amazing?
Expand ...
Is the WAB the driver of seeding? Why are we so concerned? Is it not just one variable?
 
HooperScooper said:
I would say yes. As of right now Duke has a 0.52 WAB lead, 8.82 to UConn's 8.29. Duke losing to Michigan is not going to cost them a lot of WAB points. For example, the Duke's neutral site loss to Texas Tech was -0.25 so let's say the loss to Michigan is only -0.15.

UConn beating Creighton is not going to gain many WAB points. My guess is about 0.20. But beating Villanova on the road will give UConn a decent amount of WAB points. The prior win over Villanova at home was worth 0.45. The road wins over middling and bad BE teams such as Georgetown, Butler, Xavier, PC, etc. were all worth around 0.50 WAB. So I'd say the win at Villanova would be worth about 0.65.

So 0.65 plus 0.20 is 0.85, which would put UConn at 9.14. Duke would go from 8.82 to 8.67.

Does all this make sense?
Expand ...
Yes-Becoming clearer -Thank you for the clarification. WABABILITY of wins and losses appears to be a metric we've got to embrace as we head to Selection Sunday.
 
RuffRuff said:
Do you think Duke losing to the number one team in the country, leader in every metric there is, is going to hurt Duke, on a neutral floor? They're expected to lose. What is up with this place - should vanilla frosting be put on every post?

If they win, I could see some saying they're the overall one.
Expand ...
No one is saying that. Duke losing to Michigan wouldn’t devalue them at all. But it certainly doesn’t add to their value. Losing a game never adds to your value. While us winning at Nova would add to our value. And if you think us being ahead of Duke is crazy, please share the 3 Duke wins that compare to our win at Kansas and neutral wins against Florida and Illinois. (I’ll give you a hint — those three Duke wins don’t exist.)

The real question is having just had posted an article by a real sportswriter saying the resume metrics are more important than predictive metrics, why are you on a thread about resume metrics if they’re not important to you? (The question is rhetorical for those who have given up on your sincerity by the way.)
 
businesslawyer said:
No one is saying that. Duke losing to Michigan wouldn’t devalue them at all. But it certainly doesn’t add to their value. Losing a game never adds to your value. While us winning at Nova would add to our value. And if you think us being ahead of Duke is crazy, please share the 3 Duke wins that compare to our win at Kansas and neutral wins against Florida and Illinois. (I’ll give you a hint — those three Duke wins don’t exist.)

The real question is having just had posted an article by a real sportswriter saying the resume metrics are more important than predictive metrics, why are you on a thread about resume metrics if they’re not important to you? (The question is rhetorical for those who have given up on your sincerity by the way.)
Expand ...
They're obviously important. Have you seen the NET? They're 2 and we are 10. The gap there is pretty sizeable. They'll also have some chippy games left against the ACC that would probably have equivalent or better value than the remaining BE for us.
 
.-.
RuffRuff said:
Is the WAB the driver of seeding? Why are we so concerned? Is it not just one variable?
Expand ...
Here, read the article I posted above in this post.

HooperScooper

Post in thread 'WAB (Wins Above Bubble)'

Good article on WAB. There's a discussion about UConn in here. They explain why UConn's WAB is much more important than their NET and why the NET is 10 for a team they're projecting as a 1 seed.

If you only have the bandwidth to master one (metric) before Selection Sunday, consider learning Wins Above Bubble. Introduced officially last season to the NCAA Tournament selection process, WAB is a resume-based metric rapidly becoming a staple of the college hoops lexicon.

"I think fans would be probably pretty surprised that the selection of the at-large teams was probably more highly...
  • Like
 
RuffRuff said:
They're obviously important. Have you seen the NET? They're 2 and we are 10. The gap there is pretty sizeable. They'll also have some chippy games left against the ACC that would probably have equivalent or better value than the remaining BE for us.
Expand ...
The NET is not intended to be used this way. It is not used for ranking/seeding teams, it was created as a way to organize the teams into quadrants on the team sheet and then the resume/predictive metrics and records in quadrants are what is used to seed the teams
 
  • Like
Reactions: caw
navery12 said:
The NET is not intended to be used this way. It is not used for ranking/seeding teams, it was created as a way to organize the teams into quadrants on the team sheet and then the resume/predictive metrics and records in quadrants are what is used to seed the teams
Expand ...
So are you waying the WAB is the most predictive index? What would be the point of the NET used for organizing teams if not used itself as a metric?
 
RuffRuff said:
They're obviously important. Have you seen the NET? They're 2 and we are 10. The gap there is pretty sizeable. They'll also have some chippy games left against the ACC that would probably have equivalent or better value than the remaining BE for us.
Expand ...
I have seen the NET. The NET rewards teams for margins. If two teams play 3 times at a neutral site, and Team A wins twice by 1 point and Team B wins the 3rd game by 20, the NET says Team B is better and that Team A’s winning the series is due to good fortune. Literally, Ken Pom has a “luck” metric that “explains” why teams win more games than their NET rankings say they should. Please post the link to the MLB, NFL or NBA standings that use total margin of points or runs to determine who deserves to be rewarded for the better season as opposed to wins and losses. If you can’t point me to where those rankings are used (other than as a predictive metric), maybe it’s time to drop it?

And yes, comparing teams across the country, mere wins and losses without adjusting for schedule isn’t fair either. What WAB and the other resume metrics to is reward wins and punish losses after adjusting them for schedule. Without caring about margins.
 
RuffRuff said:
So are you waying the WAB is the most predictive index? What would be the point of the NET used for organizing teams if not used itself as a metric?
Expand ...
What is wrong with you? Seriously? Again, not a single person is saying that. WAB is not a predictive metric. That’s why it’s called a resume metrics and not a predictive metric. Teams don’t make the playoffs in MLB, NFL or NBA based on how they would be predicted to do in playoffs. They make the playoffs, and get seeded, based on their actual wins and losses. That’s what a resume metrics is. It rewards wins and punishes losses adjusted for schedules but not for margins.

It’s really not that hard. And I’m getting tired of pretending that you’re not trolling.
 
.-.
RuffRuff said:
So are you waying the WAB is the most predictive index? What would be the point of the NET used for organizing teams if not used itself as a metric?
Expand ...
Not at all, WAB is not predictive. It's a resume/results based metric.

IMG_1859.jpeg


The point of grouping teams by NET is that it's borderline impossible to jump back and forth between every ranking across 6 different metrics for every team to gauge if a win is good. Is beating the 30th NET team really that much better than 31st? No. But there has to be a cutoff somewhere to group them
 
RuffRuff said:
So are you waying the WAB is the most predictive index? What would be the point of the NET used for organizing teams if not used itself as a metric?
Expand ...
I guess you still haven’t read the article I posted.
 
We will lose around 0.8 WAB from this loss, falling to 4th. And more importantly, the gap between us and Houston is now quite small, and they have a massive lead in the predictive metrics, meaning they will likely be the 4th 1-seed on Saturday in the bracket reveal.
 
At this point we seem like a 2 seed at best. The team that played Creighton tonight will have a tough time winning @Nova, home vs SJU & Hall. The next 3 game stretch is our toughest of the conference season and it's coming at a time UConn is playing it's worst basketball.
 
1 seed would be unlikely now, but didn't move in the NET, still 4th in WAB. Q2 loss.
 
.-.
BFieldHusky said:
WAB (2/22)

1. Michigan 10.0
2. Duke 9.8
3. Arizona 9.2
4. UConn 8.3
5. Purdue 7.1
6. Houston 6.9
7. Nebraska 6.5
8. Gonzaga 6.5
9. Illinois 6.4
10. Iowa St. 6.1
Expand ...
Sizeable gap - Purdue isn’t getting a 1. I’d say back in control. If they can win out 1 is theirs, after a perfect Saturday.
 
RuffRuff said:
Sizeable gap - Purdue isn’t getting a 1. I’d say back in control. If they can win out 1 is theirs, after a perfect Saturday.
Expand ...
Yes, because our predictive metrics are a little worse, we need a fairly big gap. And right now we've got it.
 
auror said:
Yes, because our predictive metrics are a little worse, we need a fairly big gap. And right now we've got it.
Expand ...
If you listened to the unveiling, you could sense there was heavy emphasis put on top 3 wins. It will be tough for anyone to compete with UConn if all else is close. It was actually a bit surprising given all of ours are first months or so. Hurley is going to stock the early season every year.
 
.-.
You must log in or register to reply here.

Forum statistics

Threads
168,154
Messages
4,554,958
Members
10,438
Latest member
UConnheart


Top Bottom