NCAA selection thread | Page 19 | The Boneyard

NCAA selection thread

Digging in a little deeper, I think leaning so heavily on metrics alone for seeding was a mistake by the Committee. There was some serious over-seeding, especially at the 5 and 6 lines (Michigan, Oregon, Missouri, Mississippi, Illinois), while under seeding teams like Creighton/Louisville and UConn. A case could be made that Purdue and Texas A&M are over seeded too.

The metrics have a tendency to reward teams for quality losses while punishing teams for non-quality wins. I am not anti-metrics, but it feels like it is a massive driver in seedings whne it should just be a factor. Does anyone really think Mississippi or Oregon is a threat to go to the Final 8? While this may seem like fun for the SEC and Big 10 conferences to get more teams with big seeds, 1 seed’s like Florida and Auburn now potentially have really tough second round matchups.
 
Stanford is actually a spot worse than Colorado in Pomeroy, and only 4 spots ahead of Colorado in NET. So how is UNC losing to Stanford at home better than WVU losing to Colorado at a neutral site? Yeah, it seems counterintuitive with Stanford being 12-8 in the ACC and Colorado being 3-17 in the Big 12. But the ACC is really terrible, and Colorado wasn't really as bad as their 3-17 Big 12 record suggests. Colorado lost a bunch of close conference games, and as this board knows, they beat UConn OOC.

So, WVU's worst loss is actually better than UNC's worst loss, and WVU's collection of wins is so vastly better than UNC's wins that it's not funny? How can anybody justify this?
 
nelsonmuntz said:
Digging in a little deeper, I think leaning so heavily on metrics alone for seeding was a mistake by the Committee. There was some serious over-seeding, especially at the 5 and 6 lines (Michigan, Oregon, Missouri, Mississippi, Illinois), while under seeding teams like Creighton/Louisville and UConn. A case could be made that Purdue and Texas A&M are over seeded too.

The metrics have a tendency to reward teams for quality losses while punishing teams for non-quality wins. I am not anti-metrics, but it feels like it is a massive driver in seedings whne it should just be a factor. Does anyone really think Mississippi or Oregon is a threat to go to the Final 8? While this may seem like fun for the SEC and Big 10 conferences to get more teams with big seeds, 1 seed’s like Florida and Auburn now potentially have really tough second round matchups.
Expand ...

The committee usually claims that the NET is actually meant to judge a team's opponents rather than the team itself. Basically, your opponent's NET ranking is supposed to matter, but your own NET ranking really isn't supposed to matter.

Other than WVU and UNC, they really didn't seed that much based on a team's own NET ranking. For example, Gonzaga got an #8 seed despite being #8 in NET. Memphis got a #5 seed despite being #50 in NET. Gonzaga would be a big favorite over Memphis on a neutral court, but Memphis still got a much better seed due to having much more impressive wins than Gonzaga did. As somewhat less extreme examples, VCU got an 11 seed despite being #31 in NET, and UCSD got a 12 seed despite #35 in NET.

Really, I don't think that UNC's NET ranking had that much to do with why UNC made the tournament, even though that's the excuse they're using. Bubba Cunningham was going to put his team in the tournament unless they had an overall losing record. Even if UNC was 18-17 and ranked #102 in NET, Cunningham probably would have put UNC in the tournament. Especially since this was already going to be his last year as committee chair, so he didn't even have to worry about being removed as committee chair for picking UNC.
 
alphasigko said:
Marquette also did not lose to the likes of Colorado, Dayton and Seton Hall.

So, offering views on my team that oppose your views qualifies as "hate"? Methinks you have your head where it should not be.
Expand ...

Marquette also lost to Dayton OOC. And Seton Hall was a conference game, so that already would be registered in how UConn finished one game ahead of Marquette in Big East play.

UConn's neutral site loss to Colorado was only a Quad 2 loss and not really as bad as it seems. (Since Colorado lost a bunch of close games, as I discussed on my post about WVU.) However, it is definitely worse than Marquette's loss to Iowa State.

Not sure that beating Wisconsin, Georgia and Purdue is really that much better than beating Texas, Baylor and Gonzaga.

I really don't see a lot in the OOC performances that justifies putting Marquette as a seed higher, when UConn went 2-0 head to head and finished a game better in conference play.
 
Orange said:
Marquette also lost to Dayton OOC. And Seton Hall was a conference game, so that already would be registered in how UConn finished one game ahead of Marquette in Big East play.

UConn's neutral site loss to Colorado was only a Quad 2 loss and not really as bad as it seems. (Since Colorado lost a bunch of close games, as I discussed on my post about WVU.) However, it is definitely worse than Marquette's loss to Iowa State.

Not sure that beating Wisconsin, Georgia and Purdue is really that much better than beating Texas, Baylor and Gonzaga.

I really don't see a lot in the OOC performances that justifies putting Marquette as a seed higher, when UConn went 2-0 head to head and finished a game better in conference play.
Expand ...
They found the cheat code.

Schedule Big 10 teams OOC. High probability that we will win and they will be metric warriors as the season progresses.
 
Hunt for 7 said:
Independent panel will not work. How can we think the selection committee is a valid entity. Cunningham the guy running the thing got a bonus when UNC made the tournament.

There is no way to create a bigger conflict of interest than that there scenario.

Wow I wondered how UNC got in now we know.

This should end all debate about how corrupt the NCAA has become. This is a clear smoking gun. Something needs to change or no credibility will be left.
Expand ...
This is all so stupid because had there been even one massive upset/Cinderella in conference tournaments, we wouldn't be having this conversation.

This was the deepest the committee has ever been able to go into the bubble. None of the teams on the outside really had any business being in, regardless what you think about UNC.
 
JohnFSilver said:
The Big East seeding was fair. They even got five into the tourney.

UConn is on the 8 line, if it beat Creighton would probably have jumped to 7 line ahead of Kansas, marquette, Missouri. If it won the Big East tourney, I think a case for a No. 6. But they lost in the semifinals. UConn's an 8 and deservedly so. .

As far as lack of respect for Big East...The conference is terrible in the bottom half. Seton Hall, Providence, Butler, DePaul were bad.

The size of the league is also an issue. I think we are going to see these teams expand. The SEC got in 14 teams, while the Big East only has 11 teams. I think conferences are going to have to get to 20 for hoops. These smaller conferences playing home and homes is just killing the metrics.

Huskies with 8 games in conference against bad teams this year.
Expand ...
And what happened to Kansas and Missouri? I have no love for Calipari but Kansas seeding was a joke. Drake! So one SEC team in Missouri is done, hoping we vanquish Oklahoma for a second. err third thank you Gonzaga for dismissing Georgia.
 
HuskyHawk said:
How could I imagine that an OU team that went 6-12 in conference would (a) get in and (b) be a 9 seed. I don't think I've ever seen anything as ridiculous as the SEC getting 14 teams in. Good for UConn except being in Florida's bracket.
Expand ...
That is the issue with SOS…play their top 5 in the SEC once or twice and your SOS jumps so you ride the wave. UConn was -4.5 then 5.5 and even saw a flash of 6.5 so betting action is on UConn. Maybe people were tired of betting against us and losing but my flight back home about an hour before tippoff…just enough time to grab some beers, ribs, and watch tonights tilt.
 
6Nattys4Us said:
And what happened to Kansas and Missouri? I have no love for Calipari but Kansas seeding was a joke. Drake! So one SEC team in Missouri is done, hoping we vanquish Oklahoma for a second. err third thank you Gonzaga for dismissing Georgia.
Expand ...
Texas out also.
 
6Nattys4Us said:
That is the issue with SOS…play their top 5 in the SEC once or twice and your SOS jumps so you ride the wave. UConn was -4.5 then 5.5 and even saw a flash of 6.5 so betting action is on UConn. Maybe people were tired of betting against us and losing but my flight back home about an hour before tippoff…just enough time to grab some beers, ribs, and watch tonights tilt.
Expand ...

You could see it in KenPom in real time, where they would go up for a loss to Tennessee at home while UConn would go down for a win over PC on the road. A win should never hurt a team.
 
6Nattys4Us said:
And what happened to Kansas and Missouri? I have no love for Calipari but Kansas seeding was a joke. Drake! So one SEC team in Missouri is done, hoping we vanquish Oklahoma for a second. err third thank you Gonzaga for dismissing Georgia.
Expand ...
Xavier already took care of Texas
 
nelsonmuntz said:
Digging in a little deeper, I think leaning so heavily on metrics alone for seeding was a mistake by the Committee. There was some serious over-seeding, especially at the 5 and 6 lines (Michigan, Oregon, Missouri, Mississippi, Illinois), while under seeding teams like Creighton/Louisville and UConn. A case could be made that Purdue and Texas A&M are over seeded too.

The metrics have a tendency to reward teams for quality losses while punishing teams for non-quality wins. I am not anti-metrics, but it feels like it is a massive driver in seedings whne it should just be a factor. Does anyone really think Mississippi or Oregon is a threat to go to the Final 8? While this may seem like fun for the SEC and Big 10 conferences to get more teams with big seeds, 1 seed’s like Florida and Auburn now potentially have really tough second round matchups.
Expand ...
I think in terms of how I expected them to perform, I think you identified 7 teams I think are not worth their seed: Michigan, Oregon, Missouri, Mississippi, Illinois, Purdue, A&M. I think they also overseeded Louisville and Clemson. I thought that before, but of course it looks like hindsight now.

I think you're right about one of the flaws of metrics. I wonder if there's a way to cap the benefit of a win at something like 20 or 25. Beating a bad team by 45 rather than 35 really doesn't make any difference, but it does play a fairly large role in the metrics.
 
Marat said:


Dwight Howard picks UConn to make it to the title game.
Expand ...
Side note: Dwight Howard went from underrated to massively overrated among the online NBA community. People who watched him in his prime mostly undervalued him and some of his All-NBAs are because there were no other great centers, and people who didn't watch him in his prime just tout his inflated accolades and his one deep run he only made because of the Garnett injury, LeBron having terrible teammates, and an otherwise terrible East. Strange career.
 
nelsonmuntz said:
You could see it in KenPom in real time, where they would go up for a loss to Tennessee at home while UConn would go down for a win over PC on the road. A win should never hurt a team.
Expand ...
If you're trying to make a predictive system, then a team should absolutely be able to make their power rating worse with a loss.

Questioning whether a 10-12 game non-conference season is enough to establish proper baselines for teams before zero sum conference games start is a fair, but unavoidable concern. These systems are not perfect, but they are better than most alternatives given the constraints.
 
Honestly, the most accurate statistical analysis of the game is the point spread.
 
auror said:
If you're trying to make a predictive system, then a team should absolutely be able to make their power rating worse with a loss.

Questioning whether a 10-12 game non-conference season is enough to establish proper baselines for teams before zero sum conference games start is a fair, but unavoidable concern. These systems are not perfect, but they are better than most alternatives given the constraints.
Expand ...
It would be nice to see something different. I do my own mental math plus eye test, but I look at an under 5 point loss to a good team neutral or road as not a negative. Possibly even a positive. Losing at home to anyone is a negative. Losing by double figures to anyone is a big negative, although worse if the team is not good or if it is at home. In short, a win should not add more than a loss takes away if the circumstances are similar.

When a team like OU has (making this up) 16 games against quality opponents, winning 5 of them shouldn't be viewed as some kind of positive accomplishment. That's terrible.
 
auror said:
With a win I meant.
Expand ...
I liked both. lol. A predictive system absolutely should account for how well you play, win or loss.

Ultimately, resume metrics should far more heavily weigh wins and losses.
 
You must log in or register to reply here.

Online statistics

Members online
239
Guests online
2,603
Total visitors
2,842
Totals may include hidden visitors.

Share This Page

Forum statistics

Threads
163,956
Messages
4,376,620
Members
10,168
Latest member
CTFan142


.
..
Top Bottom