NET Update post Purdue

[Dano564]

Where is the cut line ? There is not a cut line established for NET, there was not for RPI either. Plus, people that are in the game as a lifetime devotion, like coaches voted for something better. Something not to be decided by hobbyists who only follow their own team with a bias.

There kinda was a cut line with RPI. If you were top 40 and major conference, you were in.
Below 40 - 65, you had a chance.
It wasn't a rule, but there were few if any instances that fell outside of that.

 

[Dano564]

Said nothing is perfect but measuring these things is required and to do so it was determined by the coaches that this was better than RPI and it is better. Where exactly would you or any fan rank the Gophers right now ? Keep in mind how little you see of other teams and therefore know very little about them. The brain thinks highly of what it wants to be true, we are all humans. Bet if all the coaches watched all the film of all the games they would come up with a great top 50, but as you get more towards 50 you will get more variety as those margins are razor thin. Instead of all the bitching why does someone not come up with a solution that all humans will love equally ! There is not one ! People that do it for a living came up with something better. Research how poor RPI works, Research exactly what NET does, now share with me a better idea.

Coaches evidently don't know crap about math.
A better idea is an unbiased system that rewards you for wins and Strength of Schedule. Like RPI for example.

 

[Some guy]

Coaches evidently don't know crap about math.
A better idea is an unbiased system that rewards you for wins and Strength of Schedule. Like RPI for example.

I like RPI for that reason. I think RPI should’ve been recalibrated.


Current RPI

RPI = (WP * 0.25) + (OWP * 0.50) + (OOWP * 0.25)

What RPI should be recalibrated to

RPI = (WP * 0.50) + (OWP * 0.30) + (OOWP * 0.20)


I know that system would heavily incentivezed scheduling easy wins. But then you couple combine it with the quad wins system to mitigate that and pick teams based on resume based on the new RPI, quad wins, and efficiency metrics.

 

[Gopherbbdude]

The NET ranking is not the "end all be all", it is in the box of tools they will use

 

[Ope3]

The Penn State NET score is just embarrassing.

They must be the best example of a team who loses close often and when they win they win by large margins.
It's puzzling to have a team with an overall losing record in the area of being "in contention" for a tournament bid.

If Penn State played the Gophers on a neutral court what would the point spread be? My hunch is the Nits would be a slight favorite.

They have a favorable NET because the beat Michigan & Maryland decisively, none of there non-conf losses were to dregs except maybe Bradley and have not got blown out often. They now have as many Big 10 road wins as the Gophers.

 

[builtbadgers]

Coaches evidently don't know crap about math.
A better idea is an unbiased system that rewards you for wins and Strength of Schedule. Like RPI for example.

Better read up on exactly what makes up the RPI.

 

[Some guy]

If Penn State played the Gophers on a neutral court what would the point spread be? My hunch is the Nits would be a slight favorite.

They have a favorable NET because the beat Michigan & Maryland decisively, none of there non-conf losses were to dregs except maybe Bradley and have not got blown out often. They now have as many Big 10 road wins as the Gophers.

For sure. Let’s only look at the results of those games but ignore losses to Bradley, de Paul, home Rutgers

They’ve lost 17 times. 17.
Not 7.
Not 10.
17
like more than half the time.

Like if they play a game this year, they usually lose.

 

[Some guy]

Better read up on exactly what makes up the RPI.

Yeah, RPI overemphasizes SOS IMO.

 

[Dano564]

Better read up on exactly what makes up the RPI.

Very well familiar with it.

 

[builtbadgers]

For sure. Let’s only look at the results of those games but ignore losses to Bradley, de Paul, home Rutgers

They’ve lost 17 times. 17.
Not 7.
Not 10.
17
like more than half the time.

Like if they play a game this year, they usually lose.

Again, if i look at all of the information, every measure, wins and losses is where i start and between PSU and the Gophers for a spot in the field it is simply not close, i would pick the Gophers and so will the committee. NET is only one thing that will be looked at, one tool and can be a tiebreaker but it will not come into play between these two teams.

 

[alchemy2u]

The metric they use to determine who is selected for the tournament directly changes the way the game is played. Coaches will change who they play and how the game is played to maximize their chances to get into the tournament!

The NCAA has to take into consideration what is best for basketball. Because they are changing the game by the metric they use.

-if a metric favors wins by 10 points then coaches will play starters until the end and not give other players PT.
- if a metric favors close loses as much as wins then that will change who you schedule.
- if a metric favors shooting percentage then that will change how you play
- if a metric favors defensive effectiveness then that will change how you play.
- if a metric favors.... etc..
- if the metric favors one style of play over winning then it will dictate how teams can play (effectively changing the rules of the game).

NET seems to be more concerned about how you play a game with arbitrary weights of the factors. So it will dramatically change the way the game is played. It looks like they are trying to be too cute in taking judgement out of the equation.

RPI was based upon metrics that are good for the game. It takes into account wins and who you played. It favors teams that play tougher opponents and win, which is exactly what we want in the game of basketball.

Obviously, not everyone will be happy with any metric, but at least use one that doesn’t negatively change the way the game is played!


Sent from my iPad using Tapatalk

 

[Some guy]

Again, if i look at all of the information, every measure, wins and losses is where i start and between PSU and the Gophers for a spot in the field it is simply not close, i would pick the Gophers and so will the committee. NET is only one thing that will be looked at, one tool and can be a tiebreaker but it will not come into play between these two teams.

The problem is the NET isn’t one thing they look at. It is two things.
It is what determines what a quality win is.

8 teams have gotten quad 1 wins by beating penn state.
3 teams have gotten quad 1 wins by beating Minnesota.

That’s messed up

 

[Dano564]

If Penn State played the Gophers on a neutral court what would the point spread be? My hunch is the Nits would be a slight favorite.

They have a favorable NET because the beat Michigan & Maryland decisively, none of there non-conf losses were to dregs except maybe Bradley and have not got blown out often. They now have as many Big 10 road wins as the Gophers.

They have played 30 games this season, and they won 13.

 

[Ope3]

Again, if i look at all of the information, every measure, wins and losses is where i start and between PSU and the Gophers for a spot in the field it is simply not close, i would pick the Gophers and so will the committee. NET is only one thing that will be looked at, one tool and can be a tiebreaker but it will not come into play between these two teams.

Agreed. If Penn St was playing the Gophers heads, I'd go with Penn St. That does not mean they are more NCAA worthy. Gophers have a big edge, mostly based on the Vancouver & US Bank results.

 

[Some guy]

Agreed. If Penn St was playing the Gophers heads, I'd go with Penn St. That does not mean they are more NCAA worthy. Gophers have a big edge, based on results.

FIFY

 

[PoockItInfor6]

We’d beat Penn State 75%-80% of the time on a neutral floor...

 

[Ope3]

We’d beat Penn State 75%-80% of the time on a neutral floor...

I highly doubt that. They Gophers won by 1 at the Barn, at full strength healthwise.

 

[MinnGopher]

Is there a cheatsheet anywhere that would have updated NET rankings next to our opponent and the result? Or even their NET ranking at the time

 

[stocker08]

I highly doubt that. They Gophers won by 1 at the Barn, at full strength healthwise.

PSU also shot nearly 50% from 3pt range. Taking one isolated game and using it to base future results is foolish. Especially in basketball where a hot shooting team can overcome a talent gap on any given day. Keep in mind...our game against PSU came a few days after losing by 27 to Illinois. Bet you figured that our 11 point win over Illinois two weeks later was the upset of the year.


Sent from my SM-N960U using Tapatalk

 

[GopherOhana]

PSU also shot nearly 50% from 3pt range. Taking one isolated game and using it to base future results is foolish. Especially in basketball where a hot shooting team can overcome a talent gap on any given day. Keep in mind...our game against PSU came a few days after losing by 27 to Illinois. Bet you figured that our 11 point win over Illinois two weeks later was the upset of the year.


Sent from my SM-N960U using Tapatalk

Ignoring the one primary data point would be even more foolish. The models suggest the two teams are close. The one game supports this. There is no data that suggest the gophers would beat PSU on a neutral court 75% of the time. Fortunately this isn’t the mlb and we don’t have to watch the two teams play 19 times

 

[Dano564]

Is there a cheatsheet anywhere that would have updated NET rankings next to our opponent and the result? Or even their NET ranking at the time

http://warrennolan.com/basketball/2019/team-net-sheet?team=Minnesota

and here is Penn St's.
http://warrennolan.com/basketball/2019/team-net-sheet?team=Penn-State

If you look closely, you will see we have a better winning percentage overall.
This is mainly because when we play opponents ranked anywhere from 1 to 353, we win more often.

 

[gopherguy15]

PSU also shot nearly 50% from 3pt range. Taking one isolated game and using it to base future results is foolish. Especially in basketball where a hot shooting team can overcome a talent gap on any given day. Keep in mind...our game against PSU came a few days after losing by 27 to Illinois. Bet you figured that our 11 point win over Illinois two weeks later was the upset of the year.


Sent from my SM-N960U using Tapatalk

Have you watched Penn St over the last month or so? I'm no fan of Penn St, but they are playing great right now. Wouldn't surprise me at all if they won a couple games at the Big Ten Tournament next week. Here are their games since the loss to MSU on 1/13:

Lost to Iowa by 7
Lost at MN by 1
Lost to Rutgers by 4
Lost to Purdue by 9
Won at NW by 7
Lost at OSU by 4
Beat Michigan by 6
Lost at Purdue by 12
Beat NE by 24
Won at IL by 7
Beat Maryland by 17
Lost at WI by 4
Won at Rutgers by 1

That sure looks like a team that's competing very well right now...

 

[howeda7]

You have used this argument many times in support of NET, and it has absolutely nothing to do with the efficacy of a tool. People make decisions all the time that may be good or bad. To my mind, NET has enough ridiculous outliers to cast doubt on its ability to be an effective evaluator of team performance - as do the other tools. I don't believe there is a perfect metric. But NET doesn't seem demonstrably better.

+1. Nebraska is a case study in the Net being garbage. It's very obvious that once you build up a certain body of work, nothing you do matters much. You can lose 13 out 15 games and only drop into the 40's. That's a broken tool. Fix it, or go back to the RPI.

 

[SelectionSunday]

Are people actually watching games lately, or just looking at the standings?

If the Gophers and Penn State improbably meet up in Chicago, I think it's likely a pick 'em or a 1-2 point spread either way.

There is a significant difference between the Gophers and Nittany Lions resumes/records, yes, but there's not much difference if any between the teams.

 

[Urbandale]

If the Gophers and Penn State improbably meet up in Chicago, I think it's likely a pick 'em or a 1-2 point spread either way.

Are people actually watching games lately, or just looking at the standings?

There is a significant difference between the Gophers and Nittany Lions resumes & records, yes, but there's not much any difference between the teams.

You are 100% correct. But an NCAA tournament bid shouldn't be based on a predictor (what would happen in a game tomorrow) but on an evaluator (what has the team accomplished).

 

[Dano564]

If the Gophers and Penn State improbably meet up in Chicago, I think it's likely a pick 'em or a 1-2 point spread either way.

Are people actually watching games lately, or just looking at the standings?

There is a significant difference between the Gophers and Nittany Lions resumes & records, yes, but there's not much any difference between the teams.

By your definition is there a big difference between us and Florida St then, "between the teams"?

To me they are essentially the same, but Florida St won more games is all.

 

[SelectionSunday]

You are 100% correct. But an NCAA tournament bid shouldn't be based on a predictor (what would happen in a game tomorrow) but on an evaluator (what has the team accomplished).

I'm not arguing for or against the NET. What I 'm saying in this specific instance, I see Penn State ranked #47 in the NET and the Gophers #54 and based what I've seen recently from the two teams, I don't think, wow, that's a grave injustice. And quite frankly, I don't care, because Penn State has no chance for an at-large bid, while the Gophers have a great chance.

That said, Ill be more than ready to rip the NET if NC State (if it fails to beat a couple of the big boys in ACC tourney) and its over-inflated NET ranking receives an at-large bid over the Gophers.

 

[stocker08]

Ignoring the one primary data point would be even more foolish. The models suggest the two teams are close. The one game supports this. There is no data that suggest the gophers would beat PSU on a neutral court 75% of the time. Fortunately this isn’t the mlb and we don’t have to watch the two teams play 19 times

Maybe. But their 13-17 record tells a much different story.

 

[Dano564]

Most inflated NET's over RPI among top 50'ish team

Team - NET - RPI
North Carolina State 31 117
Florida 34 68
Penn State 49 77
Lipscomb 46 71
Ole Miss 37 61
Ohio State 43 63
Clemson 40 58
Iowa State 17 34
Baylor 36 50
Virginia Tech 12 25
Wisconsin 15 28
Wofford 14 27
Texas 35 47

Most inflated RPI's over NET.

Team NET RPI
Harvard 116 43
Arizona State 68 40
UNCG 58 33
Toledo 60 36
Davidson 69 46
Temple 57 35
Hofstra 67 48
Kansas 18 1
Saint John's 61 45
Minnesota 54 39
VCU 32 20
Washington 33 21
New Mexico State 47 37

 

[JimmyJamesMD]

NC State would have to be worried with a RPI of 117? Wow