NCAA makes change to NET

[JFCats22]

https://www.ncaa.com/news/basketbal...etball-committee-announces-change-net-2020-21

The NCAA Division I Men’s Basketball Committee announced that beginning with the 2020-21 season, the NCAA Evaluation Tool will be changed to increase accuracy and simplify it by reducing a five-component metric to just two. The remaining factors include the Team Value Index (TVI), which is a result-based feature that rewards teams for beating quality opponents, particularly away from home, as well as an adjusted net efficiency rating.

The adjusted efficiency is a team’s net efficiency, adjusted for strength of opponent and location (home/away/neutral) across all games played. For example, a given efficiency value (net points per 100 possessions) against stronger opposition rates higher than the same efficiency against lesser opponents and having a certain efficiency on the road rates higher than the same efficiency at home.

No longer will the NET use winning percentage, adjusted winning percentage and scoring margin. The change was made after the committee consulted with Google Cloud Professional Services, which worked with the NCAA to develop the original NET.

 

[ganner918]

It seemed like they initially threw everything in the model with little rationale. I'm not sure it was predictive of anything. This new model looks like kenpom with boosts for good wins.

 

[kydmb99]

Remember how awful seedings were before the advent of computer rankings?

Yeah, me neither.

 

[YardCat]

Wow, I'm sure this is something not good for UK. Very seldom do any rule changes help us and Cal?

 

[BassProCat]

So is that good or no?

 

[502 Wildcat]

I'm no expert on these things, but it sounds like it doesn't change much for Power 5 teams like us. Rather, teams like Gonzaga that feast on and blowout cupcakes all season will no longer mysteriously hold serve in the Top 5 of the NET.

The winning percentage and margin of victory components were very friendly to the Gonzagas of the world.

 

[ganner918]

I'm no expert on these things, but it sounds like it doesn't change much for Power 5 teams like us. Rather, teams like Gonzaga that feast on and blowout cupcakes all season will no longer mysteriously hold serve in the Top 5 of the NET.

The winning percentage and margin of victory components were very friendly to the Gonzagas of the world.

Gonzaga was #2 in KenPom last year though, and that's 100% just adjusted efficiency, which is a big part of what the NET will be apparently. Interesting that they're moving this way toward a model that is more heavily "how strong is your team," something theoretically more predictive. Because the narrative has been "strength of resume" over "strength of team" for seeding and the RPI and initially the NET reflected this.

 

[GonzoCat90]

It shouldn't change all that much, this just makes it less redundant and removes the stupid boost for winning by more than 10.

 

[CB3UK]

I'm no expert on these things, but it sounds like it doesn't change much for Power 5 teams like us. Rather, teams like Gonzaga that feast on and blowout cupcakes all season will no longer mysteriously hold serve in the Top 5 of the NET.

The winning percentage and margin of victory components were very friendly to the Gonzagas of the world.

Exactly what I was thinking. Hopefully it adds incentive for us to add a few additional good P5 schools to our OOC.

 

[TJ4BigBlue]

I thought the changes coming from the NCAA would be every UK loss counts as a double bump up for the opposing team and a double bump down for us. As well any of our wins, would count as neutral statistics and data!

 

[shutzhund]

https://www.ncaa.com/news/basketbal...etball-committee-announces-change-net-2020-21

The NCAA Division I Men’s Basketball Committee announced that beginning with the 2020-21 season, the NCAA Evaluation Tool will be changed to increase accuracy and simplify it by reducing a five-component metric to just two. The remaining factors include the Team Value Index (TVI), which is a result-based feature that rewards teams for beating quality opponents, particularly away from home, as well as an adjusted net efficiency rating.

The adjusted efficiency is a team’s net efficiency, adjusted for strength of opponent and location (home/away/neutral) across all games played. For example, a given efficiency value (net points per 100 possessions) against stronger opposition rates higher than the same efficiency against lesser opponents and having a certain efficiency on the road rates higher than the same efficiency at home.

No longer will the NET use winning percentage, adjusted winning percentage and scoring margin. The change was made after the committee consulted with Google Cloud Professional Services, which worked with the NCAA to develop the original NET.

If the NCAA thinks this is the answer then I'm against it.

 

[shutzhund]

Exactly what I was thinking. Hopefully it adds incentive for us to add a few additional good P5 schools to our OOC.

Great idea. Take team that has to start over with 4-5 freshmen every year and load their butts up with stronger teams at the beginning of a season.

Any other ideas.

 

[CB3UK]

Great idea. Take team that has to start over with 4-5 freshmen every year and load their butts up with stronger teams at the beginning of a season.

Any other ideas.

 

[Creed Bratton]

But when UK would win on the road and drop 5 spots the experts on here told people who said that made no sense they were stupid. Seems like the NCAA even recognized the model was broken or they wouldn't have fixed it.

 

[westerncat]

Guess we will have to wait and see how this will work for next season.

 

[saxonburgcat]

No longer will the NET use winning percentage, adjusted winning percentage and scoring margin. The change was made after the committee consulted with Google Cloud Professional Services, which worked with the NCAA to develop the original NET.

Great...pay a professional service to develop a worthless model and then pay them again to fix it.

The original idea was as brilliant as the guy who put the screendoor on his submarine.

 

[The_Answer1313]

You know what would be the best but they will never do

Take all of the respected computer systems out there.......and AVERAGE them.

The average of the systems will be more accurate than any specific model.

 

[Ugoff]

I thought the changes coming from the NCAA would be every UK loss counts as a double bump up for the opposing team and a double bump down for us. As well any of our wins, would count as neutral statistics and data!

You may be thinking of the AP poll.

 

[CatOfDaVille]

But when UK would win on the road and drop 5 spots the experts on here told people who said that made no sense they were stupid. Seems like the NCAA even recognized the model was broken or they wouldn't have fixed it.

Came here for this comment.

Pepperidge Farms remembers.

 

[Smashcat]

Palm takeaways.
https://www.cbssports.com/college-b...uild-the-march-madness-bracket-key-takeaways/

 

[The_Answer1313]

It really did not make any sense whatsoever to have that scoring margin component. It made no sense to cap it and it made no sense to even have it considering you were already taking into account net efficiency lol.

 

[irishcat1965]

What I think UK should do is make sure the non conference games don’t have any opponents above the 150-200 range. Having teams in the high 200s piss off the computers.

 

[ganner918]

What I think UK should do is make sure the non conference games don’t have any opponents above the 150-200 range. Having teams in the high 200s piss off the computers.

If we do play them, we need to clobber them. Like win by 40.

 

[JPScott]

Using win percent was always a stupid idea. It's the type of idea that the RPI was based on and it was a stupid idea then. It screams some non-technical administrator's idea of how a ranking model should work.

I was railing against this more than 20 years ago.

Link to Early Argument against RPI

In reality, won-loss percentage reduces the criticality and importance of individual games and reduces to a single amorphous data point which isn't particularly useful and tends to heavily overvalue teams with gaudy records while not providing any useful way of ranking the relatively large number of teams with respectable but not great win % that are closer to the mean.

Glad the NCAA is dropping it, but I can't say I've ever trusted the NCAA to get this right. As someone already suggested, they would be better off just taking the various top ranking models and polls and just averaging them. Or have Las Vegas seed the teams and then bracket them accordingly. There's really no need for a committee of self-anointed administrators either.

 

[silentsaregolden2.0]

Why would it need changing? I mean Purdue was clearly better than us despite their 16-15 record.

My bad, I think that was according to Kenpom. I get my sucky computer models mixed up sometimes.

 

[Wildcat-in-STL]

This is like when I had a toilet all nighter after Taco Bell. So I changed my diet and ate Indian food, curry and spent another bathroom all nighter.


NET was sh*tty. And now they are just gonna switch to a new level of sh*tty because the NCAA gets Everything wrong

 

[wildcatsboston1984]

So is that good or no?

Have to wait and see how it works out since we don’t have access to the formula to run it against previous seasons data.

 

[wildcatsboston1984]

I'm no expert on these things, but it sounds like it doesn't change much for Power 5 teams like us. Rather, teams like Gonzaga that feast on and blowout cupcakes all season will no longer mysteriously hold serve in the Top 5 of the NET.

The winning percentage and margin of victory components were very friendly to the Gonzagas of the world.

Actually for NET margin of victory was capped at 10 so that actually hurt them given they would frequently win conference games by more than that

 

[wildcatsboston1984]

It shouldn't change all that much, this just makes it less redundant and removes the stupid boost for winning by more than 10.

There was no boost for winning by more than 10. The margin of victory metric was capped at 10.

 

[wildcatsboston1984]

You know what would be the best but they will never do

Take all of the respected computer systems out there.......and AVERAGE them.

The average of the systems will be more accurate than any specific model.

There’s already a website that does it

 

[JerseyCat84]

Or whatever helps Duke.

 

[kybassfan]

Just write the names of all the teams, one to an individual sheet of paper. Look the other way while someone tapes a quarter on Pukes page and throw the pages down the stairs. Further you make it the higher you are ranked.

 

[EastCoastYoungCat]

Basically beat somebody worth a damn and especially away from home and it will matter. Otherwise, it won’t matter much.

 

[The_Answer1313]

There’s already a website that does it

Unforunately the NCAA doesn't use it tho

 

[The_Answer1313]

Also FWIW altho the scoring margin component was capped at 10, the net efficiency was not.

So beating a bunch of teams by a ton of points was still being counted.

 

[JPScott]

Unforunately the NCAA doesn't use it tho

True. The truth is that there really is no reason to even have a selection committee. One could very easily just use the average ratings to figure out which teams make the field and how they should be seeded.

After that, a relatively straightforward program could be put together to place the seeded teams largely along the S-curve with some conditions such as trying to minimize travel if possible, ensuring teams from same conference don’t meet each other in early rounds, BYU doesn’t play on Sunday etc.

I’ve long been a proponent of dropping the selection committee & moving to an automated system. There’s a lot of benefits to doing so, including saving a ton of money and being able to run the model quickly that they can wait until all games are completed & they can provide weekly or even daily updates as the season progresses.

 

[morgousky]

How about that time when we didn’t use stupid computer models? Worked pretty good didn’t it?

how about getting FORMER BASKETBALL PLAYERS to use common sense and throw out the hatchet man academic with his stupid algorithms and computer models?

well that wouldn’t work because the insider bullshit wouldn’t be as effective.

Academics read books and graphs well. There’s no place for it in sports. Until it’s stopped the insanity will never end.

 

[hotelblue]

It really did not make any sense whatsoever to have that scoring margin component. It made no sense to cap it and it made no sense to even have it considering you were already taking into account net efficiency lol.

i remember talking about that. amazing they edited it.

 

[JPScott]

How about that time when we didn’t use stupid computer models? Worked pretty good didn’t it?

how about getting FORMER BASKETBALL PLAYERS to use common sense and throw out the hatchet man academic with his stupid algorithms and computer models?

well that wouldn’t work because the insider bullshit wouldn’t be as effective.

Academics read books and graphs well. There’s no place for it in sports. Until it’s stopped the insanity will never end.

I invite you to take a look at the combined models & compare them to what the committee comes up with on a given year. They’re largely very similar, and where they’re not the models generally make more sense from a basketball standpoint IMO.

 

[morgousky]

I invite you to take a look at the combined models & compare them to what the committee comes up with on a given year. They’re largely very similar, and where they’re not the models generally make more sense from a basketball standpoint IMO.

Jon I’m just saying, you take Barkeley and Issel and I promise you the selections would make sense for the first time in years.