From RPI to NET

[Bills_06]

No more RPI for tournament seeding, now the NCAA will use the NCAA Evaluation Tool (NET).  Curious to see how this compares to RPI and get the Wiz's thoughts on it.  

Link

[ACE]

4 minutes ago, Bills_06 said:

No more RPI for tournament seeding, now the NCAA will use the NCAA Evaluation Tool (NET).  Curious to see how this compares to RPI and get the Wiz's thoughts on it.  

Link

Some "mid-majors" were doing too well with RPI rankings, so slowly the NCAA had to figure out how to make it less important. I suspect this new system is designed to make it more difficult for mid-majors. And if mid-majors start to crack the code on this new system, they'll then minimize the importance of NET and turn to the "eye test."

[slufan13]

16 minutes ago, ACE said:

Some "mid-majors" were doing too well with RPI rankings, so slowly the NCAA had to figure out how to make it less important. I suspect this new system is designed to make it more difficult for mid-majors. And if mid-majors start to crack the code on this new system, they'll then minimize the importance of NET and turn to the "eye test."

I agree with this but I also think the RPI sucked. 

[HoosierPal]

This will be interesting to see how they use the NET in the Quadrant System, which has been based on RPI.  I guess they could flip NET in for RPI.

Quad system summary: 

The quadrant system is broken down into four sections, with the top two quadrants having the most importance.

Quadrant one wins are those home games vs. teams RPI ranked 1-30, neutral games vs. 1-50, and away games vs. 1–75. For example, a home win against a top 30 team would qualify as a quadrant one victory.

A Quadrant 2 win are those at Home against the RPI 31–75, neutral 51–100, and away victories over the RPI 76–135.

[Cowboy]

-this line had me paying particular attention.....In addition, the NET will also tweak some data to "ensure fairness."   right, to ensure fairness, I'll buy that along with a certain bridge

[ACE]

3 minutes ago, Cowboy said:

-this line had me paying particular attention.....In addition, the NET will also tweak some data to "ensure fairness."   right, to ensure fairness, I'll buy that along with a certain bridge

"ensure fairness" = "eye test"    

[HoosierPal]

8 minutes ago, ACE said:

"ensure fairness" = "eye test"    

Or putting it in English. Last year, Syracuse's 23-14 just looked better to the committee than St. Mary's 30-6 and Middle TN 25-8.

[Pistol]

I guess I'll need to let it play out to see how it actually compares to other ratings systems, but it seems like the best thing to be is a power conference team that plays a good amount of road games. And those are pretty rare in the current landscape, so I'll at least give it a chance to play out before seeing if it's just another way to bury non-power conference teams.

[Pistol]

Also, a big incentive for the NCAA here is that they own the NET. The RPI was a third-party controlled metric.

They're going to release updated NET rankings weekly starting in late November or early December. Part of me wonders whether this is just going to be like the college football playoff system and that the NCAA will stick to it rigidly when the Tournament brackets are being announced. Might take some drama out of the process, for better or worse.

[Glorydays2013]

1 minute ago, Pistol said:

Also, a big incentive for the NCAA here is that they own the NET. The RPI was a third-party controlled metric.

They're going to release updated NET rankings weekly starting in late November or early December. Part of me wonders whether this is just going to be like the college football playoff system and that the NCAA will stick to it rigidly when the Tournament brackets are being announced. Might take some drama out of the process, for better or worse.

Seems like a college football playoff setup. I think this wont really hurt Mid- Majors since RPI helped out teams in power conferences, if my understanding is right. If teams can't perform in their conference going .500 they shouldn't get at large bids. 

[kshoe]

You would think that the NCAA would release the NET results from 2017-18 to go with this announcement. Might help people better understand what is in the formula and which teams will move up or down based on this new methodology. 

[davidnark]

The RPI was an antiquated formula that was not considered statistically predictive. We will need a season or two of data to understand what types of teams and conferences will most benefit from the change in algorithms, but everyone was long overdue for a more statistically relevant methodology. 

[brianstl]

2 hours ago, HoosierPal said:

This will be interesting to see how they use the NET in the Quadrant System, which has been based on RPI.  I guess they could flip NET in for RPI.

Quad system summary: 

The quadrant system is broken down into four sections, with the top two quadrants having the most importance.

Quadrant one wins are those home games vs. teams RPI ranked 1-30, neutral games vs. 1-50, and away games vs. 1–75. For example, a home win against a top 30 team would qualify as a quadrant one victory.

A Quadrant 2 win are those at Home against the RPI 31–75, neutral 51–100, and away victories over the RPI 76–135.

 

Quote

The quadrant system, which had some critics in its first year of use, will remain in place with the same lines of delineation in the NET and on the team sheets going forward.

 

[Pistol]

52 minutes ago, kshoe said:

You would think that the NCAA would release the NET results from 2017-18 to go with this announcement. Might help people better understand what is in the formula and which teams will move up or down based on this new methodology. 

Yeah, it would be helpful to have that as a comparison.

51 minutes ago, davidnark said:

The RPI was an antiquated formula that was not considered statistically predictive. We will need a season or two of data to understand what types of teams and conferences will most benefit from the change in algorithms, but everyone was long overdue for a more statistically relevant methodology. 

+1

[Compton]

Kinda sucks for the conferences that just totally rejiggered their approach to scheduling (including not knowing how some games are scheduled until much later in the season) to achieve higher RPIs for their top teams. 

[The Wiz]

11 hours ago, Bills_06 said:

No more RPI for tournament seeding, now the NCAA will use the NCAA Evaluation Tool (NET).  Curious to see how this compares to RPI and get the Wiz's thoughts on it.  

Link

I use the same 6 factors that the NCAA uses in their formula.... game results...SOS...game location....scoring margin...Net off and def efficiency and quality of wins and losses. The NET system will also use more than a dozen other factors  including a few RPI factors.    So the NET should be pretty close to what I do....Right?   Maybe...but most likely not.Their weightings will probably be different...Plus I don't have a "Tweak" in my system..."to ensure fairness"...I like to  think my system is fair....I have no need to "ensure" any outcome. For those who are going to sit back and see how this goes...it may be hard to follow as the NCAA will also be watching and tweaking every year...one of the reasons they got rid of their 3rd party...to "ensure" things are going the way they should go. 

An observation...It is interesting to me that results won't be posted till early Dec ...hmmm...that sounds familiar...you mean like the 8th game.

Bottomline....The NCAA will do the same thing I do using the same data . However without an objective  3rd party and little or no transparency  ...instead of coming up with "Wiz results" they will come up with "desired results"  My guess is that this will not bode well for non Power 5 conferences.

[dlarry]

12 hours ago, Compton said:

Kinda sucks for the conferences that just totally rejiggered their approach to scheduling (including not knowing how some games are scheduled until much later in the season) to achieve higher RPIs for their top teams. 

Is that why they did it?

The little guy found away to better it's odds of making the tournament so the NCAA had to put a stop to it?

[kshoe]

9 hours ago, The Wiz said:

I use the same 6 factors that the NCAA uses in their formula.... game results...SOS...game location....scoring margin...Net off and def efficiency and quality of wins and losses. The NET system will also use more than a dozen other factors  including a few RPI factors.    So the NET should be pretty close to what I do....Right?   Maybe...but most likely not.Their weightings will probably be different...Plus I don't have a "Tweak" in my system..."to ensure fairness"...I like to  think my system is fair....I have no need to "ensure" any outcome. For those who are going to sit back and see how this goes...it may be hard to follow as the NCAA will also be watching and tweaking every year...one of the reasons they got rid of their 3rd party...to "ensure" things are going the way they should go. 

An observation...It is interesting to me that results won't be posted till early Dec ...hmmm...that sounds familiar...you mean like the 8th game.

Bottomline....The NCAA will do the same thing I do using the same data . However without an objective  3rd party and little or no transparency  ...instead of coming up with "Wiz results" they will come up with "desired results"  My guess is that this will not bode well for non Power 5 conferences.

The more I think about this the more upset I get that the NCAA hasn't released past seasons as examples of how the rankings would be different. Surely they have back tested this for a number of years (I'd hope 10 or more) so why not share with the world what the data looks like? If you don't want to give away the secret sauce of the formula then fine, let people like the Wiz try to re-engineer it, but you owe it to the schools and the fans to give some context on how things have changed.

Unfortunately, I'm afraid the real reason they haven't released the past data is what the Wiz says above in bold...they have found another way to favorably bias the big schools and don't want to show the past data that proves it.

[kshoe]

14 hours ago, Compton said:

Kinda sucks for the conferences that just totally rejiggered their approach to scheduling (including not knowing how some games are scheduled until much later in the season) to achieve higher RPIs for their top teams. 

I have to think that the rejiggering affect will still be beneficial under the new metric, whatever it is. The rejiggering is designed so that good teams in the conference get more games against the other good teams in their conference. That should be universally beneficial whether you are using the RPI, NET, Sagarin, whatever.

[ACE]

23 hours ago, Bills_06 said:

No more RPI for tournament seeding, now the NCAA will use the NCAA Evaluation Tool (NET).  Curious to see how this compares to RPI and get the Wiz's thoughts on it.  

Link

Tool is in the name... enough said.

[Old guy]

Just keep this in mind when you are thinking of the new rules. If they have introduced a "tweak to insure fairness" they can do with it anything they wish to do. Fudge factors are created to insure that things stay the same, that is the way they are expected to be.

Do you remember or ever heard the song "Henry the VIII I am." I loved their use of this line: "Second verse same as the first." This may well be a situation where the second verse will be the same as the first, depending on how they apply their "tweak."

[billiken_roy]

17 minutes ago, Old guy said:

Just keep this in mind when you are thinking of the new rules. If they have introduced a "tweak to insure fairness" they can do with it anything they wish to do. Fudge factors are created to insure that things stay the same, that is the way they are expected to be.

Do you remember or ever heard the song "Henry the VIII I am." I loved their use of this line: "Second verse same as the first." This may well be a situation where the second verse will be the same as the first, depending on how they apply their "tweak."

wow a rare hermits hermits reference.   i would have bet the farm that billikens.com would not be able to effective use hermits hermits in discussion.   old guy is amazing.  

[almaman]

29 minutes ago, billiken_roy said:

wow a rare hermits hermits reference.   i would have bet the farm that billikens.com would not be able to effective use hermits hermits in discussion.   old guy is amazing.  

Old guy is Dandy I ain’t Telling you Why. 

[Billiken Rich]

Herman's Hermits?

[billiken_roy]

44 minutes ago, Billiken Rich said:

Herman's Hermits?

Rich i was trying to show i wasnt old enough to know better.   dont tell anyone