NCAA Replacing RPI for Ranking Teams

[cmgoods]

 

[Buffnik]

This means that my RPI thread is going to be an even bigger PITA this year.

Not just switching to the NET methodology, but apparently we also have to look at KenPom, Sagarin, ESPN BPI and Kevin Pauga Index.

Would be a good year to win the Pac-12 Tourney as this sorts its way out.

 

[Goose]

Everything I've read so far (granted, not a whole lot) is scaring me on this one. No transparency and no real explanation of what makes the formula up are bad, but the refusal to show how it would have worked in previous years is the scariest to me.

Curious to get @jgisland 's thoughts on this.

 

[Darth Snow]

Smells like corruption and or stupid

 

[BehindEnemyLines]

This means that my RPI thread is going to be an even bigger PITA this year.

Not just switching to the NET methodology, but apparently we also have to look at KenPom, Sagarin, ESPN BPI and Kevin Pauga Index.

Would be a good year to win the Pac-12 Tourney as this sorts its way out.

That espin rater is craptastic. I hope the ncaa doesn't use it as something that feeds their new toy.

 

[ahoelsken]

This means that my RPI thread is going to be an even bigger PITA this year.

Not just switching to the NET methodology, but apparently we also have to look at KenPom, Sagarin, ESPN BPI and Kevin Pauga Index.

Would be a good year to win the Pac-12 Tourney as this sorts its way out.

Sounds like we need to play a tougher OOC than what we did this past year as well.

 

[torero]

Random thought based on no knowledge of anything: maybe this is why we don't have a schedule yet? Had some plans, but if those plans were built on building RPI, this decision changed some of those plans?

 

[CUFan]

Powered by AI and therefore not "readable" sounds like complete and utter BS.

 

[Buffnik]

Random thought based on no knowledge of anything: maybe this is why we don't have a schedule yet? Had some plans, but if those plans were built on building RPI, this decision changed some of those plans?

That makes sense to me. Especially since some of those metrics discount home wins at altitude, which hurts us badly with strength of record. Tad probably looking to play some winnable road games.

 

[hokiehead]

a bit more on the criteria and the approval process before adoption:

The NCAA Evaluation Tool, or NET, will rely on game results, strength of schedule, game location, scoring margin, net offensive and defensive efficiency and the quality of wins and losses.

In late July, use of the NET model was approved following several months of consultation between the Division I Men's Basketball Committee, the National Association of Basketball Coaches, top basketball analytics experts and Google Cloud Professional Services.
...
In order to "ensure fairness" components such as game dates, the order in which contests are placed was omitted from the development of NET, making early- and late-season equal in importance. The model also caps winning margins at 10 points in accounting for the event of a blowout score.

SI link

edit: scoring margin being a factor is interesting (capped at 10pts)

 

[Highlander]

Looks like they are using Machine Learning to try to analyze more statistical data. The refusal to show how previous years rankings would have compared is disturbing. ML can do amazing things, but you have to train your model correctly. The refusal to show the comparison infers deviation, and they don't want to have to defend that publicly.

 

[Buffnik]

All this controversy and I don't think there's any college basketball fan who would say we have ever had an NCAA Tourney that wasn't loaded with compelling matchups and great games. I think that this has gotten way too complicated. All we really need is a criteria that's transparent. Stat geeks will still argue over whether it's the best criteria, but I don't think most of us care. Neither do the coaches. Just tell us what a team's going to be rated on and then follow those guidelines.

 

[ahoelsken]

That makes sense to me. Especially since some of those metrics discount home wins at altitude, which hurts us badly with strength of record. Tad probably looking to play some winnable road games.

I still find it a little odd that we're not playing UNC or DU. We need the bodybag games, and its cheap to bring those two in here.

 

[Buffalo Brad]

All I can say is ugh.....****ing with something that works pretty well seems very dumb but we are talking the NCAA

 

[Goose]

All this controversy and I don't think there's any college basketball fan who would say we have ever had an NCAA Tourney that wasn't loaded with compelling matchups and great games. I think that this has gotten way too complicated. All we really need is a criteria that's transparent. Stat geeks will still argue over whether it's the best criteria, but I don't think most of us care. Neither do the coaches. Just tell us what a team's going to be rated on and then follow those guidelines.

100% the bold.

 

[dyemeduke]

Everything I've read so far (granted, not a whole lot) is scaring me on this one. No transparency and no real explanation of what makes the formula up are bad, but the refusal to show how it would have worked in previous years is the scariest to me.

Curious to get @jgisland 's thoughts on this.

..........wow... ... ...just wow...smh.

 

[Scotch]

All this controversy and I don't think there's any college basketball fan who would say we have ever had an NCAA Tourney that wasn't loaded with compelling matchups and great games. I think that this has gotten way too complicated. All we really need is a criteria that's transparent. Stat geeks will still argue over whether it's the best criteria, but I don't think most of us care. Neither do the coaches. Just tell us what a team's going to be rated on and then follow those guidelines.

100% the bold.

Yup. My knee jerk is to compare this with the BCS, but the huge difference is this really only affects the bubble teams trying to get in. Seeding may be controversial, but the cream will rise to the top after the first weekend regardless. Still though, major headaches are coming for programs like CU.

 

[jsmilkst]

Yeah, this is 100% not a thing that requires ML. Trying to be predictive in this scenario seems pointless, a defined algorithm that you can complain about is sufficient. In the end though, it's probably only going to affect the last 4 spots, so I guess whatever.

 

[dyemeduke]

I agree with you guys that are saying this will ultimately be inconsequential except for maybe 4-8 teams every year, and the tournament is consistently pretty good despite whatever metric is used. While VCU may be an outlier of being a "first four" in and making it to the FF, it usually does not happen. Still, tournament money can mean a lot to some of these bubble programs...a more complete, transparent answer would be appreciated and beneficial.

BUT...I find this ridiculous. I WISH, I could have the freedom at work to tell my clients and bosses - HEY, I have this figured out, and this is the answer. BUT...I can't tell you how I arrived at such answer, nor can I demonstrate how it would have worked on past cases. But, trust me, it's a solid answer and makes sense.... ....even though it's much different from what we've been doing for the last several years.

And considering how much of a cluster**** the NCAA is to me, and many others... .....it's just baffling.

 

[Buffnik]

 

[Brett Fulton]

I agree with you guys that are saying this will ultimately be inconsequential except for maybe 4-8 teams every year, and the tournament is consistently pretty good despite whatever metric is used. While VCU may be an outlier of being a "first four" in and making it to the FF, it usually does not happen. Still, tournament money can mean a lot to some of these bubble programs...a more complete, transparent answer would be appreciated and beneficial.

BUT...I find this ridiculous. I WISH, I could have the freedom at work to tell my clients and bosses - HEY, I have this figured out, and this is the answer. BUT...I can't tell you how I arrived at such answer, nor can I demonstrate how it would have worked on past cases. But, trust me, it's a solid answer and makes sense.... ....even though it's much different from what we've been doing for the last several years.

And considering how much of a cluster**** the NCAA is to me, and many others... .....it's just baffling.

Totally agree with this first sentence. When you are dealing with 68 teams, it is pretty hard to get it all wrong. But the new metric will likely mess up plans for some bubble teams. If I had to guess, I would expect the new and improved bubble to somehow favor larger conference teams. So I expect to see more 17ish win Big10 teams, and less Ohio Valley 24 win teams dancing this year.

I think this one is only for CIA, politicians and apparently NCAA leaders.