http://www.espn.com/mens-college-basketball/story/_/id/18468008/ncaa-tournament-officials-meet-analytics-experts-consider-creating-new-metric
Would love if they replaced RPI. So many better metrics out there.
I don't like the fact that they are considering "a new metric."
Since Gavitt himself states "current advanced metrics feature their own flaws," and "analytics experts he'll meet next week have similar goals but different theories about the minutiae of a new metric," why do they have to come up with some grand, unified idea?
Why don't they just "institutionalize" the four metrics for the Committee to refer to as part of its process?
How about expanding their criteria to include inviting teams with a rich college BB history that currently aren't good enough to qualify due their record and on court performance. lol
Quote from: Dr. Vinnie Boombatz on January 13, 2017, 01:43:38 PM
I don't like the fact that they are considering "a new metric."
Since Gavitt himself states "current advanced metrics feature their own flaws," and "analytics experts he'll meet next week have similar goals but different theories about the minutiae of a new metric," why do they have to come up with some grand, unified idea?
Why don't they just "institutionalize" the four metrics for the Committee to refer to as part of its process?
I would guess that the appeal with having one metric is transparency--for better or worse, each team and its fans knows where it stands in the RPI. With four different metrics, it will be unclear if a team has disparate rankings across the metrics how it will be treated.
That said, at the end of it all, there is always going to be some subjectivity of the selection committee that means that whatever metric(s) you use, they may ignore in certain situations. So, I think have multiple metrics would not be that problematic.
Quote from: Dr. Vinnie Boombatz on January 13, 2017, 01:43:38 PM
I don't like the fact that they are considering "a new metric."
Since Gavitt himself states "current advanced metrics feature their own flaws," and "analytics experts he'll meet next week have similar goals but different theories about the minutiae of a new metric," why do they have to come up with some grand, unified idea?
Why don't they just "institutionalize" the four metrics for the Committee to refer to as part of its process?
What are the four? RPI, KenPom, Sagarin, and?
The entire discussion is going to be about "how can we merge all of these metrics (RPI, BPI, Sag, Pom) into a single metric" because newsflash, all of those metrics are already being considered by the committee individually.
Quote from: Aughnanure on January 13, 2017, 01:57:46 PM
What are the four? RPI, KenPom, Sagarin, and?
From the article: "Jeff Sagarin (Sagarin), Kevin Pauga (KPI), Ken Pomeroy (KenPom.com) and Ben Alamar (ESPN's BPI)."
Seems like it would be easy enough....add the team's ratings for the four systems and divide by four and *Voila* you have a ranking.
Quote from: Dr. Vinnie Boombatz on January 13, 2017, 01:43:38 PM
I don't like the fact that they are considering "a new metric."
Since Gavitt himself states "current advanced metrics feature their own flaws," ....
Yet, I get crucified for saying the same thing :-\
Quote from: Dr. Vinnie Boombatz on January 13, 2017, 02:26:29 PM
From the article: "Jeff Sagarin (Sagarin), Kevin Pauga (KPI), Ken Pomeroy (KenPom.com) and Ben Alamar (ESPN's BPI)."
Ahh, so actually read it. Thanks.
Quote from: BrewCity83 on January 13, 2017, 04:26:06 PM
Seems like it would be easy enough....add the team's ratings for the four systems and divide by four and *Voila* you have a ranking.
Not that easy given that absolute rank does not reflect relative rank. In other words, to simply average them together would require each metric to be calibrated so that the relative difference between #49 and #50 in one metric is exactly the same as the relative difference between #49 and #50 in all of the other metrics. I'm going to suppose that's statistically impossible.
I'd rather stick with RPI.
Quote from: brandx on January 13, 2017, 05:09:52 PM
Yet, I get crucified for saying the same thing :-\
I won't crucify you, I definitely think this is a questionable proposition.
I'm mixed on this. RPI sucks, but I'd agree with Gavitt's take that there are likely flaws in all the metrics. I think my biggest worry is that the NCAA will try to create their own definitive super-metric and in 10-15 years it will be just as outdated as any other. Getting all the best metric minds today might be a good short-term fix, but will it be a long-term positive?
Also, while I love the idea of a more transparent system, part of the fun of Selection Sunday is the debate over who is most deserving. If all the debate comes down to one straight line metric, I think it takes some of that fun out of the day and you won't get the surprise and excitement of an at-large bid.
I'll be honest, I want to get in to the Tourney this year however we can, but if we end up snagging one of the last bids and get that stunned "we made it moment" I think that can be as much fun as getting a win in the Tournament can be. If you know that your ranking in that system is all that matters, what's the point of watching the Selection Show other than to see where you are going?