Duke or Georgia

[CocoHusky]

There were 2 double digit seeds (10 Oregon & 12 Quinnipiac) that made it to the sweet sixteen last year. For there to be more than two this year, 11 Creighton would have to beat 3 UCLA or 12 FGCU would have to beat 4 Stanford. I suppose it’s possible but if it doesn’t happen we are almost exactly where we last year. Have not seen anything in the tournament yet that was drastically different from results in the regular season or was easily predictable by the RPI. If you don't believe in the RPI then you probably don't believe in seedings. But to make up your own seedings to justify "conference dominance or weakness is a r e a c h.

 

[Bear11]

There were 2 double digit seeds (10 Oregon & 12 Quinnipiac) that made it to the sweet sixteen last year. For there to be more than two this year, 11 Creighton would have to beat 3 UCLA or 12 FGCU would have to beat 4 Stanford. I suppose it’s possible but if it doesn’t happen we are almost exactly where we last year. Have not seen anything in the tournament yet that was drastically different from results in the regular season or was easily predictable by the RPI. If you don't believe in the RPI then you probably don't believe in seedings. But to make up your own seedings to justify "conference dominance or weakness is a r e a c h.

The RPI is not a valid way to rank teams or an accurate predictor of results. This has been proven time and again, with analysts on the men's side showing that it is one of the worst predictors of tournament success among the many metrics out there. We follow it because the committee uses it, not because it is a good metric. Look at how Sagarin/Massey ranked the teams that were "upset" and compare to RPI, and you will see the flaws. Every one was ranked higher by RPI.

 

[CocoHusky]

The RPI is not a valid way to rank teams or an accurate predictor of results. This has been proven time and again, with analysts on the men's side showing that it is one of the worst predictors of tournament success among the many metrics out there. We follow it because the committee uses it, not because it is a good metric. Look at how Sagarin/Massey ranked the teams that were "upset" and compare to RPI, and you will see the flaws. Every one was ranked higher by RPI.

Think about what you are saying. The RPI is invalid but the committee continues to use it-probably just to aggravate certain fans. There are other metrics out there that would have better predicted the “upsets” and it is those metrics that the committee should use for seeding? So if the committee uses those other metrics there should be no more “upsets” in the tournament? As a huge UCONN fan sign me up and tell Mississippi state that UCONN want’s a rematch from 2017 and tell Baylor not to bother showing up for the Championship game if they make it that far this year. Upsets happen regardless of which model you use.

 

[Bear11]

Think about what you are saying. The RPI is invalid but the committee continues to use it-probably just to aggravate certain fans. There are other metrics out there that would have better predicted the “upsets” and it is those metrics that the committee should use for seeding? So if the committee uses those other metrics there should be no more “upsets” in the tournament? As a huge UCONN fan sign me up and tell Mississippi state that UCONN want’s a rematch from 2017 and tell Baylor not to bother showing up for the Championship game if they make it that far this year. Upsets happen regardless of which model you use.

Of course there will always be upsets. But if one metric consistently rates certain higher teams compared to another, and then those teams lose, you have to reconsider that metric. Sagarin and Massey are indisputably better at ranking teams than RPI. They aren't "politically correct" but they are better. Simply google "RPI basketball flaws" and you will get a host of articles explaining this.

Sagarin has Georgia, Mizzou, and LSU at 30, 34, and 45 respectively. RPI has them at 26, 24, and 29. RPI had Ohio State at 6, Sagarin at 10. Oregon State is 42 in RPI and 12 in Sagarin. There is a consistent pattern here, and it does not favor RPI.

 

[CocoHusky]

Of course there will always be upsets. But if one metric consistently rates certain higher teams compared to another, and then those teams lose, you have to reconsider that metric. Sagarin and Massey are indisputably better at ranking teams than RPI. They aren't "politically correct" but they are better. Simply google "RPI basketball flaws" and you will get a host of articles explaining this.
Sagarin has Georgia, Mizzou, and LSU at 30, 34, and 45 respectively. RPI has them at 26, 24, and 29. RPI had Ohio State at 6, Sagarin at 10. Oregon State is 42 in RPI and 12 in Sagarin. There is a consistent pattern here, and it does not favor RPI.

Ah... Google, don't know why I did not think of that. When I google I find an equal number of articles advocating for or stating a case against RPI as I do for or against Sagarin. Not I don't know what to think. I also think the committe uses RPI.

 

[Bear11]

Ah... Google, don't know why I did not think of that. When I google I find an equal number of articles advocating for or stating a case against RPI as I do for or against Sagarin. Not I don't know what to think. I also think the committe uses RPI.

I would be interested to see these articles advocating for RPI. I have not seen them.

Basketball RPI: Why it's a lousy way to pick teams for the NCAA Tournament.
The NCAA Is Modernizing The Way It Picks March Madness Teams
The RPI is Not the Real Predictive Indicator
Analysis | RPI appears to have too much influence in the NCAA tournament seeding process
March Madness: Why RPI Is Not a Great Measurement of a Team's Ability
The Ratings Percentage Index Myth
RPI doesn't (and shouldn't) matter

 

[CTtransplant]

I was planning a more in-depth write up on this later, but the RPI is inherently biased toward the SEC and ACC because they only play 16 conference games. I don't know if this was the intent behind the reduced schedules, but it works. It provides a slight boost, and with the small margins in the middle of the RPI, it can move teams up a seed line or 2. Massey has the SEC 4th, and Sagarin 5th. I believe this is an accurate assessment.

Please do the write up. I do not immediately see why this would be so, and would appreciate you giving us more details.

 

[CTtransplant]

I don't have the time to lay out all the math now, but playing 2 fewer conference games means that each ACC and SEC team averages about 1 more win per year. When your entire conference has an additional win, the 75% of RPI that is who you play gets a slight boost. I am not arguing that one conference IS better, but that some have certain advantages in the RPI before any games are played. Since the committee seems to rely heavily on RPI, especially further down the bracket, this is tremendously important. I don't know if it is intentional, but the SEC and ACC are gaming the RPI. We can see this as 3 SEC teams lost to lower seeds, with a couple more close shaves. All of these teams were overseeded according to Massey/Sagarin, but were boosted by a strong, inflated RPI.

Now I see where you are coming from, and you make an interesting point. I hope you find the time to quantify the effect a little more.