How come we are seeded #7 even though our BPI is 15?

[TyroneJames]

If our BPI is that high, we should be a 3 or 4 seed.

 

[HooperScooper]

Because metrics don't matter for seeding or selection. Here, I'll post this again.

Q: "Why is [insert your team here] seeded in your bracket much lower than its metrics?"

Talk to the Palm: Because metrics aren't that important. To see what is important, read this link here.

But here is a relatively recent example. In 2017, Wichita State fans were killing me because I had them squarely on the bubble while it was in the top 10 in KenPom. The Shockers won the MVC Tournament, ended up 8th in KenPom's rankings and earned a 10-seed in the bracket. If all you have is metrics, you have nothing. The resume has to back it up.

Even Ken Pomeroy thinks using his metrics in the selection process is a bad idea. At a meeting we attended when the NCAA was exploring replacing the RPI, he told the NCAA folks that they should not use his ratings because he is not measuring what they are trying to reward.

Talk to the Palm: Bracketology mailbag explores the race to No. 1 seeds in 2021 NCAA Tournament

Bracketology Expert Jerry Palm answers your questions about the top seeds in March Madness

www.cbssports.com

 

[auror]

Because metrics don't matter for seeding or selection. Here, I'll post this again.

Q: "Why is [insert your team here] seeded in your bracket much lower than its metrics?"

Talk to the Palm: Because metrics aren't that important. To see what is important, read this link here.

But here is a relatively recent example. In 2017, Wichita State fans were killing me because I had them squarely on the bubble while it was in the top 10 in KenPom. The Shockers won the MVC Tournament, ended up 8th in KenPom's rankings and earned a 10-seed in the bracket. If all you have is metrics, you have nothing. The resume has to back it up.

Even Ken Pomeroy thinks using his metrics in the selection process is a bad idea. At a meeting we attended when the NCAA was exploring replacing the RPI, he told the NCAA folks that they should not use his ratings because he is not measuring what they are trying to reward.

Talk to the Palm: Bracketology mailbag explores the race to No. 1 seeds in 2021 NCAA Tournament

Bracketology Expert Jerry Palm answers your questions about the top seeds in March Madness

www.cbssports.com

They matter to an extent, especially for seeding rather than. selection. The chairman said as much after the selection this year. But it's not the only thing that matters, and won't jump you way up past where your resume dictates you should be.

The fact that we got a 7 seed was definitely because of the metrics. We had an 9 seed resume.

 

[billybud]

3-3 vs Quad 2 played in, I think.

 

[bendm]

Even Ken Pomeroy thinks using his metrics in the selection process is a bad idea. At a meeting we attended when the NCAA was exploring replacing the RPI, he told the NCAA folks that they should not use his ratings because he is not measuring what they are trying to reward.

so what is kenpom and NET measuring in its metrics versus what is the ncaa trying to reward in terms of resumes, and where is the disconnect between the two?

what accounts for these large disparities between metrics and resumes e.g. wichita in 2017 and UConn this year?

imo the metrics should be used to inform/interprete the resume. for instance, at the end of the season you rerank every team by some combination of the metrics and then use those rankings to determine quadrant results and SOS etc

 

[pepband99]

BPI is also an ESPN invention - not sure why anyone would think it has any official capacity at all...

 

[dennismenace]

If metrics means nothing he must be alluding to the Big East as being a weak conference for this year. Is there an overall SOS data available for Uconn? The Big East as a whole composite SOS?

 

[auror]

so what is kenpom and BPI measuring in its metrics versus what is the ncaa trying to reward in terms of resumes, and where is the disconnect between the two? shouldnt the metrics provide a check on the resumes and vice versa?

what accounts for these large disparities between metrics and resumes e.g. wichita in 2017 and UConn this year?

imo the metrics should inform the resume. for instance, at the end of the season you rank every team by some combination of the metrics and use those numbers to determine quadrant results

KenPom & BPI primarily input is margin of victory, adjusted for opponent and location.
Resume ranks like Strength of Record, KPI, and WAB are fed just with record and opponent and location, not with score.
The NET was designed to be a combination of the two types.

So the latter is more about who you beat, and the former is also how you beat them. KenPom and the like are more accurate predictors of future performance, but the committee has often said that it's about the most deserving teams, not the best teams that get in. So the latter describe wins/losses more accurately, so they are preferred for selection. And then the committee keeps the former predictive metrics in mind when seeding so as not to screw a good team by putting a team with a unremarkable resume but better actual team in a terribly under-seeded position. But they don't go so far as to move teams like Loyola Chicago into a 3 seed, they just bump them up a couple seedlines.

 

[HooperScooper]

They matter to an extent, especially for seeding rather than. selection. The chairman said as much after the selection this year. But it's not the only thing that matters, and won't jump you way up past where your resume dictates you should be.

The fact that we got a 7 seed was definitely because of the metrics. We had an 9 seed resume.

Yeah, I think you captured this better than I did. Palm's answer did say, "metrics aren't that important," rather than they're not used at all.

 

[uconngb]

If our BPI is that high, we should be a 3 or 4 seed.

Because we didn’t play Fairfield and Quinnipiac

 

[Leebo]

I have never heard a non-ESPN employee talk about BPI in the college basketball world.