Why is the RPI still a thing?

[ckhokie]

I honestly think it's just a vehicle for the committee to pick and choose who they want in the tourney.

I was noticing on Lunardi's Braketology today that VT, after beating NCST and winning @UVA this week, we're 'trending down' into an 11 seed.

We're now 56 in the RPI and I was looking at teams ahead of us. Nebraska has a worse record, a worse SOS, and is 0-5 against top 25 teams. VT is 2-3. Nebraska is ahead of us in RPI.
Buffalo is 11! spots in front of us, are 17-7 (we're 18-7) and their best win is Toledo who is 71, next best win is 124...

I get it, to many of us RPI means nothing, but the committee still references it, and will continue to legitimize their selections using it. How is this logical?

 

[Ickey Shuffle]

Yeah it needs to be thrown away just like the AP. Every single person should just use Kenpom. I like their emphasis on quadrants though.

 

[Ickey Shuffle]

2018 Pomeroy College Basketball Ratings

 

[Hitman Hart]

I agree with the OP, even though I've made a shit ton of posts today referencing the RPI.

I also agree with @Trench Mob in that kenpom is excellent.

 

[ericd7633]

I am probably the only one, but I still really like the RPI, in terms of measuring a resume. The only thing I don't like, and this is probably the case with most metrics is that I don't think you should get more "credit" for beating a team ranked 250 as opposed to bearing a team ranked 335. I think there should be a cap as far as SOS is concerned at least in terms of playing truly awful competition. Any bubble team is going to beat a 250 team or the 340 team. Let's not give one team more credit in that scenario. I just have always liked the RPI. I realize there are some instances that don't make a lot of sense in terms of valuing a team, but the committee does a good job of looking past that.

In regards to the OP's concerns. There is no way Buffalo will get selected as an at large team. They simply haven't beaten anybody of significance. I don't have a problem with where their slotted in the RPI though, because they did have a really good Non conference SOS.

 

[jontaejones]

I am probably the only one, but I still really like the RPI, in terms of measuring a resume. The only thing I don't like, and this is probably the case with most metrics is that I don't think you should get more "credit" for beating a team ranked 250 as opposed to bearing a team ranked 335. I think there should be a cap as far as SOS is concerned at least in terms of playing truly awful competition. Any bubble team is going to beat a 250 team or the 340 team. Let's not give one team more credit in that scenario. I just have always liked the RPI. I realize there are some instances that don't make a lot of sense in terms of valuing a team, but the committee does a good job of looking past that.

In regards to the OP's concerns. There is no way Buffalo will get selected as an at large team. They simply haven't beaten anybody of significance. I don't have a problem with where their slotted in the RPI though, because they did have a really good Non conference SOS.

I'm with you.

RPI measures who you have beaten.

KenPom and BPI measures things like how much you beat up on weaker teams. That might be a measure of the TRUE talent of a team. But I don't think it should be taken into account when seeding teams. UNLESS there's like a tie-breaker or something where Auburn has a 25-4 record and Michigan State has a 25-4 record and both have similar profiles.

The thing RPI doesn't take into account (and correct me if I'm wrong here) is road/neutral victories.

But I believe the NCAA came up with a new formula to take that into account.

 

[jontaejones]

I agree with the OP, even though I've made a shit ton of posts today referencing the RPI.

I also agree with @Trench Mob in that kenpom is excellent.

KenPom is good but too much emphasis on score. Like Villanova would beat a team 105-85 and their KenPom D would sink like 15 points. They got penalized for beating a team by 20 points?

That belies the fact that our defense is actually pretty bad this year, but still ... pace of play and blowing a team out and letting up or bringing in walk ons are factors in how many points per possession you give up.

I think a combination of RPI/BPI would be the best gauge of how good a team really is and likely to win the thing. But they still shouldn't use that for seeding purposes.

 

[Hitman Hart]

KenPom is good but too much emphasis on score. Like Villanova would beat a team 105-85 and their KenPom D would sink like 15 points. They got penalized for beating a team by 20 points?

That belies the fact that our defense is actually pretty bad this year, but still ... pace of play and blowing a team out and letting up or bringing in walk ons are factors in how many points per possession you give up.

I think a combination of RPI/BPI would be the best gauge of how good a team really is and likely to win the thing. But they still shouldn't use that for seeding purposes.

Kenpom is a great tool to use during the bubble season, where you're trying to look up how good a team actually is. Take Washington for example. They were looking mighty pretty last year, but their kenpom was still bad. This past week they lost games against Oregon and Oregon State.

and if you give up 85 points to a team in CBB during regulation, your D ranking should drop.

 

[jontaejones]

I honestly think it's just a vehicle for the committee to pick and choose who they want in the tourney.

I was noticing on Lunardi's Braketology today that VT, after beating NCST and winning @UVA this week, we're 'trending down' into an 11 seed.

We're now 56 in the RPI and I was looking at teams ahead of us. Nebraska has a worse record, a worse SOS, and is 0-5 against top 25 teams. VT is 2-3. Nebraska is ahead of us in RPI.
Buffalo is 11! spots in front of us, are 17-7 (we're 18-7) and their best win is Toledo who is 71, next best win is 124...

I get it, to many of us RPI means nothing, but the committee still references it, and will continue to legitimize their selections using it. How is this logical?

I will admit your guys RPI is a little bit of an anomaly. It should definitely be a little bit higher. Anyway, I think .500 in ACC gets you guys in. Being .500 in the ACC and beating the #1 overall seed on the road would be criminal to leave out.

 

[ckhokie]

In regards to the OP's concerns. There is no way Buffalo will get selected as an at large team. They simply haven't beaten anybody of significance. I don't have a problem with where their slotted in the RPI though, because they did have a really good Non conference SOS.

But who cares if they're not winning? At the end of the day, it's a rating system. Should Buffalo be considered 11 spots better than a team who has 10 'better' wins because the scheduled (but lost to) better teams? I understand why their non-conf schedule might be an important data point, but that shouldn't be a driver behind how you're gauging how good teams are comparatively, especially with the egregious discrepancies. I only really considered VT bc we'll likely be called a bubble team again, so there are likely many more as or more egregious.

 

[ckhokie]

I will admit your guys RPI is a little bit of an anomaly. It should definitely be a little bit higher. Anyway, I think .500 in ACC gets you guys in. Being .500 in the ACC and beating the #1 overall seed on the road would be criminal to leave out.

I don't think we're a shoe-in until we're 10-8, unfortunately. Too many years we 'should' have been in and we're left out, and RPI is always referenced.

The RPI is a terrible metrics guys can cite, legitimizing in their mind that playing some shitty teams (and winning) in November is a bigger knock than going .500 or better in the toughest conference in the nation

 

[ericd7633]

But who cares if they're not winning? At the end of the day, it's a rating system. Should Buffalo be considered 11 spots better than a team who has 10 'better' wins because the scheduled (but lost to) better teams? I understand why their non-conf schedule might be an important data point, but that shouldn't be a driver behind how you're gauging how good teams are comparatively, especially with the egregious discrepancies. I only really considered VT bc we'll likely be called a bubble team again, so there are likely many more as or more egregious.

They are 17-7 though. They just don't have any wins of substance. Which is why a team with that profile, despite that high RPI number is going to be a team that gets left out. They may have had a chance had they run the table, but they suffered two terrible losses. And Buffalo only scheduled one game against a 250+ team, while Va Tech scheduled 6 such games. But that's why teams with Buffalo's resume in the past haven't gotten in.

 

[ericd7633]

I'm with you.

RPI measures who you have beaten.

KenPom and BPI measures things like how much you beat up on weaker teams. That might be a measure of the TRUE talent of a team. But I don't think it should be taken into account when seeding teams. UNLESS there's like a tie-breaker or something where Auburn has a 25-4 record and Michigan State has a 25-4 record and both have similar profiles.

The thing RPI doesn't take into account (and correct me if I'm wrong here) is road/neutral victories.

But I believe the NCAA came up with a new formula to take that into account.

RPI does take into account road/neutral/home locations.

 

[ckhokie]

They are 17-7 though. They just don't have any wins of substance. Which is why a team with that profile, despite that high RPI number is going to be a team that gets left out. They may have had a chance had they run the table, but they suffered two terrible losses. And Buffalo only scheduled one game against a 250+ team, while Va Tech scheduled 6 such games. But that's why teams with Buffalo's resume in the past haven't gotten in.

Do you think Buffalo is better than VT? Or maybe more simply, do you think based on the work each team has done this year, they deserve to be rated 11 spots higher as part of any rating system?

 

[rmilia1]

RPI does take into account road/neutral/home locations.

Yeah that's a tiallu my biggest issue with RPI. They put too much value on road wins . You get basically 1.3 times as much credit for a road win as you do a home since . You're better off beating number 35 on the road than you are beating number 5 at home. P5 teams would actually be better off playing a SoS of 150 ooc and playing majority road/neutral games as opposed to a top 50 SoS with a lot of home games. You get more value beating Drake, UNCG and St Joe's on the road than you do beating Maryland, Butler and Oregon all at home

 

[ericd7633]

Do you think Buffalo is better than VT? Or maybe more simply, do you think based on the work each team has done this year, they deserve to be rated 11 spots higher as part of any rating system?

I think VT is better. I think VT is more deserving of an at large bid. The biggest reason why VT is a bubble team(and rightfully so IMO) is that they scheduled awful in the non conference. VT almost literally played the easiest Non conference schedule in the country.

I'm fine with buffalo being rated higher in this system as it's constituted. Like I said earlier, anything after 250+ should all be considered the same because then you wouldn't get these outliers such as VT and Buffalo.

 

[ericd7633]

Yeah that's a tiallu my biggest issue with RPI. They put too much value on road wins . You get basically 1.3 times as much credit for a road win as you do a home since . You're better off beating number 35 on the road than you are beating number 5 at home. P5 teams would actually be better off playing a SoS of 150 ooc and playing majority road/neutral games as opposed to a top 50 SoS with a lot of home games. You get more value beating Drake, UNCG and St Joe's on the road than you do beating Maryland, Butler and Oregon all at home

I actually don't mind getting more credit for winning games on the road. And the economics make it virtually impossible for P5 teams to play a majority of their OOC games on the road, but I do understand where you're coming from.

 

[ckhokie]

I think VT is better. I think VT is more deserving of an at large bid. The biggest reason why VT is a bubble team(and rightfully so IMO) is that they scheduled awful in the non conference. VT almost literally played the easiest Non conference schedule in the country.

I won't argue with that. Our Non-conf schedule wasn't good.
However, we beat teams we should have, and then are playing decently well in conference. The tourney should be about getting the best teams, not cherry-picking 6 games that were played in November and saying "those weren't hard enough for you. You don't get in"

 

[rmilia1]

I actually don't mind getting more credit for winning games on the road. And the economics make it virtually impossible for P5 teams to play a majority of their OOC games on the road, but I do understand where you're coming from.

Yeah obviously it's not fiscally possible but there needs to be some kind of cutoff where a road win ceases to add so much value. Beating number 150 on the road shouldn't be equivalent to beating number 60 at home. I don't know where that cutoff should be but there's gotta be a way to do it.

 

[ericd7633]

I won't argue with that. Our Non-conf schedule wasn't good.
However, we beat teams we should have, and then are playing decently well in conference. The tourney should be about getting the best teams, not cherry-picking 6 games that were played in November and saying "those weren't hard enough for you. You don't get in"

It's not about cherry picking 6 games in the non conference. But because of all those games against sub 150 competition, Va Tech is only 6-6 against the top 100 of the RPI. Along with a sub 100 loss to St Louis. So obviously as great as the UVA win is, the St Louis loss just about cancels it out. Is a team that can go .500 against the top 100 deserving? If it were today, I'd say probably, but it's why Va Tech is squarely on the bubble.