R
Pretty interesting and it'll be interesting to see how it all shakes out. Especially considering they don't really say how they parse two different kinds of measurements - Predictive and Results oriented. If they are combining the two types of measurements that will be very, very difficult to replicate. Do we know if they'll keep posting it weekly like they did with the RPI?
InGoodCo
Well-Known Member
i will be curious if this will stop teams from being able to game the system like the RPI - IE - playing teams with bad RPI's that are like 175 to 250 but avoiding 250+ teams and keeping a high ranking. IMO - once you get past the to 200, they are all fairly bad basketball teams. Also - does this completely make the RPI obsolete? Can we put to rest that if you don't have an RPI of whatever you're out or will that still be a determining factor?
Fryowa
Administrator
That'll never happen. Every ranking system since the dawn of ranking systems has been argued and had holes shot in it from both sides. It's just going to go from RPI to whatever the new one shakes out to be.
InGoodCo
Well-Known Member
right, and I agree - this will just take the place of that - i just wonder again, like, should we not even use RPI moving forward?
SpiderRico
Well-Known Member
I can tell you already that the 10 point cap margin is BS. Say you have 2 teams that have a common opponent that was pretty good. Team A is beating that opponent by 30 points with 6 minutes to go, scrubs for Team A are put in and they end up winning by 20...that's only going to count as a 10 point win. Contrast that with Team B, who struggles with the same opponent....down at the half by 4, but Team B ends up wearing the opponent down in the 2nd half and wins by 10. Once again, it counts as the same 10 point win as Team A, but it was nowhere near the same result. The explanation given is that the NCAA doesn't want teams running up the score. Noble, but misguided, if you're trying to judge how good a team is.
I agree the margin of victory is the biggest problem.
What I'll be really curious to see is how much different this ends up being from the RPI, and how much it might have influenced who made the tournament in previous years. I honestly think it wouldn't make a huge difference.
What I'll be really curious to see is how much different this ends up being from the RPI, and how much it might have influenced who made the tournament in previous years. I honestly think it wouldn't make a huge difference.
deanvogs
Well-Known Member
Hopefully they will start using MANY metrics and that way you don't rely on just one. As you said they all have holes, so why rely on a single one? It hasn't made sense to rely so much on the RPI for years now.
IMO they are going to lean hard into this one.
If you read the article, they basically provided 3 results metrics, (KPI, SOR, and RPI) and 3 predictive/performance metrics (KenPom, Sagarin, BPI) to the committee last season. We don't know how things will be weighted currently but this new NET metric is attempting to provide some sort of middle ground between Preditive and Results based metrics. I think that is about as good as they can do without having the members debate the vagaries between BPI and KPI.
I actually like the idea of using one metric primarily. That way we have a really good idea of who they will select, and if there is a huge outlier, it's something that can be really questioned and investigated.
This is presuming 2 things - we have at least a pretty good idea of what the NET formula is, and the committee makes itself available for real questioning in regard to who they pick for the tourney.
If you read the article, they basically provided 3 results metrics, (KPI, SOR, and RPI) and 3 predictive/performance metrics (KenPom, Sagarin, BPI) to the committee last season. We don't know how things will be weighted currently but this new NET metric is attempting to provide some sort of middle ground between Preditive and Results based metrics. I think that is about as good as they can do without having the members debate the vagaries between BPI and KPI.
I actually like the idea of using one metric primarily. That way we have a really good idea of who they will select, and if there is a huge outlier, it's something that can be really questioned and investigated.
This is presuming 2 things - we have at least a pretty good idea of what the NET formula is, and the committee makes itself available for real questioning in regard to who they pick for the tourney.
