Net vs rpi

[OTRMUSKIE]

CAn someone way way way smarter then me explain why the NET is a better indicator then RPI or vice versa. From looking at the NET it appears to be a reasonable way of deciding how well a team has played vs the RPI which really have too much credit for just wins and not how well you actually played. So is the NET superior compare with other ranking systems?

[X-band '01]

The NET is designed to factor in scoring margin up to a certain point - it won't go beyond a 10-point win or loss, for example. The RPI strictly focuses on who you scheduled and where you scheduled it.

[xubrew]

I don't know if it's "better" or "worse." It depends on how you're using it. If all either one is really being used for is determining how data is sorted onto a page, I personally don't think it matters THAT much.

If you're wondering what the differences are, the RPI is extremely basic. It's basically your record, the records of the teams you've played, and the records of the teams they've played.

The NET has five components that are much more complex. It's offensive and defensive efficiency, a flat winning percentage, and adjusted winning percentage (based on location), margin of victory, and wins over quality opponents. What's also unknown is how each of those variables are factored in. In truth, it's impossible to explain how much each of the five variables are weighted because they are weighted differently for each team. There is sort of an artificial intelligence piece to it that factors in outliers and weighs them less heavily. For instance, if a team has the following rankings in the five different cagegories: 4, 7, 5, 2, 85, it will deemphasize the 85 because it will assume that it is an outlier and not indicative of the team's overall strength.

But, at the end of the day, the committee makeup hasn't change and the committee procedure hasn't changed. It's going to come down to how hard was it for a team to win the games that they won. I suppose the NET can better measure that. That's what we've been told, anyway. It's also supposed to better indicate good teams from outside the power conferences that may not be winning games against top 50 teams, but are winning games that are hard to win, such as on the road against a team that's really good at home.

So, we shall see.

IN short, I don't think it really matters that much which one they use. The selections and seedings will be about the same.