Quote:
Originally Posted by HongKongPhooey
According to quad a win at Kansas is same level as a win a UConn they need to rethink these quads.
|
Reposting some stuff from a couple years ago when the groups first came out:
"RPI groups are simply designed to initially sort teams. Sorting hundreds of teams would be next to impossible for a human if you didn't have a small number of criteria to compare. But when splitting hairs between a few teams, then it's easier to compare more details. Still, the number of Group 1 wins is really important even if they're all at the bottom of Group 1, because it puts you on the right side of the first process used to sort teams."
"oh a win at duke is the same as a win at UCF now?...
That's obviously one of the drawbacks, but the system was more designed to answer questions like "is a home win against Maryland better than a win at SMU?" Previously, the answer required seeing that Maryland is top 50 and SMU is not, and then trying to mentally account for the fact that SMU was played on the road to come up with an idea of which is better. A human can't go through that process thousands of times and remember the conclusions when sorting. The new Group system tells you from the very beginning that winning at SMU is better than beating Maryland at home. It should leave less room for human value judgments."