Reply to thread

Message: [QUOTE="Bone Dog, post: 4498773, member: 12088"] I must admit to having searched the NCAA NET page looking for some sort of SoS factor in the NET, and I can't find one, at least not a direct factor. This has led me to conclude in other posts that SoS (or quality of opponent, as you put it) is expected to fall out indirectly from other factors it does include, and to do so better as more data is collected. This is possible and leads me to believe the NCAA thinks that, although the NET will tend to be inaccurate in the early season, perhaps wildly so, it will improve as the season wears on, and by late February will be more or less accurate. As one example, LSU's NET ranking for this week can hardly be reconciled with their SoS. But it may turn out to be more accurate in a month. If I'm right about this, then it isn't useful to complain about NET inaccuracies in December and January, since the tool isn't designed to be accurate at that time. It also shouldn't matter as long as the seeding decisions for the tournament are not influenced by early season ranking data. All things considered, RPI looked like a more generally reliable tool than NET, even if it had its own flaws. Reading between the lines, I wonder if the reason the NCAA abandoned it was out of a fear that it would encourage gambling. [/QUOTE]