It isn't very. NET has meaning (to the Selection Committee) at selection time, but it is more volatile than other metrics during the season. Since we don't know the formula, we can't say why that's the case (I suspect schedule strength has a lot to do with it) but NET tends fluctuate considerably up and down during the season.
As I pointed out in another thread, as the season progresses the NET tends to converge on the Pomeroy ranking. By the end of the season they are quite close together, at least for teams with NCAA aspirations. The point is that the Pomeroy rank is less volatile, thus is a better measure during the season of where the NET might end up.
If you look at last season's final rankings as of Selection Sunday, the Pomeroy rank was within 3, on average, of the NET rank for the top 25 NET teams. For most teams it was even closer, as the average difference was thrown off by large differentials for four of the top 25 teams: Houston (NET=4, POM=15) , Wofford (13 to 19), Buffalo (15 to 22) and Cincinnati (25 to 32) .
The differences between NET and Pomeroy ranking diverge a bit as we move higher in rank, but are still quite close in the "NCAA range" of the NET. The teams ranked 26 to 40 had NET and Pomeroy rankings within 4, on average. Nine of the 15 were within 3. The only teams with differences greater than 5 were UCF, Mississippi, Texas, and NM State.
The divergence between teams 41 to 60 in NET were more pronounced, averaging 8 slots of difference. But the teams chosen from that range are obviously chosen for reasons other than their NET ranking.
As I pointed out in another thread, as the season progresses the NET tends to converge on the Pomeroy ranking. By the end of the season they are quite close together, at least for teams with NCAA aspirations. The point is that the Pomeroy rank is less volatile, thus is a better measure during the season of where the NET might end up.
If you look at last season's final rankings as of Selection Sunday, the Pomeroy rank was within 3, on average, of the NET rank for the top 25 NET teams. For most teams it was even closer, as the average difference was thrown off by large differentials for four of the top 25 teams: Houston (NET=4, POM=15) , Wofford (13 to 19), Buffalo (15 to 22) and Cincinnati (25 to 32) .
The differences between NET and Pomeroy ranking diverge a bit as we move higher in rank, but are still quite close in the "NCAA range" of the NET. The teams ranked 26 to 40 had NET and Pomeroy rankings within 4, on average. Nine of the 15 were within 3. The only teams with differences greater than 5 were UCF, Mississippi, Texas, and NM State.
The divergence between teams 41 to 60 in NET were more pronounced, averaging 8 slots of difference. But the teams chosen from that range are obviously chosen for reasons other than their NET ranking.
