I'm not sure season-long Pomeroy numbers are all that helpful at this point for Duke, a team that seems to reinvent themselves every two weeks.
In the last 25 days, Duke has completely overhauled their defensive scheme, played 4 games without their highest usage player, moved their starting pg off the ball, returned their highest usage player, and now seemingly will move their starting pg back to the primary ballhandler spot.
The defense is actually the big constant here and I think we can say that even if there is some negative three point regression, it's a very good college basketball defense. But I can't get hung up on whether we get passed by Purdue or jump back in front of them or whatever. It's not really meaningful from a predictive standpoint when there's such uncertainty from half to half. In the last five halves before tonight's second half, I'd imagine Duke was somewhere between 70 and 100 in adjusted offensive efficiency. They follow that up with a 1.4 in tonight's second half.
Obviously very positive, but who the fuck knows what the next game brings.