I had a little spare time, so I whipped up the tempo free stats for the first five conference games, and divided them between the two losses and three, much referable wins:

Now, this is a small sample size, and some of the differences are obvious - hey, UNC's opponents scored more in the losses than the wins! - but there are a couple of interesting things that pop out. First of all is the assist rate. The Heels got much more efficient at getting points off their passes, and had more assists in each of their three wins than there losses. As a result, they also had significantly fewer offensive rebounds, since when the shots go in, there's nothing to rebound.

Carolina also changed their defense a bit, increasing the number of turnovers they caused without increasing their steals. It looks like, although it's hard to tell in only a few games, that they're no longer overweighting the defense for fast break opportunities as much as shutting down their opponents scoring. There's still a lot of fast breaks to be had off defensive rebounds, but not nearly the overplaying that often gives up points.

Or at least that's what has jumped out at me in the last few games. With a sample size of only five, the numbers are far more open to interpretation. I'll see if it holds up as the season continues.