i have a theory about those unders you are playing in the NBA this early in the season. my theory is generally scoring will be higher this year than in previous years for two reasons.
1. these players have been playing pick up games for 5 months which has no team defense. those games are all scoring. since there wasn't a preseason teams aren't playing good defense. of course over the course of the first month this will return back to normal.
2. it is harder to play good defense than offense. it takes more energy. teams will take off on defense during this season to conserve energy.
of course the books will adjust to this, but early in the season when your model gives you an under maybe you should take this into account because i am assuming the data you have isn't take this into consideration.